Biosynthesis of Heparan Sulfate

!"#!$! %#""&#!&&"#% '(' '')*#!++ ! ! "# $ % & ' % % (& & )! % *+ ,& - . '&+

+ "#+ ./, ( 0 1% &+ 2 + + 34 + + 51 6 78$$38#33$87+

0 % )01* ' & & % ' ' & )' * + % % ' ' &'& & ' + 01 &9 ' % ' '

' & ' & & + ,& & - & & % 9 01 & ' + ( 9 % & 01 & 9 & ./, % % + 5 & & ./, % % % ./, ./," ./,: ./,:" ./,:; & % 01 & % + 5 & % & % % 01 9 ' & - ' ./, ./,<./," + ./, % ' & & ./,<./," & & % ' - %% + 1 % ./, ./," 62 & % 01 & ' &+ 5 % - & ./,: 62 ' 01 & + = % ./, 8 - & ./," & ' & -& % ./," ./,: & %% 01 & ' + 1 % 01 % % % - & & & . & - & & ' % & 01 & & & -8 % )" > ? *+ ,& & % & 01 & - '& % & % & -8 01 & - & 01 & &'& % % ' + %% & -8 - % !@!" 9 .A & & +

!"# $%&!'#( )

B "#

5116 #$8#"# 51 6 78$$38#33$87 8# )& << +?+< C D 8#* To my family, especially to Krzysztof LIST OF PAPERS

This thesis is based on following papers:

I Busse M. and Kusche-Gullberg M. In vitro polymerization of heparan sulfate backbone by the EXT proteins. J. Biol. Chem. (2003) 278, 41333-41

II Yamada S., Busse M., Ueno M., Kelly O.G., Skarnes W.C., Sugahara K., Kusche-Gullberg M. Embryonic fibroblast with a gene trap mutation in Ext1 produce short heparan sulfate chains. J. Biol. Chem. (2004) 279, 32134-41.

III Busse M. and Kusche-Gullberg M. EXT1, EXT2 and EXTL3 siRNA gene silencing result in altered heparan sulfate chain length. Manuscript.

IV Busse M. and Kusche-Gullberg M. Heparan sulfate from fibroblasts carrying a hypomorphic mutation in Ext1 exhibit altered interaction with FGF2. Manuscript.

Reprints were made with permission from the publisher. Contents

INTRODUCTION...... 5 BACKGROUND ...... 6 PROTEIN GLYCOSYLATION ...... 6 Diversity of glycosylations...... 6 Glycosyltransferases...... 7 PROTEOGLYCANS ...... 8 Glycosaminoglycans ...... 8 Core proteins of heparan sulfate proteoglycans ...... 10 Functions of heparan sulfate proteoglycans...... 12 BIOSYNTHESIS OF HEPARAN SULFATE ...... 12 GAG-protein linkage region ...... 12 Biosynthesis of the heparan backbone...... 13 Modification of the heparan backbone ...... 13 Interactions between heparan sulfate biosynthetic enzymes ...... 15 Deficiencies in heparan sulfate biosynthetic enzymes ...... 15 EXT FAMILY ...... 16 Discovery of the EXT family of proteins ...... 16 Glycosyltransferase activities of the EXT family of proteins...... 17 Structure of the EXT family of proteins...... 19 Human EXT orthologs ...... 21 Hereditary multiple exostoses and heparan sulfate...... 22 PRESENT INVESTIGATIONS ...... 26 AIM...... 26 RESULTS...... 26 DISCUSSION ...... 29 In vitro polymerization of the heparan backbone...... 30 What is the heparan sulfate polymerase in vivo?...... 30 Mechanism of polymerization ...... 31 Role of EXT2 in heparan sulfate biosynthesis...... 31 Role of EXTL3 in heparan sulfate biosynthesis...... 31 Polymerization and N-sulfation...... 32 Importance of heparan sulfate chain length for HS-protein interactions...... 32 FUTURE PERSPECTIVES ...... 32 ACKNOWLEDGEMENTS ...... 34 REFERENCES...... 36 ABBREVIATIONS

CS chondroitin sulfate ECM extracellular matrix ER endoplasmic reticulum EXTs members of EXT family of proteins EXT1 exostosin 1 EXT2 exostosin 2 EXTL1 exostosin like 1 EXTL2 exostosin like 2 EXTL3 exostosin like 3 FGF fibroblast growth factor GAG glycosaminoglycan Gal galactose GalNAc D-D-N-acetylgalactose Gal-TI galactosyltransferase I Gal-TII galactosyltransferase II GlcA D-D-glucuronic acid GlcA-TI glucuronyltransferase I GlcA-TII glucuronyltransferase II GlcNAc D-D-N-acetylglucosamine GlcNAc-TI N-acetylglucosaminyltransferase I GlcNAc-TII N-acetylglucosaminyltransferase II GPI glycosylphosphatidyl-inositol Hh hedgehog HME hereditary multiple exostoses HS heparan sulfate HSPG heparan sulfate proteoglycan IdoA iduronic acid Ihh Indian hedgehog NDST N-deacetylase/N-sulfotransferase OST O-sulfotransferase PAPS 3’-phosphoadenosine-5’-phosphosulfate PG proteoglycan UDP uridine 5’-diphosphate XT xylosyltransferase Xyl xylose INTRODUCTION

Proteoglycans (PGs) have gained immense attention during recent years, since it has become evident that they play pivotal roles in processes such as signaling and developmental patterning but also pathogenic events. Most PGs are composed of core proteins with covalently attached glycosaminoglycan (GAG) chains. GAGs are unbranched polysaccharides that consist of alternating units of hexuronic acid and hexosamine. Probably the most extensively studied GAG is heparan sulfate (HS). The HS chain can be modified by epimerization and sulfation, resulting in high structural heterogeneity. HSPGs mediate many protein-protein interactions, important for a number of biological processes such as organogenesis, angiogenesis, blood coagulation, growth factor signaling and lipid metabolism. Recognition of the mechanisms controlling the biosynthesis of HS, therefore, is of great importance for understanding the functions of HS.

5 BACKGROUND

PROTEIN GLYCOSYLATION

Posttranslational modifications of proteins are very often essential for their proper functions. Glycosylation – attachment of one or more mono/oligosaccharides is one of the main modifications in the secretory pathway.

Diversity of glycosylations Protein glycosylation can occur on asparagine (in the consensus sequence Asn-X-Ser/Thr), via the nitrogen atom of an amide group - therefore referred to as N-glycosylation or on serine, threonine, hydroxylysine, hydroxyproline and tyrosine, via the oxygen atom of a hydroxyl group – referred to as O- glycosylation. Glycosylation is often required for the proper folding, activity and stability of proteins. N-glycosylation starts in the endoplasmic reticulum (ER) by the addition of a branched oligosaccharide composed of 14 specific monosaccharides; two N-acetylgalactosamines (GalNAc), nine mannoses and three glucoses, derived from a dolichol-oligosaccharide precursor. This oligosaccharide is further modified through trimming by glucosidases and mannosidases and addition of other sugar units by glycosyltransferases in the ER and Golgi apparatus. After these modifications, N-glycans share a common pentasac- charide core structure and based on the composition of the additional sugar units they can be divided into subclasses: high-mannose, complex and hybrid, containing high mannose and complex units. O-glycosylation, in contrast to N-glycosylation, where the initial oligosaccharide is transferred in one single step, is a step-by-step process; all sugars units are added sequentially from corresponding nucleotide sugars. There are many possible ways to O-glycosylate proteins (Spiro, 2002; Van den Steen et al., 1998). In mucins, GalNAc is the first sugar added to regions in proteins that are rich in prolines, serines or threonines. In the epidermal growth factor and coagulation factors, serine and threonine residues can be

6 modified by the addition of mannose or fucose (alone or as inner sugars of oligosaccharide structures). Hydroxylysine of collagens can be substituted with galactose or glucose-galactose structures. Glycogenin, the core protein for glycogen molecules, has the unusual property of catalyzing its own glycosylation, attaching a glucose to a tyrosine residue in the protein. The attached glucose is believed to serve as the primer required by glycogen syn- thase (Lomako et al., 2004). Among the most fascinating forms of protein- sugar conjugates are the PGs, in which the GAG chains are O-linked to a serine by the first sugar in the chain, xylose (Xyl). Additional forms of glycosylation have been described, such as C- glycosylation, attachment of D-mannose to tryptophan via a C-C bond (Furmanek and Hofsteenge, 2000); P-glycosylation, addition of sugar to protein through a phosphodiester bond; the glycosylphosphatidyl-inositol (GPI) membrane anchor, formed between mannose and the C-terminus of a protein via phosphoethanolamine (Spiro, 2002). One example of a GPI- anchored protein is the HSPG glypican (described below). Some types of protein glycosylation are believed to take place outside the secretory pathway, e.g. the addition of GalNAc to serine or threonine in nuclear and cytosolic proteins (Hart, 1997).

Glycosyltransferases Glycosyltransferases catalyze the biosynthesis of glycoconjugates by the transfer of sugar moieties from activated donor molecules to specific acceptors. The acceptor can be either a growing oligosaccharide, a protein or a lipid. Many Golgi-localized glycosyltransferases share the common topology of type II-transmembrane proteins. They consist of a short N-terminal cytoplasmic tail, followed by a transmembrane domain, a stem region and a large C-terminal catalytic part (Figure 1) (Breton and Imberty, 1999; Paul- son and Colley, 1989). According to the catalytic mechanism by which they perform reactions, glycosyltransferases can be divided into two groups: retaining and inverting. Retaining enzymes transfer UPD-sugar onto an acceptor without a change in the anomeric configuration of carbon C1 on the do- nor sugar. In contrast, inverting glycosyltransferases change the anomeric configuration of the carbon C1 on the donor sugar. Additionally, glycosyltransferases are grouped into different families based on the sequence simi- larity (Carbohydrate-Active-enZYmes (CAZy) homepage: http://afmb.cnrs- mrs.fr/CAZY/).

7 Figure 1. Common topology of Golgi glycosyltransferases: CT- cytoplasmic tail, TM-transmembrane domain, SR-stem region, CD-catalytic domain.

PROTEOGLYCANS

Proteoglycans are a class of O-glycosylated proteins, with one or more GAG chains attached to the core protein. The functions of PGs depend both on the attached GAG chains and on the core proteins. PGs bind a whole range of proteins and are essential for many intermolecular interactions.

Glycosaminoglycans GAGs are linear polysaccharides, composed of alternating hexuronic acid: glucuronic acid (GlcA) or iduronic acid (IdoA) and hexosamine: glucosamine (GlcNAc) or galactosamine (GalNAc) (Figure 2). Based of the nature of this repeating disaccharide unit, GAGs can be divided into four main groups. HS/heparin is composed of alternating hexuronic acid and GlcNAc, which can be modified with sulfate groups at several positions. Chondro- itin/dermatan sulfate (CS/DS) is composed of hexuronic acid and GalNAc and can be modified with sulfate groups in various positions. Keratan sulfate (KS) lacks the hexuronic acid moiety and is build of alternating, partially sulfated galactose and GlcNAc units. Hyaluronan, in contrast to the other GAGs, is not linked to a core protein and is composed of repeating GlcA and GlcNAc units that are not modified by sulfation. GAGs have been reported to be present in a variety of animal organisms. The nematode, Caenorhabditis elegans, synthesizes non-sulfated chondro-

8 itin but sulfated HS chains (Toyoda et al., 2000; Yamada et al., 1999). Sul- fated heparin-like GAGs are found in many other invertebrates, except for sponges (Medeiros et al., 2000). Nevertheless, nonsulfated GAGs are present in sponges: cell-cell adhesion in these organisms is mediated by a particular family of PGs, named spongicans, the glycan subunit structures of which resemble hyaluronan (Fernandez-Busquets and Burger, 2003). GAGs are also produced by bacteria – for example the E. coli K5 strain, associated with urinary tract infection in humans, produces a capsule of extracellular polysaccharides, which are structurally identical with mammalian non- modified heparan chains (Figure 2). As the capsule polysaccharide is similar or identical to the host GAG, the immune response to bacterial infection is very limited and thus renders the bacteria more virulent (Jann and Jann, 1992). The advantage of such a bacterial strain for researchers is the avail- ability of the K5 polysaccharide, which has been used to study HS biosynthesis and for production of heparin-related drugs (Lindahl et al., 2005).

Figure 2. Common structure of different glycosaminoglycans; R: H, COCH3 or SO3- R1: OH or OSO3-.

9 Core proteins of heparan sulfate proteoglycans HSPGs are produced along the secretory pathway (Prydz and Dalen, 2000; Uhlin-Hansen et al., 1997) and most are secreted out of the cell. Secreted PGs can be found either associated with the cell membrane – cell surface PGs (glypicans and syndecans) or in the extracellular matrix (e.g. perlecan) (Figure 3). The HS-attachment site in the core protein generally consists of a Ser-Gly dipeptide with one or more flanking acidic amino acids. Hydrophobic amino acids in close proximity to the attachment site have been suggested to promote HS synthesis (Esko and Zhang, 1996).

Intracellular HSPG Serglycin is present in the secretory granules of mast cells, where it is involved in storage of proteases (Henningsson et al., 2002). This PG is modified with a highly sulfated variant of HS (heparin). Some HSPGs have been reported to localize in the nucleus, probably as a result of the internalization of cell surface PGs (Kolset et al., 2004).

Cell surface HSPG Glypicans are covalently linked to the plasma membrane by a GPI anchor. They are localized within rafts, small lipid patches mainly built up of choles- terol and sphingomyelin. Lipid rafts are suggested to act as sites of cell signaling events (Grzybek et al., 2005). Glypicans are rich in cysteines in their extracellular domain; the position of the 14 cysteine residues is conserved among the glypicans indicating potential structural similarities. HS attachment sites in glypican are clustered close to the plasma membrane, followed by a globular ectodomain. Glypicans can be cleaved within the GPI anchor by phospholipases, or in the protein by proteases, resulting in the release of their ectodomains. Six glypican isoforms have been found in vertebrates, two in D. melanogaster and one in C. elegans (Bernfield et al., 1999; Hacker et al., 2005; Kramer and Yost, 2003). Syndecans, of which four isoforms have been found in vertebrates, one in the fly and one in the worm, are type I transmembrane proteins, with GAG attachment sites most often localized in the N-terminal part of the protein. The short, conserved cytoplasmic domain contains tyrosines, which can be phosphorylated. Syndecans are shed by proteolytic cleavage at the site close to their transmembrane domain. The released ectodomains can still act in HSPG-dependent processes, as agonists or antagonists of ligand binding (Bernfield et al., 1999; Hacker et al., 2005; Kato et al., 1998; Kramer and Yost, 2003).

10 Extracellular HSPG There are three main HSPGs present in the extracellular matrix (ECM): perlecan, agrin, and collagen XVIII (Iozzo, 2005). Perlecan is a large PG with HS attachment sites localized at the N-terminal and C-terminal parts of the protein. Perlecan can interact through its core protein and/or its HS chains with many ECM components such as HS-binding growth factors, laminins, collagens and integrins. Agrin is a regulator of the formation of postsynaptic structures and the neuromuscular junction. Collagen XVIII is expressed in the basement membrane region that connects basement membrane with subadjacent connective tissue.

Figure 3. Schematic depiction of different heparan sulfate proteoglycans. Perlecan is an example of an extracellular matrix HSPG, glypicans and syndecans are found on the surface of many cells. Serglycin is a heparin-PG found in storage granules of connective tissue type mast cells.

11 Functions of heparan sulfate proteoglycans HSPGs interact with many different proteins, such as proteases and their inhibitors, growth factors and morphogens, chemokines and adhesion proteins and can therefore mediate and regulate various cellular events. Regula- tion of cell growth and differentiation by HSPGs is particulary marked during development. Lack of proper HS results in a series of developmental defects and organisms devoid of HS die during early embryonic stages (see chapter “Deficiencies in heparan sulfate biosynthetic enzymes”). HSPGs also play central roles in pathological conditions such tumor metastasis, virus/bacterial adhesion and invasion (Sasisekharan et al., 2002; Spillmann, 2001) The significance of HSPGs for cell-signaling has become an extensively studied field. Signaling molecules that are active during development, such as fibroblast growth factors (FGFs), Wingless (Wnt/Wg), hedgehog (Hh), transforming growth factor – beta (TGF-E) and bone morphogenic protein (BMP), need HSPG for proper function. HSPGs serve as co-receptors for growth factors mediating their efficient binding and activation of the cognate receptors. HSPGs also participate in the generation of morphogen gradients (Bernfield et al., 1999; Blair, 2005; Hacker et al., 2005; Whitelock and Iozzo, 2005). It has been observed that some defects caused by the mutations in glypican and in syndecan are overlapping but some are unique, suggesting that not only the HS chains but also the protein core is essential for the biological functions of HSPGs (Rawson et al., 2005). Signaling through the core protein of syndecans has been shown to regulate cytoskeletal organization (Yoneda and Couchman, 2003). Perlecan is an essential molecule for embryonic development, since mice deficient in this protein die at embryonic day 10 to 12 (E10 - E12) or just after birth, with major defects in cartilage formation (Arikawa-Hirasawa et al., 1999; Costell et al., 1999).

BIOSYNTHESIS OF HEPARAN SULFATE

GAG-protein linkage region HS and CS chain biosynthesis is initiated by the formation of the so called GAG-protein linkage region - a tetrasaccharide linked to a serine residue in the core protein (GlcAE1-3GalE1-3GalE1-4XylE1-O-Ser). The formation of

12 the HS/CS-protein linkage region is a multi-step reaction. First, a xylose residue is added by xylosyltransferase (XT) (Gotting et al., 2000) in the late endoplasmic reticulum compartments and in the Golgi apparatus (Vertel et al., 1993). Then, two distinct galactosyltransferases, Gal-TI and Gal-TII, each add one galactose (Almeida et al., 1999; Bai et al., 2001; Okajima et al., 1999). The last sugar in the HS- and CS-protein linkage region is GlcA, transferred by glucuronyltransferase I (GlcA-TI) (Kitagawa et al., 1998). It has been found that the linkage region can by modified by the phopshoryla- tion of the xylose residue, a modification common for both HS and CS. On the other hand, the sulfation of the second galactose has been found to occur only in CS chains. Both modifications may be important in determining whether HS or CS is synthesized onto the linkage region (Ueno et al., 2001). Interestingly, it has been observed that not all linkage regions are used as primers for GAG biosynthesis (Nadanaka et al., 1998).

Biosynthesis of the heparan backbone The synthesis of the HS backbone is initiated by the addition of a single D- GlcNAc residue to the polysaccharide-protein linkage region by GlcNAc-TI. This reaction can be catalyzed in vitro by EXTL2 and EXTL3, proteins belonging to the EXT family (Kim et al., 2001; Kitagawa et al., 1999). Alter- natively, it has been speculated that such transfer may be catalyzed by the EXT1/EXT2 (EXT1/2) complex (Kim et al., 2003). After the addition of the first GlcNAc, HS chain elongation is believed to occur by the alternate addition of GlcA and GlcNAc to the non-reducing end of the polymer. The corresponding transfer reactions are also catalyzed by members of the EXT family, namely EXT1, EXT2, EXTL1 and EXTL3 (EXTL1 and EXTL3 can transfer only GlcNAc residues) (reviewed in (Esko and Selleck, 2002; Sugahara and Kitagawa, 2002; Zak et al., 2002) (Figure 5).

Modification of the heparan backbone As the HS chain is polymerized, it also undergoes modifications (Esko and Selleck, 2002; Habuchi et al., 2004; Kusche-Gullberg and Kjellen, 2003; Lindahl et al., 1998) (Figure 4). First, the N-acetyl group of selected GlcNAc residues is replaced with a sulfate group, in a reaction catalyzed by bifunctional enzymes: glucosaminyl N-deacetylase/N-sulfotransferase (NDST) proteins (Aikawa and Esko, 1999; Aikawa et al., 2001; Eriksson et al., 1994; Kjellen et al., 1992; Kusche- Gullberg et al., 1998; Wei et al., 1993). Four isoforms of NDST exist in vertebrates (NDST1-4). In mouse, NDST1 and 2 are widely expressed both during embryonic development and in adult animals, whereas isoforms 3 and 4 are expressed mainly during embryonic development. N-sulfation has been shown to promote further modifications to the heparan backbone (Kjellen,

13 2003). Epimerization of GlcA into IdoA is catalyzed by glucuronyl C-5 epimerase (Crawford et al., 2001; Li et al., 1997; Li et al., 2001).

Figure 4. Schematic picture of heparan sulfate modifications.

2-O-sulfation, of mainly IdoA but also some GlcA, is catalyzed by 2-O- sulfotransferase (2OST) (Kobayashi et al., 1997; Rong et al., 2000). 6-O- and 3-O-sulfation of glucosamine units are performed by the respective sulfotransferases (6OST and 3OST). In vertebrates, epimerase and 2OST occur in single isoforms, whereas 6OST and 3OST have been found in multiple isoforms. All 6OST isoforms show similar enzymatic specificities (Smeds et al., 2003). Interestingly, 3OST catalyzes the rarest modification reaction; addition of sulfate groups at carbon 3 of glucosamine units, but it has the highest number of isoforms among HS biosynthesis enzymes (Shworak et al., 1999). HS chains can also contain a few glucosamine units with a free amino group, but the mechanism of their formation is enigmatic (van den Born et al., 1995; Westling and Lindahl, 2002). Recent findings show that HS sulfation patterns can be additionaly remodeled by the removal of 6-O sulfate groups by extracellular HS 6-O endosulfatases (Ai et al., 2003).

14 It is notable that within the HS chain, only a fraction of the potential substrates are modified and the end product is a polysaccharide of considerable structural heterogeneity.

Interactions between heparan sulfate biosynthetic enzymes Since HS displays high structural diversity, one may ask what are the regulatory mechanisms determining HS structure? Questions about regulation lead to queries about the enzymes: do they communicate with each other in order to produce such a complex molecule? Do they interact with each other? Some studies have shown that indeed some interactions occur between HS biosynthesis enzymes. It has been proposed that the first enzymes in the HS- biosynthetic machinery: XT and Gal-TI form a complex (Silbert and Sugu- maran, 2002). On the other hand, XT and Gal-TII, the next enzyme adding the second galactose, seem not to interact with each other (Pinhal et al., 2001). Similary, glycosyltransferases involved in chain elongation, EXT1 and EXT2 have been shown to form a complex. This complex is believed to be a functional HS polymerase (Kobayashi et al., 2000b; McCormick et al., 2000; Senay et al., 2000). 2OST and the epimerase can also form a complex and their interaction appears to be required for translocation of both the epimerase and 2OST to the Golgi apparatus (Pinhal et al., 2001). In contrast, absence of 2OST does not alter the localization of any of the 6OST isoforms (Nagai et al., 2004). Do the above data suggest that these enzymes work together in a complex, necessary for the production of properly modified HS chains? Al- though it can be tempting to speculate that HS biosynthetic enzymes form a complex, a so-called gagosome, there is no evidence for its existence, except for described interactions between some of the HS-synthesizing enzymes. In contrast, some observations counter this theory. Overexpression of HS biosynthetic enzymes, in many cases, results in altered HS structure. For example, cells overexpressing NDST1 or NDST2 synthesize longer, over-N- sulfated HS chains (Pikas et al., 2000). If the control of HS structure had been restrained by tight interactions between the biosynthetic enzymes, such effects would probably have not been observed. The HS producing enzymes may nevertheless communicate with each other in some way to form properly modified chains. It seems that some regulatory pathways do exist, since HS from different tissues have different structures (Ledin et al., 2004).

Deficiencies in heparan sulfate biosynthetic enzymes HS is involved in many developmental processes (see above). The importance of this molecule during embryogenesis has been shown in experiments with mice deficient in HS synthesizing enzymes.

15 Lack of functional EXT1 or EXT2 abolishes HS production and causes embryonic lethality. EXT1 and EXT2-deficient mice fail to gastrulate and die around embryonic day 8 (E8) (Lin et al., 2000; Stickens et al., 2005). Conditional disruption of EXT1 in the developing mouse brain results in severe brain defects (Inatani et al., 2003). Organisms deficient in the mammalian EXTL proteins are not available yet, but inactivations of the rib-2 gene (the ortholog of human EXTL3) in C. elegans abolishes HS biosynthesis and development in worms stops at gastrulation (Morio et al., 2003). Early developmental defects have also been observed in mice lacking HS- modifying enzymes. Most NDST1 knockout mice die soon after birth, although some die during the embryonic period. They exhibit lung, skull and eye defects and a dramatic reduction in HS N-sulfate content (Fan et al., 2000; Ringvall et al., 2000). NDST1-deficient mice also display developmental brain and facial defects (Grobe et al., 2005). NDST2 knockout mice are viable and fertile but unable to synthesize sulfated heparin (Forsberg et al., 1999; Humphries et al., 1999). Embryos deficient in both NDST1 and NDST2 die between E3.5 and E6.5 (Holmborn et al., 2004). Epimerase knock-out mice do not produce any IdoA and die neonataly. The main causes of death are lack of kidneys, defects in lung and skeleton development, but many vital organs, like brain, heart, liver develop normally (Li et al., 2003). 2OST-deficient mice, similar to the epimerase knock-out, die at birth, showing renal agenesis, abnormalities in the skeleton and eye, but have no lung defects (Bullock et al., 1998). Mice deficient in 6OST have not been reported yet, but a morpholino-mediated functional knockdown of 6- OST in zebrafish results in disturbed muscle development (Bink et al., 2003) and disturbed angiogenesis (Chen et al., 2005). Surprisingly, deficiency in 3OST1 (the enzyme responsible for the production of the majority of the HS sequence necessary for its anticoagulant properties) does not result in an obvious procoagulant phenotype and the mice are both viable and fertile (Shworak et al., 2002).

EXT FAMILY

The human EXT family of proteins (EXTs) comprises five glycosyltransferases: EXT1, EXT2, EXTL1, EXTL2, and EXTL3. EXTs are implicated in HS chain elongation.

Discovery of the EXT family of proteins It has been suggested for many years that the two glycosyltransferase activities required for HS elongation, GlcA-TII and GlcNAc-TII, are located in a single protein (Lidholt et al., 1992). Subsequently, a protein that catalyzed

16 the in vitro transfer of both GlcA and GlcNAc to the appropriate substrates was purified from bovine serum (Lind et al., 1993). Cloning of the corresponding cDNA identified the GlcA- and GlcNAc- transferase as EXT2 (Lind et al., 1998). Using an expression cloning strategy, EXT1 has been found to be involved in HS biosynthesis by restoring HS-dependent herpes simplex infectivity to an HS deficient cell line (McCormick et al., 1998). The indispensable function of these two proteins in HS biosynthesis has been shown by mice knockout studies. EXT1 and EXT2-deficient embryos exhibit the same developmental abnormalities; they fail to gastrulate and HS biosynthesis is abolished (Lin et al., 2000; Stickens et al., 2005). Further- more, the two EXT proteins form a complex with enhanced glycosyltransferase activity (McCormick et al., 2000; Senay et al., 2000). Association of EXT1 and EXT2 has been suggested to be essential for their translocation from ER to the Golgi apparatus where HS biosynthesis takes place (McCormick et al., 2000). In another study, however, it has been shown that untagged, overexpressed EXT1 or EXT2 are localized both to the ER and the Golgi apparatus, and co-expression of EXT1 and EXT2 only reduced the ER residency (Kobayashi et al., 2000b). Three other members of the EXT family: EXTL1, EXTL2 and EXTL3, have been cloned (Saito et al., 1998; Van Hul et al., 1998; Wise et al., 1997; Wuyts et al., 1997), and identified as glycosyltransferases harboring GlcNAc transferase activities. EXTL proteins, therefore, have also been suggested to be involved in HS biosynthesis (Kim et al., 2001; Kitagawa et al., 1999; Pedersen et al., 2003). EXT1 and EXT2 are ubiquitously and concomitantly expressed in mice, both during development and in the adult organisms. In contrast, EXTL1 expression appears to be restricted to the brain and skeletal muscle. EXTL2 and EXTL3 are also ubiquitously expressed (Kitagawa et al., 1999; Saito et al., 1998; Wise et al., 1997).

Glycosyltransferase activities of the EXT family of proteins In common with many other glycosyltransferases, EXTs, use UDP-sugars as donor substrates. They can catalyze the transfer of GlcA and/or GlcNAc (Figure 5) on the growing heparan chain by invertion and retention mechanisms, respectively.

GlcNAc-TI Biochemical studies indicate that the addition of the first GlcNAc unit to the GlcA in the polysaccharide protein-linkage region (GlcA-Gal-Gal-Xyl-O- Ser) can be catalyzed by EXTL2 and EXTL3 (Kim et al., 2001; Kitagawa et al., 1999). It has also been postulated that the human EXT1/2 complex is able to perform this reaction, since recombinant EXT1/2 has been shown to polymerize HS backbone onto substrates in vitro, mimicking the protein-

17 linkage region that lacks the first GlcNAc (Izumikawa et al., 2005; Kim et al., 2003).

GlcA –TII and GlcNAc-TII After the addition of a single GlcNAc residue, elongation proceeds by the alternating addition of GlcA and GlcNAc to the non-reducing end of the growing chain. The incorporation of GlcA residues can be catalyzed by EXT1, EXT2 and the EXT1/2 complex. The complex has been found to possess the highest GlcA-TII activity (McCormick et al., 2000; Senay et al., 2000). As with the GlcA-TII reaction, GlcNAc-TII transfer can be catalyzed by EXT1, EXT2 and EXT1/2. Further, EXTL1 and EXTL3 have also been shown to posses GlcNAc-TII activity (Kim et al., 2001; Lind et al., 1998; McCormick et al., 2000; Senay et al., 2000).

Figure 5. Glycosyltransferase activities of human EXT proteins.

All these reactions, catalyzed in vitro by recombinant proteins, are necessary for the initiation and elongation of the HS saccharide backbone, but the respective functions of the EXT family members in vivo are still elusive. Fur- thermore, the roles of EXTL1 and EXTL2 in HS polymerization have not been conclusively demonstrated.

18 Structure of the EXT family of proteins Human EXTs are type II transmembrane proteins, with a short N-terminal cytoplasmic tail, a single transmembrane-spanning domain, a stem region and a large catalytic domain (Figure 1).

Table 1. Properties of human EXT proteins

Number Molecular Isoelectric Chromsome Tissue Protein of aa weight point (pI) localization specificity residues (kD) EXT1 746 86 9.2 chr.8 q24.11 ubiquitous

EXT2 718 82 6.1 chr.11 p11.2 ubiquitous brain, EXTL1 676 75 8.5 chr.1 p36.11 muscle EXTL2 330 38 9.1 chr.1 p21.2 ubiquitous

EXTL3 919 105 6.1 chr.8 p21.1 ubiquitous

The molecular masses of the EXT proteins are from 37 kD to 105 kD (Table 1). The pI values of the different proteins lie in a broad range between 6 and 9. EXTs are putatively glycosylated since several N-glycosylation sites occur in their amino acid sequences. The multiple sequence alignment of human EXTs showed that the proteins are most similar in their C-terminal part (Figure 6). Several conserved cysteine residues are found in all EXTs, suggesting that they share a common structural fold. One or more DXD motifs, which provide the key elements for binding of the UDP-sugar donors through the coordination of a divalent cation, are present in all EXT proteins. In EXT1 the GlcA-T domain is situated in the N-terminal part of protein (after the transmembrane domain) (Wei et al., 2000). The amino acid sequence of EXTL2, the shortest member of the family, catalyzing the addition of GlcNAc onto the linkage region, aligns with C-terminal parts of the other EXT proteins. The C-terminal part, therefore, probably contains the GlcNAc-T domain. EXTL3 is the largest member of the EXT family and contains an N-terminal fragment with no homology with the other family members. This fragment could be responsible for binding Reg (regenerating gene) protein, as rat EXTL3 has been identified as the Reg receptor that mediates the signal for pancreatic E-cell regeneration (Kobayashi et al., 2000a). The N-terminal part of EXTL3 has also been shown to be involved in TNF-D-induced NF-NB activity (Mizuno et al., 2001). So far, only the crystal structure of EXTL-2 has been solved, in the presence of the donor substrates UDP-GlcNAc and UDP-GalNAc and the acceptor substrate GlcA-

19 Gal-O-naphthalenemethanol (Pedersen et al., 2003). In fact, for EXTL2, UDP-GalNAc appears to be a better donor molecule than UPD-GlcNAc (Kitagawa et al., 1999; Pedersen et al., 2003), which may suggest a function for this enzyme other than the initiation of HS biosynthesis.

Figure 6. Alignment of the amino acid sequences of the human EXT family members. Conserved amino acids within the C-terminal part are shaded. Alignment was prepared using ClustalW (http://www.ebi.ac.uk/clustalw/) and ESPripr 2.1 (http://prodes.toulouse.inra.fr/ESPript/cgi-bin/ESPript.cgi).

20 Human EXT orthologs The human EXT family of proteins has highly homologous counterparts in other mammals and vertebrates, as well as in lower invertebrates (Table 2). EXT orthologs are present in simple invertebrates, such as C. elegans. This is not surprising, since this organism produces HS (Toyoda et al., 2000; Yamada et al., 1999). Interestingly, only the EXT1 and EXTL3 orthologs, rib-1 and rib-2, respectively, have been found in the worm, suggesting that at least in this animal they are sufficient for the production of HS chains. No transferase activity has been detected for rib-1, but it is important to note that rib-1 is about half as long as human EXT1 and corresponds to the N- terminal part of EXT1, which potentially comprises the GlcA-T domain. Rib-1 lacks a signal sequence, a transmembrane domain, and has only one of the conserved DXD-motifs. Rib-2, the EXTL3 ortholog, catalyzes the transfer of GlcNAc in vitro onto acceptors mimicking initiation and polymerization reactions (Kitagawa et al., 2001). Worms deficient in rib-2 lack HS and exhibit impaired development (Morio et al., 2003). In Drosophila, tout velu (TTV, ortholog of EXT1), sister of TTV (SOTV, the EXT2 counterpart), and brother of TTV (BOTV, the counterpart of EXTL3) have all been shown to participate in HS biosynthesis (Bellaiche et al., 1998; Bornemann et al., 2004; Han et al., 2004; Kim et al., 2002; Takei et al., 2004; The et al., 1999). BOTV is an initiator of chain elongation and a complex of TTV and SOTV, like the mammalian EXT1 and EXT2, has been suggested to be the Drosophila HS polymerase. In contrast to the mammalian counterparts, Drosophila EXT1/2 complex has no ascribed GlcNAc-TI activity (Izumikawa et al., 2005). dak and box genes encode the zebrafish EXT2 and EXTL3. Both proteins are required for HS synthesis and therefore necessary for axon sorting in the optic tract and for FGF10 signaling during limb development (Lee et al., 2004; Norton et al., 2005). Interestingly, in zebrafish, three EXT1 genes have been cloned (ext1a-c). This unusually high number is probably due to the genome duplications that occurred in bony fish. The three different EXT1 mRNAs are differentially expressed during development, which may suggest functional divergence (Siekmann and Brand, 2005). Many members of the EXT family are cloned and the primary structure of other transcripts can be predicted from genomic sequences. Phylogenetic analysis of some of the known and predicted sequences has been performed to study the relationships between EXT family members from different eu- karyotic organisms (Busse and Wicher, manuscript in preparation). As shown in Figure 7, the different EXTs cluster together with respective counterparts among species rather than with other family members. EXTL2, which probably first appeared in vertebrates, is the most distinct EXT. This finding supports the hypothesis that EXTL2 may not play a role in HS biosynthesis. EXTL1 is found in all investigated mammals (including marsupi- als) but not in lower vertebrates. EXTL1 is the most similar to mammalian EXT1, suggesting that it arose from the duplication of the EXT1 gene.

21 EXT2 is absent in C. elegans, but is present in Drosophila and higher animals. EXT1, EXTL1 and EXT2 form a cluster suggesting a common origin. C.elegans rib-2 is the most distinct EXTL3, indicating that this protein has different functions in HS biosynthesis in comparison with EXTL3 in other species.

Hereditary multiple exostoses and heparan sulfate Hereditary multiple exostoses (HME) is a human autosomal disorder charac- terized by skeletal malformations that manifests as bony, benign tumors near the ends of long bones, and which in a few cases transforms into chon- drosarcomas or osteosarcomas (Hennekam, 1991; Pierz et al., 2002). Link- age analysis initially identified three different loci for HME; EXT1 on chromosome 8 (8q24.1), EXT2 on chromosome 11 (11p11-13) and EXT3 on chromosome 19 (19p) (Ahn et al., 1995; Le Merrer et al., 1994; Stickens et al., 1996; Wuyts et al., 1996). The third locus, EXT3, is controversial and has not been verified. As EXT1 and EXT2 are glycosyltransferases involved in HS biosynthesis (Lind et al., 1998; McCormick et al., 1998), it may imply that impaired HS biosynthesis is part of the etiology of HME. This idea is supported by the finding of HS abnormalities in exostosis growth plates - the chondrocyte zones of exostosis growth plates have no detectable HS and a diminished and abnormal distribution of perlecan (Hecht et al., 2002). Fur- thermore, exostosis chondrocytes display diminished levels of EXT1 and EXT2 (Bernard et al., 2001). Many frameshift mutations in both EXT1 and EXT2, resulting in truncated forms of the proteins as well as missense mutations in EXT1 have been linked to HME (Wuyts and Van Hul, 2000). The effect of some of the EXT1 mutations on HS synthesis was analyzed using herpes simplex virus infectivity as an indicator of HS expression on the cell surface of an EXT1- deficient cell line. Interestingly, expression of some mutated forms of EXT1 did rescue the EXT1-dependent HS polymerase activity of the mutant cell line, possibly indicating some other undefined function of EXT1 in the development the of HME (Cheung et al., 2001). Using a yeast two hybrid system, the interactions of the C-terminal fragment of EXT2 with tumor necro- sis factor type 1-associated protein and an isoform of human UDP- GalNAc:polypeptide N-acetylgalactososaminyltransferase were shown to be disturbed by HME-related mutations (Simmons et al., 1999). Development of HME can be attributed to improper Indian hedgehog (Ihh) signaling, as has been proposed from studies on mice producing reduced amounts of EXT1. No exostosis, however, was observed in these mice (Hilton et al., 2005; Koziel et al., 2004). On the other hand, ext2 heterozy- gote mice have normal Ihh signaling but develop exostosis (Stickens et al., 2005).

22 23 Figure 7. Phylogenetic analysis of some EXT proteins. The amino acid sequences were aligned using CLUSTALW. Phylogenetic tree was constructed using program MEGA 3.1 with the maximum parsimony method.

24 Hypermethylation of the ext1 promoter leads to elimination of HS biosynthesis in leukemia and non-melanoma skin cancer. Reintroduction of EXT1 into cancer cells induces tumor-suppressor like features (Ropero et al., 2004). Mutations in EXT1 (related to exostosis) have also been linked with autism associated with mental retardation, and may suggest involvement of EXT1 in the development of mental disorders (Li et al., 2002). This sugges- tion may be supported by studies on conditional knockouts of EXT1 in mice, which have shown that HS is essential for brain development (Inatani et al., 2003).

25 PRESENT INVESTIGATIONS

AIM

The aim of this study was to investigate the roles of human EXT1, EXT2 and EXTL3 in heparan sulfate biosynthesis.

RESULTS

In vitro polymerization of heparan sulfate backbone by the EXT proteins (PaperlI) In this study, the ability of EXT1, EXT2 and EXT1/2 complex to elongate the HS carbohydrate backbone in vitro was investigated. We expressed truncated, secreted forms of EXT1, EXT2, and EXT1 and EXT2 together (EXT1/2) in human embryonic kidney (HEK) 293 cells. The proteins were affinity captured from culture media on agarose beads that were later used as an enzyme source in the assays. In order to analyze the catalytic properties of the EXT proteins, two main assays were employed the transferase assay, which quantifies the transfer of single sugars to acceptor molecules (Fig 8) and the polymerization assay, in which the alternating addition of GlcA and GlcNAc to the size-defined acceptor was studied (Fig 9). To study if EXT1, EXT2 or EXT1/2 could polymerize HS chains in vitro, affinity isolated proteins were incubated with small amounts of radiolabeled, size-defined K5 polysaccharide acceptors (GlcA-[GlcNAc-GlcA]4- 3 3 [ H]aManR or [GlcNAc-GlcA]5-[ H]aManR) and a large excess of UDP- sugars (in molar ratio 1:1000). The reaction products were analyzed by gel chromatography.

26 Figure 8. GlcA-TII reaction. Affinity captured EXT proteins were incubated with radiolabeled UDP-GlcA and GlcNAc [GlcA-GlcNAc]n oligosaccharide acceptors. Next, radiolabeled oligosaccharides were separated from unincor- porated UDP-sugars by gel chromatography and quantified by scintillation- counting. For similar analysis of GlcNAc-TII activity, radiolabeled UDP- GlcNAc and [GlcA-GlcNAc]n acceptors were used.

Figure 9. Polymerization reaction. Affinity captured EXT proteins were incubated with radiolabeled, size-defined K5 oligosaccharide acceptors and a mixture of unlabeled UDP-sugars. Labeled products were analyzed by gel chromatography.

27 Alternating addition of GlcNAc and GlcA was observed both for EXT1 alone and for the EXT1/2 complex. However, EXT1 formed generally longer chains than EXT1/2 and also the mode of sugar incorporation in both cases was different. EXT1 appeared to incorporate GlcA and GlcNAc with the same efficiency whereas the EXT1/2 complex seemed to incorporate GlcA more efficiently than GlcNAc. Results from single sugar transfer assays indicated lower GlcNAc-transferase activity for the EXT1/2 complex than for EXT1 alone. EXT2 was able to catalyze the addition of a single GlcA to the [GlcNAc- 3 GlcA]5-[ H]aManR acceptor, but no significant chain elongation was observed.

Embryonic fibroblast with a gene trap mutation in Ext1 produce short heparan sulfate chains (Paper II) In this study HS synthesized by cultured fibroblasts from mice with a gene trap mutation in Ext1 (Ext1Gt/Gt) was examined. In contrast to the Ext1-deficient mice generated by gene targeting that die at day E8.5, the Ext1Gt/Gt mutant mice survive to E14.5 and have a less severe phenotype. In contrast to the complete loss of HS in embryonic stem cells from the Ext1-deficient mice generated by the gene targeting, the Ext1Gt/Gt fibroblasts still produced HS chains. These chains, however, were signifi- cantly shorter then those synthesized by wild–type fibroblasts (20 and 70 kDa, respectively). The presence of HS in mutant cells could be explained by the fact that the mutant cells carry a hypomorphic mutation in Ext1 and therefore small amounts of normal Ext1 transcript are produced in these cells (Koziel 2004). Consistently low but detectable levels of both GlcNAc and GlcA transferase activities were observed in the Ext1Gt/Gt cell extracts. Disaccharide analysis revealed that the sulfation pattern of HS produced by mutant cells was apparently normal. Several signaling molecules, including members of the FGF, Wing- less/Wnt, and Hh families, require HSPG at the cell surface for optimal signaling activity. Whereas biochemical studies using cell culture systems and in vitro binding assays have provided data concerning the minimal growth factor binding epitope on HS, the influence of HS chain length for in vivo growth factor binding in tissues is not known. Our data suggest that not only the sulfation pattern but also the length of the HS chains is a critical determinant of normal mouse development.

EXT1, EXT2 and EXTL3 siRNA gene silencing result in altered heparan sulfate chain length (Paper III) In order to clarify the respective functions of EXT1, EXT2 and EXTL3 in HS biosynthesis, the effect of reduced or increased levels of human EXT1,

28 EXT2 and EXTL3 proteins on HS chain elongation was analyzed. Small interfering RNAs (siRNAs) directed against the human EXT1, EXT2 or EXTL3 mRNAs were introduced into HEK293 cells in order to suppress the expression of respective protein and stable clones overexpressing soluble EXT proteins were developed to increase the protein amounts. Cells transfected with EXT1 or EXT2 siRNAs synthesized shorter HS chains as compared with cells transfected with control siRNA. In contrast, cells transfected with EXTL3 siRNA produced longer HS chains. Overex- pression of soluble EXT1, alone or co-expressed with EXT2, resulted in increased chain length, whereas overexpression of soluble EXT2 or EXTL3 has no detectable effect on HS chain elongation. Additionally, no complex formation between soluble EXTL3 and the EXT1/2 complex was observed by gel filtration analysis.

Heparan sulfate from fibroblasts carrying a hypomorphic mutation in Ext1 exhibit altered interaction with FGF2 (Paper IV) In a previous study (Paper II), HS structure in cultured fibroblasts from mice with a gene trap mutation in Ext1 was examined. The Ext1Gt/Gt mutant mice survive to E14.5 and had elevated Ihh signaling during embryonic chondrocyte differentiation (Koziel et al., 2004). It has been demonstrated that HS chain length is substantially reduced in fibroblasts from these mice (Paper II). In the present study those cells were stained with the 10E4 antibodies and a difference in staining patterns between Ext1Gt/Gt and wt cells was observed. The N-sulfation patterns after nitrous acid deamination at pH 1.5 (cleavage at N-sulfated units within HS) revealed a considerably higher proportion of unmodified saccharide stretches in HS from the mutant cells. Mutant cells responded less efficiently than wt cells to the low concentrations of FGF2, as analyzed by ERK phosphorylation assay. Similarly, HS chains isolated from the mutant cells bound FGF2 more weakly than HS from the wt cells. At higher concentrations of FGF2, however, those differ- ences were reduced. In summary, we conclude that the reduced ability of Ext1Gt/Gt fibroblasts to respond to FGF2 was due to less efficient binding of the mutant HS chains to this growth factor.

DISCUSSION

29 By mediating protein-protein interactions, heparan sulfate plays a pivotal role in many processes during development and in the adult organism. The signaling pathways of many growth factors (such as FGF) are dependent on the presence of HS. It is, therefore, essential to understand HS biosynthesis in order to comprehend its function. HS biosynthesis starts by the formation of the polysaccharide protein linkage region, followed by addition of the first GlcNAc. The chain is then polymerized by the alternate addition of GlcA and GlcNAc sugar units to the nonreducing end of the growing chain. Although both GlcA-TII and GlcNAc-TII transferase activities are ascribed to EXT1, EXT2 and EXT1/2 complex, the capacity to polymerize the chain has not previously been demonstrated for any of them.

In vitro polymerization of the heparan backbone In Paper I, by using the recombinant truncated EXT proteins, it was demonstrated, for the first time, that EXT1 and EXT1/2 complex were capable of catalyzing the polymerization of oligosaccharide chains on K5 acceptors. No significant polymerase activity was observed for EXT2. In vitro polymerization by EXT1 and the EXT1/2 complex did not require membranes, core protein, or other accessory proteins. In contrast to our results, Kim and coworkers (Kim et al., 2003) have not observed any significant polymerization for EXT1, EXT2, or the EXT1/2 complex using the same oligosaccharide acceptor. This could be because they have used a different concentration ratio of acceptor to UDP-sugars, where there was no excess of UDP-sugars over the K5 acceptor. In our study, we used a 1000-fold excess of UDP- sugars over the K5 acceptor. However, Kim et al. have observed extensive polymerization for the EXT1/2 complex, but not for EXT1 and EXT2 alone, using synthetic linkage region analogs or core protein acceptors. They have thus suggested that the HS polymerization reaction requires both a core protein and the EXT1/2 complex formation.

What is the heparan sulfate polymerase in vivo? We found that polymerization products generated by EXT1 ended with either GlcNAc or GlcA, whereas those produced by EXT1/2 ended to a large extent with GlcA (Paper I). In agreement with these results, the EXT1/2 complex exhibited less GlcNAc-TII activity than EXT1 alone. This result was unexpected as previously the EXT1/2 complex has been shown to have a higher GlcNAc-TII activity than EXT1 and EXT2 alone (Senay et al., 2000). The non-reducing end of the HS chain has been suggested to have the following structure: GlcA-GlcNS6S-UA±2S-GlcNS±6-Ido2S-GlcNS±6S (Wu and Lech, 2005), which, together with our polymerization results (Paper I) supports the concept that EXT1/2 complex is the functional polymerase in

30 vivo. On the other hand, EXT1 alone was a more efficient polymerase in our in vitro system.

Mechanism of polymerization The fact that the polymerization on K5 oligosaccharides resulted in high amounts of short products and not in low amounts of long products (Paper I), suggests that in vitro EXT1 and EXT1/2 operate as non-processive glycosyltransferases, with repeated association - sugar(s) transfer - dissociation cycles. Consistently, Ext1Gt/Gt fibroblasts, which express small amounts of the Ext1 transcript, synthesized an apparently normal number of HS chains that were shorter than those produced by wt-cells, instead of producing fewer chains of normal length (Paper II).

Role of EXT2 in heparan sulfate biosynthesis Compared to EXT1, EXT2 is not an efficient glycosyltransferase or polymerase (Paper I). What then, is than the function of this protein? We showed that EXT2 modulates the enzymatic activity of EXT1 in vitro (Paper I), and seems to assist EXT1 in the polymerization process in vivo, since silencing of this protein has the same effect on the HS chain length as silencing of EXT1 (Paper III). Over-expression of EXT2, in contrast to EXT1, has no effect on HS chain length, suggesting that these two proteins do have different functions in the polymerization machinery. It has been proposed that EXT2 may be a chaperone of EXT1 (Wei et al., 2000).

Role of EXTL3 in heparan sulfate biosynthesis We observed that silencing of EXTL3 in HEK293 cells resulted in increased HS chain length compared to control cells and that overexpression of this protein had no effect on HS chain elongation (Paper III). These results are in agreement with the earlier observations that EXTL3, but not EXT1 and EXT2, possesses GlcNAcT-I activity (the ability to add the first GlcNAc residue onto the linkage region) (Kim et al., 2001; Kitagawa et al., 2001). Inhibition of EXTL3, therefore, could have reduced the number of linkage region acceptors able to serve as a substrate for the HS polymerase, eventu- ally leading to the increased length of HS chains synthesized on the remain- ing acceptors. On the other hand, Kim and coworkers (Kim et al., 2003) have observed extensive polymerization by the affinity-captured EXT1/2 complex using linkage region analog acceptors containing a GlcA residue at the non- reducing end. The authors have thus suggested that the human EXT1/2 complex can catalyze the addition of the first GlcNAc residue to the linkage region. In contrast, in vitro polymerization of HS by Drosophila enzymes, onto the same acceptor, required both EXT1/2 (TTV/SOTV) and EXTL3

31 (BOTV), indicating distinct mechanisms of HS biosynthesis in Drosophila and in man (Izumikawa et al., 2005).

Polymerization and N-sulfation It has been shown that the polymerization reaction is stimulated by N- sulfation. However, the precise coupling between HS chain elongation and its modifications has not been defined (Lidholt et al., 1989; Pikas et al., 2000). On the other hand, HS chains produced by mouse embryonic stem cells deficient in both NDST1 and NDST2 were longer then chains from control cells (Holmborn et al., 2004). The addition of recombinant NDST1 and PAPS (sulfate group donor) to the in vitro polymerization assay only slightly increased the chain elongation (Busse et al, unpublished data). Nev- ertheless, it may be that the acceptor used, a K5 decamer was too short to be a good substrate for NDST1, and therefore it is difficult to draw any conclu- sions about the role of N-sulfation for in vitro polymerization. Overexpres- sion of NDST1 or NDST2 together with EXT1, EXT2 or EXT1/2 did not affect HS chain length compared to overexpression of EXTs alone (Presto et al, manuscript in preparation).

Importance of heparan sulfate chain length for HS-protein interactions How are the studies on HS elongating enzymes related to studies on HS- protein interactions? We showed that one of the growth factors known to require HS for proper activation of its downstream signaling cascade, FGF2, activated ERK in Ext1Gt/Gt mutant cells less efficiently than in wt cells (paper IV). HS from mutant cells also bound more weakly to FGF2. Koziel et al (Koziel et al., 2004) demonstrated that Ext1Gt/Gt mice have impaired Hh signaling. The disaccharide composition of Ext1Gt/Gt HS is virtually indistin- guishable from that of wt HS (paper II). However, analysis of the amounts and distribution of N-sulfated glucosamine units, showed that the mutant HS contained higher proportions of unmodified regions than HS from wt cells (paper IV). This can be explained by higher proportions of fragments derived from the unmodified saccharide sequence adjacent to the HS-protein linkage region in shorter chains (Lyon et al., 1994). In summary, the length of HS is apparently one of the critical factors influencing HS - protein interactions.

FUTURE PERSPECTIVES

32 HS biosynthesis is a research field with great opportunities for new discover- ies. Despite the efforts of many investigators, the machinery producing such an important molecule is still poorly understood. The process of HS polymerization is still elusive. Some clues about a role of EXT1, EXT2, EXT1/2 and EXTL3 in this process were gained in the present study. However, the identity of the in vivo polymerase remains to be defined: is it EXT1, the EXT1/2 complex or are there more elements involved in this process? The function of EXT2 is very intriguing because of the necessity of this protein for HS polymerization despite its lack of significant transferase activities. In order to verify whether EXTL3 or the EXT1/2 complex initiates elongation of the HS chain, the role of EXTL3 in should be investigated in more detail. Is the chain length under some control? What are the control mechanisms behind the specific modifications of the HS chains? Do the enzymes participating in HS biosynthesis collaborate? Hopefully the ongoing studies on EXTs-NDSTs interactions will provide us with new insights into the relationships between polymerization and N-sulfation.

In science, every discovery opens new possibilities but also raises new questions. Hopefully, HS biosynthesis and function will one day no longer be a mystery. This would not only satisfy our scientific curiosity, but more importantly, may help us to find cures for HS-related diseases.

33 ACKNOWLEDGEMENTS

The work presented in this thesis was performed at the Department of Medical Bio- chemistry and Microbiology, University of Uppsala. I would like to express my sincere gratitude to everyone who has helped me to complete this thesis, with particular thanks to: Marion Kusche-Gullberg, my supervisor, for being caring, for always having time for many exciting scientific and private discussions, for many good comments on this thesis. Thank you for “sweet” introduction into the heparan sulfate field. Special thanks for spoiling me with “fancy toys”. Lena Kjellen, my co-supervisor, for understanding, for having the same ideas as me and for having more ideas than me. Thank you for helpful comments on my work. Ulf Lindahl, for many good advices and support. Thank you for GLIBS and everything else! Staffan Johansson, my examiner, for patience and support whenever I needed. My supervisors before PhD study: I have used a lot what I had learnt from you! Big thanks to Jan Potempa, Peter Andreasen, Magnus Abrahamson, Bernd Gerhartz, Hanna Rokita, Irena Dorda and Stanisáaw ZaáĊski. My co-authors, especially Kazuyuki Sugahara, for stimulating discussions, and Shuhei Yamada, for sharing the passion for both scientific and non-scientific dis- coveries! Dorothe Spilmann, Jin-Ping Li and Robert Kisilevski for good advices. Erik Fries, for valuable discussions about life in the lab, course-lab and life in gen- eral. My opponent, Tarja Kinnunen, and members of the committee, for their hard work to come through this thesis. All nice people in our and surrounding corridors, for creating pleasant atmosphere during work, especially to: Aneta, for saving my best “double clone” and “burning-drink”; Anh-Tri, for great collaboration in the course lab; Anders, for outstanding assistance during teaching, Anna; Camilla, for teaching me how to teach; Carolina; Donald, for helpful comments; Elina, for being helpful not only in the cell-room; Emanuel; Eva H, my dear seglarkamrat, for sharing the passion for sailing, skiing, hiking, gardening – to men- tion just a few; Eva G, for laughs; Evi; Fang; Feng; Gunilla; Göte; Helena – your K5 is the best!; Jenny, for nice collaboration; Johan, for talks about Szymborska; Josef, for pleasant collaboration in the course lab; Juan; Kaspars; Katarina, for fun and for sharing overheads during teaching; Kristina; Lars, for interesting eve-

34 ning conversations; Lena; Maarten; Maria J.; Maria W., for lovely discussions and African night; Martha; Mehdi; Nadja, for nice talks; Ola, Per, for good comments during seminars; Rebecka, for being, as me, a dog-lover; Sabrina, my charming roommate, for sharing not only the office; Sindhu, for understanding; Sophia, for supportive discussions; Svetlana, for dreaming together; Zhao; Åsa. Special thanks to Elisabet, for sharing happiness and despairs, for support and friendship. Anders Eklund, Håkan Abrahamsson, Pia Ek, Ted Persson, with you, teaching was a pleasure. Thanks to all students, too. Many thanks to Barbro, for great organization of GLIBS meetings; Erika for care for PhD students; Ewa; Inga-Lisa, Kerstin and Olav, for help with computer; Liisa;, Marianne; Maud, Piotr, for good advices and exciting stories; Rehné; Ulla-Maria; Tony, for the best “flower-shelf” and for supplying with necessary reagent. My friends from aikido section of UBK, for a great time together! All my friends spread around the world, especially Ola, for great, long friendship and Jasiek, for always standing by my side, even when I was wrong. “Parrots-on line” friends, for having a lot of fun! To whom this thesis is dedicated: my wonderful family! My beloved U, for support, dedication, encouragement and for giving me the oppor- tunity to make my dream to become scientist comes true! You are the best Mum in the world! My dearest Dad, thank you for teaching me the most appreciate value of being hon- est, great time spent together in the forest and night discussions about everything. My magnificent Sister, for being my best-friend! Baisa, your love, support and understanding are helping me to be a better person. Adored Borsuk, for influencing my life in the best way, being star and example for me, and having the same sense of humor! My Grandparents, for care, interest and support for my work. ElĪbieta and Janusz Wicher, for exceptional hospitality and nice time together. Grzegorz, for help in many ways. Tania, for long discussions during sailing. All my relatives, for supporting me. And, most of all, my love KrzyĞ. Without you I would not be able to finish this thesis. Thank you for immense help, support, tender, and believing in me more then I do. There are no words to express my happiness to have you by my side!

35 REFERENCES

Ahn, J., Ludecke, H. J., Lindow, S., Horton, W. A., Lee, B., Wagner, M. J., Horsthemke, B., and Wells, D. E. (1995). Cloning of the putative tumour suppressor gene for hereditary multiple exostoses (EXT1), Nat Genet 11, 137-43.

Ai, X., Do, A. T., Lozynska, O., Kusche-Gullberg, M., Lindahl, U., and Em- erson, C. P., Jr. (2003). QSulf1 remodels the 6-O sulfation states of cell surface heparan sulfate proteoglycans to promote Wnt signaling, J Cell Biol 162, 341-51.

Aikawa, J., and Esko, J. D. (1999). Molecular cloning and expression of a third member of the heparan sulfate/heparin GlcNAc N-deacetylase/ N- sulfotransferase family, J Biol Chem 274, 2690-5.

Aikawa, J., Grobe, K., Tsujimoto, M., and Esko, J. D. (2001). Multiple isozymes of heparan sulfate/heparin GlcNAc N-deacetylase/GlcN N- sulfotransferase. Structure and activity of the fourth member, NDST4, J Biol Chem 276, 5876-82.

Almeida, R., Levery, S. B., Mandel, U., Kresse, H., Schwientek, T., Bennett, E. P., and Clausen, H. (1999). Cloning and expression of a proteoglycan UDP-galactose:beta-xylose beta1,4-galactosyltransferase I. A seventh member of the human beta4-galactosyltransferase gene family, J Biol Chem 274, 26165-71.

Arikawa-Hirasawa, E., Watanabe, H., Takami, H., Hassell, J. R., and Ya- mada, Y. (1999). Perlecan is essential for cartilage and cephalic development, Nat Genet 23, 354-8.

Bai, X., Zhou, D., Brown, J. R., Crawford, B. E., Hennet, T., and Esko, J. D. (2001). Biosynthesis of the linkage region of glycosaminoglycans: cloning and activity of galactosyltransferase II, the sixth member of the beta 1,3- galactosyltransferase family (beta 3GalT6), J Biol Chem 276, 48189-95.

36 Bellaiche, Y., The, I., and Perrimon, N. (1998). Tout-velu is a Drosophila homologue of the putative tumour suppressor EXT-1 and is needed for Hh diffusion, Nature 394, 85-8.

Bernard, M. A., Hall, C. E., Hogue, D. A., Cole, W. G., Scott, A., Snuggs, M. B., Clines, G. A., Ludecke, H. J., Lovett, M., Van Winkle, W. B., and Hecht, J. T. (2001). Diminished levels of the putative tumor suppressor proteins EXT1 and EXT2 in exostosis chondrocytes, Cell Motil Cytoskeleton 48, 149-62.

Bernfield, M., Gotte, M., Park, P. W., Reizes, O., Fitzgerald, M. L., Lincecum, J., and Zako, M. (1999). Functions of cell surface heparan sulfate proteoglycans, Annu Rev Biochem 68, 729-77.

Bink, R. J., Habuchi, H., Lele, Z., Dolk, E., Joore, J., Rauch, G. J., Geisler, R., Wilson, S. W., den Hertog, J., Kimata, K., and Zivkovic, D. (2003). Heparan sulfate 6-o-sulfotransferase is essential for muscle development in zebrafish, J Biol Chem 278, 31118-27.

Blair, S. S. (2005). Cell signaling: wingless and glypicans together again, Curr Biol 15, R92-4.

Bornemann, D. J., Duncan, J. E., Staatz, W., Selleck, S., and Warrior, R. (2004). Abrogation of heparan sulfate synthesis in Drosophila disrupts the Wingless, Hedgehog and Decapentaplegic signaling pathways, Development 131, 1927-38.

Breton, C., and Imberty, A. (1999). Structure/function studies of glycosyltransferases, Curr Opin Struct Biol 9, 563-71.

Bullock, S. L., Fletcher, J. M., Beddington, R. S., and Wilson, V. A. (1998). Renal agenesis in mice homozygous for a gene trap mutation in the gene encoding heparan sulfate 2-sulfotransferase, Genes Dev 12, 1894-906.

Chen, E., Stringer, S. E., Rusch, M. A., Selleck, S. B., and Ekker, S. C. (2005). A unique role for 6-O sulfation modification in zebrafish vascular development, Dev Biol 284, 364-76.

Cheung, P. K., McCormick, C., Crawford, B. E., Esko, J. D., Tufaro, F., and Duncan, G. (2001). Etiological point mutations in the hereditary multiple exostoses gene EXT1: a functional analysis of heparan sulfate polymerase activity, Am J Hum Genet 69, 55-66.

Costell, M., Gustafsson, E., Aszodi, A., Morgelin, M., Bloch, W., Hunziker, E., Addicks, K., Timpl, R., and Fassler, R. (1999). Perlecan maintains the

37 integrity of cartilage and some basement membranes, J Cell Biol 147, 1109- 22.

Crawford, B. E., Olson, S. K., Esko, J. D., and Pinhal, M. A. (2001). Clon- ing, Golgi localization, and enzyme activity of the full-length heparin/heparan sulfate-glucuronic acid C5-epimerase, J Biol Chem 276, 21538- 43.

Eriksson, I., Sandback, D., Ek, B., Lindahl, U., and Kjellen, L. (1994). cDNA cloning and sequencing of mouse mastocytoma glucosaminyl N- deacetylase/N-sulfotransferase, an enzyme involved in the biosynthesis of heparin, J Biol Chem 269, 10438-43.

Esko, J. D., and Selleck, S. B. (2002). Order out of chaos: assembly of ligand binding sites in heparan sulfate, Annu Rev Biochem 71, 435-71.

Esko, J. D., and Zhang, L. (1996). Influence of core protein sequence on glycosaminoglycan assembly, Curr Opin Struct Biol 6, 663-70.

Fan, G., Xiao, L., Cheng, L., Wang, X., Sun, B., and Hu, G. (2000). Tar- geted disruption of NDST-1 gene leads to pulmonary hypoplasia and neonatal respiratory distress in mice, FEBS Lett 467, 7-11.

Fernandez-Busquets, X., and Burger, M. M. (2003). Circular proteoglycans from sponges: first members of the spongican family, Cell Mol Life Sci 60, 88-112.

Forsberg, E., Pejler, G., Ringvall, M., Lunderius, C., Tomasini-Johansson, B., Kusche-Gullberg, M., Eriksson, I., Ledin, J., Hellman, L., and Kjellen, L. (1999). Abnormal mast cells in mice deficient in a heparin-synthesizing enzyme, Nature 400, 773-6.

Furmanek, A., and Hofsteenge, J. (2000). Protein C-mannosylation: facts and questions, Acta Biochim Pol 47, 781-9.

Gotting, C., Kuhn, J., Zahn, R., Brinkmann, T., and Kleesiek, K. (2000). Molecular cloning and expression of human UDP-d-Xylose:proteoglycan core protein beta-d-xylosyltransferase and its first isoform XT-II, J Mol Biol 304, 517-28.

Grobe, K., Inatani, M., Pallerla, S. R., Castagnola, J., Yamaguchi, Y., and Esko, J. D. (2005). Cerebral hypoplasia and craniofacial defects in mice lacking heparan sulfate Ndst1 gene function, Development 132, 3777-86.

38 Grzybek, M., Kozubek, A., Dubielecka, P., and Sikorski, A. F. (2005). Rafts--the current picture, Folia Histochem Cytobiol 43, 3-10. Habuchi, H., Habuchi, O., and Kimata, K. (2004). Sulfation pattern in glycosaminoglycan: does it have a code?, Glycoconj J 21, 47-52.

Hacker, U., Nybakken, K., and Perrimon, N. (2005). Heparan sulphate proteoglycans: the sweet side of development, Nat Rev Mol Cell Biol 6, 530-41.

Han, C., Belenkaya, T. Y., Khodoun, M., Tauchi, M., and Lin, X. (2004). Distinct and collaborative roles of Drosophila EXT family proteins in morphogen signalling and gradient formation, Development 131, 1563-75.

Hart, G. W. (1997). Dynamic O-linked glycosylation of nuclear and cytoskeletal proteins, Annu Rev Biochem 66, 315-35.

Hecht, J. T., Hall, C. R., Snuggs, M., Hayes, E., Haynes, R., and Cole, W. G. (2002). Heparan sulfate abnormalities in exostosis growth plates, Bone 31, 199-204.

Hennekam, R. C. (1991). Hereditary multiple exostoses, J Med Genet 28, 262-6.

Henningsson, F., Ledin, J., Lunderius, C., Wilen, M., Hellman, L., and Pej- ler, G. (2002). Altered storage of proteases in mast cells from mice lacking heparin: a possible role for heparin in carboxypeptidase A processing, Biol Chem 383, 793-801.

Hilton, M. J., Gutierrez, L., Martinez, D. A., and Wells, D. E. (2005). EXT1 regulates chondrocyte proliferation and differentiation during endochondral bone development, Bone 36, 379-86.

Holmborn, K., Ledin, J., Smeds, E., Eriksson, I., Kusche-Gullberg, M., and Kjellen, L. (2004). Heparan sulfate synthesized by mouse embryonic stem cells deficient in NDST1 and NDST2 is 6-O-sulfated but contains no N- sulfate groups, J Biol Chem 279, 42355-8.

Humphries, D. E., Wong, G. W., Friend, D. S., Gurish, M. F., Qiu, W. T., Huang, C., Sharpe, A. H., and Stevens, R. L. (1999). Heparin is essential for the storage of specific granule proteases in mast cells, Nature 400, 769-72.

Inatani, M., Irie, F., Plump, A. S., Tessier-Lavigne, M., and Yamaguchi, Y. (2003). Mammalian brain morphogenesis and midline axon guidance require heparan sulfate, Science 302, 1044-6.

39 Iozzo, R. V. (2005). Basement membrane proteoglycans: from cellar to ceil- ing, Nat Rev Mol Cell Biol 6, 646-56.

Izumikawa, T., Egusa, N., Taniguchi, F., Sugahara, K., and Kitagawa, H. (2005). Heparan sulfate polymerization in Drosophila, J Biol Chem.

Jann, K., and Jann, B. (1992). Capsules of Escherichia coli, expression and biological significance, Can J Microbiol 38, 705-10.

Kato, M., Wang, H., Kainulainen, V., Fitzgerald, M. L., Ledbetter, S., Or- nitz, D. M., and Bernfield, M. (1998). Physiological degradation converts the soluble syndecan-1 ectodomain from an inhibitor to a potent activator of FGF-2, Nat Med 4, 691-7.

Kim, B. T., Kitagawa, H., Tamura, J., Saito, T., Kusche-Gullberg, M., Lin- dahl, U., and Sugahara, K. (2001). Human tumor suppressor EXT gene family members EXTL1 and EXTL3 encode alpha 1,4- N- acetylglucosaminyltransferases that likely are involved in heparan sulfate/ heparin biosynthesis, Proc Natl Acad Sci U S A 98, 7176-81.

Kim, B. T., Kitagawa, H., Tamura Ji, J., Kusche-Gullberg, M., Lindahl, U., and Sugahara, K. (2002). Demonstration of a novel gene DEXT3 of Droso- phila melanogaster as the essential N-acetylglucosamine transferase in the heparan sulfate biosynthesis: chain initiation and elongation, J Biol Chem 277, 13659-65.

Kim, B. T., Kitagawa, H., Tanaka, J., Tamura, J., and Sugahara, K. (2003). In vitro heparan sulfate polymerization: crucial roles of core protein moieties of primer substrates in addition to the EXT1-EXT2 interaction, J Biol Chem 278, 41618-23.

Kitagawa, H., Egusa, N., Tamura, J. I., Kusche-Gullberg, M., Lindahl, U., and Sugahara, K. (2001). rib-2, a Caenorhabditis elegans homolog of the human tumor suppressor EXT genes encodes a novel alpha1,4-N- acetylglucosaminyltransferase involved in the biosynthetic initiation and elongation of heparan sulfate, J Biol Chem 276, 4834-8.

Kitagawa, H., Shimakawa, H., and Sugahara, K. (1999). The tumor suppressor EXT-like gene EXTL2 encodes an alpha1, 4-N- acetylhexosaminyltransferase that transfers N-acetylgalactosamine and N- acetylglucosamine to the common glycosaminoglycan-protein linkage region. The key enzyme for the chain initiation of heparan sulfate, J Biol Chem 274, 13933-7.

40 Kitagawa, H., Tone, Y., Tamura, J., Neumann, K. W., Ogawa, T., Oka, S., Kawasaki, T., and Sugahara, K. (1998). Molecular cloning and expression of glucuronyltransferase I involved in the biosynthesis of the glycosaminoglycan-protein linkage region of proteoglycans, J Biol Chem 273, 6615-8.

Kjellen, L. (2003). Glucosaminyl N-deacetylase/N-sulphotransferases in heparan sulphate biosynthesis and biology, Biochem Soc Trans 31, 340-2.

Kjellen, L., Pettersson, I., Unger, E., and Lindahl, U. (1992). Two enzymes in one: N-deacetylation and N-sulfation in heparin biosynthesis are catalyzed by the same protein, Adv Exp Med Biol 313, 107-11.

Kobayashi, M., Habuchi, H., Yoneda, M., Habuchi, O., and Kimata, K. (1997). Molecular cloning and expression of Chinese hamster ovary cell heparan-sulfate 2-sulfotransferase, J Biol Chem 272, 13980-5.

Kobayashi, S., Akiyama, T., Nata, K., Abe, M., Tajima, M., Shervani, N. J., Unno, M., Matsuno, S., Sasaki, H., Takasawa, S., and Okamoto, H. (2000a). Identification of a receptor for reg (regenerating gene) protein, a pancreatic beta-cell regeneration factor, J Biol Chem 275, 10723-6.

Kobayashi, S., Morimoto, K., Shimizu, T., Takahashi, M., Kurosawa, H., and Shirasawa, T. (2000b). Association of EXT1 and EXT2, hereditary multiple exostoses gene products, in Golgi apparatus, Biochem Biophys Res Commun 268, 860-7.

Kolset, S. O., Prydz, K., and Pejler, G. (2004). Intracellular proteoglycans, Biochem J 379, 217-27.

Koziel, L., Kunath, M., Kelly, O. G., and Vortkamp, A. (2004). Ext1- dependent heparan sulfate regulates the range of Ihh signaling during endochondral ossification, Dev Cell 6, 801-13.

Kramer, K. L., and Yost, H. J. (2003). Heparan sulfate core proteins in cell- cell signaling, Annu Rev Genet 37, 461-84.

Kusche-Gullberg, M., Eriksson, I., Pikas, D. S., and Kjellen, L. (1998). Iden- tification and expression in mouse of two heparan sulfate glucosaminyl N- deacetylase/N-sulfotransferase genes, J Biol Chem 273, 11902-7.

Kusche-Gullberg, M., and Kjellen, L. (2003). Sulfotransferases in glycosaminoglycan biosynthesis, Curr Opin Struct Biol 13, 605-11.

Le Merrer, M., Legeai-Mallet, L., Jeannin, P. M., Horsthemke, B., Schinzel, A., Plauchu, H., Toutain, A., Achard, F., Munnich, A., and Maroteaux, P.

41 (1994). A gene for hereditary multiple exostoses maps to chromosome 19p, Hum Mol Genet 3, 717-22.

Ledin, J., Staatz, W., Li, J. P., Gotte, M., Selleck, S., Kjellen, L., and Spill- mann, D. (2004). Heparan sulfate structure in mice with genetically modified heparan sulfate production, J Biol Chem 279, 42732-41.

Lee, J. S., von der Hardt, S., Rusch, M. A., Stringer, S. E., Stickney, H. L., Talbot, W. S., Geisler, R., Nusslein-Volhard, C., Selleck, S. B., Chien, C. B., and Roehl, H. (2004). Axon sorting in the optic tract requires HSPG synthesis by ext2 (dackel) and extl3 (boxer), Neuron 44, 947-60.

Li, H., Yamagata, T., Mori, M., and Momoi, M. Y. (2002). Association of autism in two patients with hereditary multiple exostoses caused by novel deletion mutations of EXT1, J Hum Genet 47, 262-5.

Li, J., Hagner-McWhirter, A., Kjellen, L., Palgi, J., Jalkanen, M., and Lin- dahl, U. (1997). Biosynthesis of heparin/heparan sulfate. cDNA cloning and expression of D-glucuronyl C5-epimerase from bovine lung, J Biol Chem 272, 28158-63.

Li, J. P., Gong, F., El Darwish, K., Jalkanen, M., and Lindahl, U. (2001). Characterization of the D-glucuronyl C5-epimerase involved in the biosynthesis of heparin and heparan sulfate, J Biol Chem 276, 20069-77.

Li, J. P., Gong, F., Hagner-McWhirter, A., Forsberg, E., Abrink, M., Kis- ilevsky, R., Zhang, X., and Lindahl, U. (2003). Targeted disruption of a murine glucuronyl C5-epimerase gene results in heparan sulfate lacking L- iduronic acid and in neonatal lethality, J Biol Chem 278, 28363-6.

Lidholt, K., Kjellen, L., and Lindahl, U. (1989). Biosynthesis of heparin. Relationship between the polymerization and sulphation processes, Biochem J 261, 999-1007.

Lidholt, K., Weinke, J. L., Kiser, C. S., Lugemwa, F. N., Bame, K. J., Cheifetz, S., Massague, J., Lindahl, U., and Esko, J. D. (1992). A single mutation affects both N-acetylglucosaminyltransferase and glucuronosyltransferase activities in a Chinese hamster ovary cell mutant defective in heparan sulfate biosynthesis, Proc Natl Acad Sci U S A 89, 2267-71.

Lin, X., Wei, G., Shi, Z., Dryer, L., Esko, J. D., Wells, D. E., and Matzuk, M. M. (2000). Disruption of gastrulation and heparan sulfate biosynthesis in EXT1-deficient mice, Dev Biol 224, 299-311.

42 Lind, T., Lindahl, U., and Lidholt, K. (1993). Biosynthesis of heparin/heparan sulfate. Identification of a 70-kDa protein catalyzing both the D- glucuronosyl- and the N-acetyl-D-glucosaminyltransferase reactions, J Biol Chem 268, 20705-8.

Lind, T., Tufaro, F., McCormick, C., Lindahl, U., and Lidholt, K. (1998). The putative tumor suppressors EXT1 and EXT2 are glycosyltransferases required for the biosynthesis of heparan sulfate, J Biol Chem 273, 26265-8.

Lindahl, U., Kusche-Gullberg, M., and Kjellen, L. (1998). Regulated diversity of heparan sulfate, J Biol Chem 273, 24979-82.

Lindahl, U., Li, J. P., Kusche-Gullberg, M., Salmivirta, M., Alaranta, S., Veromaa, T., Emeis, J., Roberts, I., Taylor, C., Oreste, P., et al. (2005). Gen- eration of "neoheparin" from E. coli K5 capsular polysaccharide, J Med Chem 48, 349-52.

Lomako, J., Lomako, W. M., and Whelan, W. J. (2004). Glycogenin: the primer for mammalian and yeast glycogen synthesis, Biochim Biophys Acta 1673, 45-55.

Lyon, M., Deakin, J. A., and Gallagher, J. T. (1994). Liver heparan sulfate structure. A novel molecular design, J Biol Chem 269, 11208-15.

McCormick, C., Duncan, G., Goutsos, K. T., and Tufaro, F. (2000). The putative tumor suppressors EXT1 and EXT2 form a stable complex that accumulates in the Golgi apparatus and catalyzes the synthesis of heparan sulfate, Proc Natl Acad Sci U S A 97, 668-73.

McCormick, C., Leduc, Y., Martindale, D., Mattison, K., Esford, L. E., Dyer, A. P., and Tufaro, F. (1998). The putative tumour suppressor EXT1 alters the expression of cell-surface heparan sulfate, Nat Genet 19, 158-61.

Medeiros, G. F., Mendes, A., Castro, R. A., Bau, E. C., Nader, H. B., and Dietrich, C. P. (2000). Distribution of sulfated glycosaminoglycans in the animal kingdom: widespread occurrence of heparin-like compounds in invertebrates, Biochim Biophys Acta 1475, 287-94.

Mizuno, K., Irie, S., and Sato, T. A. (2001). Overexpression of EXTL3/EXTR1 enhances NF-kappaB activity induced by TNF-alpha, Cell Signal 13, 125-30.

Morio, H., Honda, Y., Toyoda, H., Nakajima, M., Kurosawa, H., and Shirasawa, T. (2003). EXT gene family member rib-2 is essential for em-

43 bryonic development and heparan sulfate biosynthesis in Caenorhabditis elegans, Biochem Biophys Res Commun 301, 317-23.

Nadanaka, S., Kitagawa, H., and Sugahara, K. (1998). Demonstration of the immature glycosaminoglycan tetrasaccharide sequence GlcAbeta1- 3Galbeta1-3Galbeta1-4Xyl on recombinant soluble human alpha- thrombomodulin. An oligosaccharide structure on a "part-time" proteoglycan, J Biol Chem 273, 33728-34.

Nagai, N., Habuchi, H., Esko, J. D., and Kimata, K. (2004). Stem domains of heparan sulfate 6-O-sulfotransferase are required for Golgi localization, oligomer formation and enzyme activity, J Cell Sci 117, 3331-41.

Norton, W. H., Ledin, J., Grandel, H., and Neumann, C. J. (2005). HSPG synthesis by zebrafish Ext2 and Extl3 is required for Fgf10 signalling during limb development, Development 132, 4963-73.

Okajima, T., Yoshida, K., Kondo, T., and Furukawa, K. (1999). Human homolog of Caenorhabditis elegans sqv-3 gene is galactosyltransferase I involved in the biosynthesis of the glycosaminoglycan-protein linkage region of proteoglycans, J Biol Chem 274, 22915-8.

Paulson, J. C., and Colley, K. J. (1989). Glycosyltransferases. Structure, localization, and control of cell type-specific glycosylation, J Biol Chem 264, 17615-8.

Pedersen, L. C., Dong, J., Taniguchi, F., Kitagawa, H., Krahn, J. M., Peder- sen, L. G., Sugahara, K., and Negishi, M. (2003). Crystal structure of an alpha 1,4-N-acetylhexosaminyltransferase (EXTL2), a member of the exostosin gene family involved in heparan sulfate biosynthesis, J Biol Chem 278, 14420-8.

Pierz, K. A., Stieber, J. R., Kusumi, K., and Dormans, J. P. (2002). Heredi- tary multiple exostoses: one center's experience and review of etiology, Clin Orthop Relat Res, 49-59.

Pikas, D. S., Eriksson, I., and Kjellen, L. (2000). Overexpression of different isoforms of glucosaminyl N-deacetylase/N-sulfotransferase results in distinct heparan sulfate N-sulfation patterns, Biochemistry 39, 4552-8.

Pinhal, M. A., Smith, B., Olson, S., Aikawa, J., Kimata, K., and Esko, J. D. (2001). Enzyme interactions in heparan sulfate biosynthesis: uronosyl 5- epimerase and 2-O-sulfotransferase interact in vivo, Proc Natl Acad Sci U S A 98, 12984-9.

44 Prydz, K., and Dalen, K. T. (2000). Synthesis and sorting of proteoglycans, J Cell Sci 113 Pt 2, 193-205.

Rawson, J. M., Dimitroff, B., Johnson, K. G., Ge, X., Van Vactor, D., and Selleck, S. B. (2005). The heparan sulfate proteoglycans Dally-like and Syndecan have distinct functions in axon guidance and visual-system assembly in Drosophila, Curr Biol 15, 833-8.

Ringvall, M., Ledin, J., Holmborn, K., van Kuppevelt, T., Ellin, F., Eriksson, I., Olofsson, A. M., Kjellen, L., and Forsberg, E. (2000). Defective heparan sulfate biosynthesis and neonatal lethality in mice lacking N-deacetylase/N- sulfotransferase-1, J Biol Chem 275, 25926-30.

Rong, J., Habuchi, H., Kimata, K., Lindahl, U., and Kusche-Gullberg, M. (2000). Expression of heparan sulphate L-iduronyl 2-O-sulphotransferase in human kidney 293 cells results in increased D-glucuronyl 2-O-sulphation, Biochem J 346 Pt 2, 463-8.

Ropero, S., Setien, F., Espada, J., Fraga, M. F., Herranz, M., Asp, J., Benas- si, M. S., Franchi, A., Patino, A., Ward, L. S., et al. (2004). Epigenetic loss of the familial tumor-suppressor gene exostosin-1 (EXT1) disrupts heparan sulfate synthesis in cancer cells, Hum Mol Genet 13, 2753-65.

Saito, T., Seki, N., Yamauchi, M., Tsuji, S., Hayashi, A., Kozuma, S., and Hori, T. (1998). Structure, chromosomal location, and expression profile of EXTR1 and EXTR2, new members of the multiple exostoses gene family, Biochem Biophys Res Commun 243, 61-6.

Sasisekharan, R., Shriver, Z., Venkataraman, G., and Narayanasami, U. (2002). Roles of heparan-sulphate glycosaminoglycans in cancer, Nat Rev Cancer 2, 521-8.

Senay, C., Lind, T., Muguruma, K., Tone, Y., Kitagawa, H., Sugahara, K., Lidholt, K., Lindahl, U., and Kusche-Gullberg, M. (2000). The EXT1/EXT2 tumor suppressors: catalytic activities and role in heparan sulfate biosynthesis, EMBO Rep 1, 282-6.

Shworak, N. W., HajMohammadi, S., de Agostini, A. I., and Rosenberg, R. D. (2002). Mice deficient in heparan sulfate 3-O-sulfotransferase-1: normal hemostasis with unexpected perinatal phenotypes, Glycoconj J 19, 355-61.

Shworak, N. W., Liu, J., Petros, L. M., Zhang, L., Kobayashi, M., Copeland, N. G., Jenkins, N. A., and Rosenberg, R. D. (1999). Multiple isoforms of heparan sulfate D-glucosaminyl 3-O-sulfotransferase. Isolation, characteri-

45 zation, and expression of human cdnas and identification of distinct genomic loci, J Biol Chem 274, 5170-84.

Siekmann, A. F., and Brand, M. (2005). Distinct tissue-specificity of three zebrafish ext1 genes encoding proteoglycan modifying enzymes and their relationship to somitic Sonic hedgehog signaling, Dev Dyn 232, 498-505.

Silbert, J. E., and Sugumaran, G. (2002). Biosynthesis of chondroitin/ dermatan sulfate, IUBMB Life 54, 177-86.

Simmons, A. D., Musy, M. M., Lopes, C. S., Hwang, L. Y., Yang, Y. P., and Lovett, M. (1999). A direct interaction between EXT proteins and glycosyltransferases is defective in hereditary multiple exostoses, Hum Mol Genet 8, 2155-64.

Smeds, E., Habuchi, H., Do, A. T., Hjertson, E., Grundberg, H., Kimata, K., Lindahl, U., and Kusche-Gullberg, M. (2003). Substrate specificities of mouse heparan sulphate glucosaminyl 6-O-sulphotransferases, Biochem J 372, 371-80.

Spillmann, D. (2001). Heparan sulfate: anchor for viral intruders?, Bio- chimie 83, 811-7.

Spiro, R. G. (2002). Protein glycosylation: nature, distribution, enzymatic formation, and disease implications of glycopeptide bonds, Glycobiology 12, 43R-56R.

Stickens, D., Clines, G., Burbee, D., Ramos, P., Thomas, S., Hogue, D., Hecht, J. T., Lovett, M., and Evans, G. A. (1996). The EXT2 multiple exostoses gene defines a family of putative tumour suppressor genes, Nat Genet 14, 25-32.

Stickens, D., Zak, B. M., Rougier, N., Esko, J. D., and Werb, Z. (2005). Mice deficient in Ext2 lack heparan sulfate and develop exostoses, Devel- opment 132, 5055-68.

Sugahara, K., and Kitagawa, H. (2002). Heparin and heparan sulfate biosynthesis, IUBMB Life 54, 163-75.

Takei, Y., Ozawa, Y., Sato, M., Watanabe, A., and Tabata, T. (2004). Three Drosophila EXT genes shape morphogen gradients through synthesis of heparan sulfate proteoglycans, Development 131, 73-82.

46 The, I., Bellaiche, Y., and Perrimon, N. (1999). Hedgehog movement is regulated through tout velu-dependent synthesis of a heparan sulfate proteoglycan, Mol Cell 4, 633-9.

Toyoda, H., Kinoshita-Toyoda, A., and Selleck, S. B. (2000). Structural analysis of glycosaminoglycans in Drosophila and Caenorhabditis elegans and demonstration that tout-velu, a Drosophila gene related to EXT tumor suppressors, affects heparan sulfate in vivo, J Biol Chem 275, 2269-75.

Ueno, M., Yamada, S., Zako, M., Bernfield, M., and Sugahara, K. (2001). Structural characterization of heparan sulfate and chondroitin sulfate of syndecan-1 purified from normal murine mammary gland epithelial cells. Common phosphorylation of xylose and differential sulfation of galactose in the protein linkage region tetrasaccharide sequence, J Biol Chem 276, 29134-40.

Uhlin-Hansen, L., Kusche-Gullberg, M., Berg, E., Eriksson, I., and Kjellen, L. (1997). Mouse mastocytoma cells synthesize undersulfated heparin and chondroitin sulfate in the presence of brefeldin A, J Biol Chem 272, 3200-6. van den Born, J., Gunnarsson, K., Bakker, M. A., Kjellen, L., Kusche- Gullberg, M., Maccarana, M., Berden, J. H., and Lindahl, U. (1995). Pres- ence of N-unsubstituted glucosamine units in native heparan sulfate revealed by a monoclonal antibody, J Biol Chem 270, 31303-9.

Van den Steen, P., Rudd, P. M., Dwek, R. A., and Opdenakker, G. (1998). Concepts and principles of O-linked glycosylation, Crit Rev Biochem Mol Biol 33, 151-208.

Van Hul, W., Wuyts, W., Hendrickx, J., Speleman, F., Wauters, J., De Boulle, K., Van Roy, N., Bossuyt, P., and Willems, P. J. (1998). Identifica- tion of a third EXT-like gene (EXTL3) belonging to the EXT gene family, Genomics 47, 230-7.

Vertel, B. M., Walters, L. M., Flay, N., Kearns, A. E., and Schwartz, N. B. (1993). Xylosylation is an endoplasmic reticulum to Golgi event, J Biol Chem 268, 11105-12.

Wei, G., Bai, X., Gabb, M. M., Bame, K. J., Koshy, T. I., Spear, P. G., and Esko, J. D. (2000). Location of the glucuronosyltransferase domain in the heparan sulfate copolymerase EXT1 by analysis of Chinese hamster ovary cell mutants, J Biol Chem 275, 27733-40.

Wei, Z., Swiedler, S. J., Ishihara, M., Orellana, A., and Hirschberg, C. B. (1993). A single protein catalyzes both N-deacetylation and N-sulfation dur-

47 ing the biosynthesis of heparan sulfate, Proc Natl Acad Sci U S A 90, 3885- 8.

Westling, C., and Lindahl, U. (2002). Location of N-unsubstituted glucosamine residues in heparan sulfate, J Biol Chem 277, 49247-55.

Whitelock, J. M., and Iozzo, R. V. (2005). Heparan sulfate: a complex polymer charged with biological activity, Chem Rev 105, 2745-64.

Wise, C. A., Clines, G. A., Massa, H., Trask, B. J., and Lovett, M. (1997). Identification and localization of the gene for EXTL, a third member of the multiple exostoses gene family, Genome Res 7, 10-6.

Wu, Z. L., and Lech, M. (2005). Characterizing the non-reducing end structure of heparan sulfate, J Biol Chem 280, 33749-55.

Wuyts, W., and Van Hul, W. (2000). Molecular basis of multiple exostoses: mutations in the EXT1 and EXT2 genes, Hum Mutat 15, 220-7.

Wuyts, W., Van Hul, W., Hendrickx, J., Speleman, F., Wauters, J., De Boulle, K., Van Roy, N., Van Agtmael, T., Bossuyt, P., and Willems, P. J. (1997). Identification and characterization of a novel member of the EXT gene family, EXTL2, Eur J Hum Genet 5, 382-9.

Wuyts, W., Van Hul, W., Wauters, J., Nemtsova, M., Reyniers, E., Van Hul, E. V., De Boulle, K., de Vries, B. B., Hendrickx, J., Herrygers, I., et al. (1996). Positional cloning of a gene involved in hereditary multiple exostoses, Hum Mol Genet 5, 1547-57.

Yamada, S., Van Die, I., Van den Eijnden, D. H., Yokota, A., Kitagawa, H., and Sugahara, K. (1999). Demonstration of glycosaminoglycans in Caenor- habditis elegans, FEBS Lett 459, 327-31.

Yoneda, A., and Couchman, J. R. (2003). Regulation of cytoskeletal organization by syndecan transmembrane proteoglycans, Matrix Biol 22, 25-33.

Zak, B. M., Crawford, B. E., and Esko, J. D. (2002). Hereditary multiple exostoses and heparan sulfate polymerization, Biochim Biophys Acta 1573, 346-55.

, *

) ' - ., ),

) , ) / ., ), 0 * 0 // /( 1 2 , , / ) 3 /4 2 ) , ( 0 2 // ) ( ) 5 -5 6 / * // - / ., ), 1 7 8 0 $"0 2 ( ) ) 96 / * // - / ., ), :1;

- ( ' (, 11 '(' '')*#!++