The most common GAG structures are chondroitin sulfate (CS), dermatan sulfate (DS), heparan sulfate (HS) and keratan sulfate (KS), hyaluronic acid (HA) and heparin. With the exception of heparin and HA, GAG chains are covalently attached at their reducing end through an Oglycosidic linkage to a serine residue or N-linked to asparagine in a core protein. The resulting macromolecule is termed a proteoglycan. Proteoglycans are distributed on the surface of adventitial reticular cells as well as within the extracellular matrix. They play a vital role in virology, immunology, biological recognition, biomolecular transport, cell and various other biological processes [1].

The most abundant of the GAGs are CS, which occur in cartilage, tendons, ligaments, and aorta. DS is a connective tissue component found in skin, blood vessels, and heart valves. Heparin is the only GAG with predominantly alpha-glycosidic links and occurs in connective tissue mast cells. HS is structurally similar to heparin and occurs in blood vessel walls and brain.

Although GAGs have a heterogeneous composition, unique structural determinants within the polysaccharide mediate their interactions with specific GAG-binding proteins.

Protein Interactions

1. Protease Inhibitors: Inhibitors of plasma serine proteases inhibitors a superfamily of proteins (termed serpins) that form 1:1 covalent complexes with the active site of their target proteases. Particularly relevant are Antithrombin III (AT) and Heparin Cofactor II (HCII). In addition to inhibiting thrombin, AT also inhibits factor Xa and other proteases of the intrinsic coagulation pathway, including factor IXa, XIa, XIIa and kallikrein. During coagulation, AT is cleaved by thrombin between Arg-393 and Ser-394, causing inactivation of the inhibitor and the target enzyme. In the presence of heparin, the inactivation of thrombin by AT is greatly accelerated, involving changes in both the conformation of heparin and AT. Upon formation of the tight inhibitor-enzyme complex, heparin dissociates and recycles to catalyze the inhibition of the next enzyme molecule with AT.

Much of the polysaccharide structure of heparin can be represented as a repeating, trisulfated disaccharide (A-B or B-C in Fig. 1). Regions containing disulfated disaccharides (C-D in Fig. 1) and disaccharides lacking a sulfate on C-6 of the glucosamine residue are typically present [3]. A specialized region in the GAG determines the anticoagulant activity. A pentasaccharide sequence E-I (Fig. 1), containing the unique trisulfated glucosamine residue (G) has been shown to represent the minimum high-affinity structure for AT [4]

It is commonly accepted that sulfation, particularly N-sulfation is crucial to heparin's ATdependant anticoagulant activity. Studies with oligosaccharides containing the E-I pentasaccharide have demonstrated that the C-6 sulfate on residue E [5], the 3-O-sulfate on residue H [6] and the two N-sulfate residues on G and I [7] are essential. However, the pentasaccharide-AT complex only inhibits free factor Xa as oligosaccharides of < 5,400 Dalton are without heparin cofactor activity for thrombin [8]. Therefore, in addition to a specific polysaccharide sequence, structure and sulfation pattern, heparin's chain length is an important determinant for thrombin inhibition.

Figure 1

In contrast to AT, HCII is specific for thrombin and does not inactivate other coagulation enzymes. Although heparin activates HCII and inhibits thrombin in solution, heparin does efficiently inhibit bound thrombin via HCII (or AT). However, HCII activated by DS mediates the efficient inhibition of bound thrombin indicating that thrombin associated with surfaces exhibits a unique conformation recognized by DS-HCII but not by AT or H/HCII [9].

DS contains fewer sulfate groups than heparin and has uronic acid residues that alternate with galactosamine rather than glucosamine. A third plasma protease inhibitor that is structurally similar to AT and HCII is leuserpin. Limited information is available as to the physiological role of this inhibitor.

2. Plasma Lipoproteins: In the presence of divalent cations, heparin precipitates serum and low density lipoproteins (LDL) from solution. The differential precipitation allows for the quantitation of high-density lipoprotein (HDL) cholesterol in the supernatant fraction [10]. The interaction between LDL and GAGs is mediated by ionic forces between the negatively charged sulfate and carboxyl groups of the GAG and the basic amino acid residues of the lipoprotein's protein constituents.

Apolipoprotein B (Apo B) and apolipoprotein E (Apo E) are protein constituents of LDL. They are the major transport vehicles of cholesterol, cholesteryl esters, phosphoplipids, and triglycerides in the circulation. The interaction of the Apo B and Apo E-containing lipoproteins with extracellular matrix components such as GAGs [11, 12], proteoglycans, collagen, and other components of basement membrane may contribute significantly to the accumulation of cholesterol and cholesteryl esters in arterial tissue and hence contribute to the pathogenesis of arteriosclerosis [13].

One could postulate that the carboxylate groups of the glutamic and aspartic amino acids are aligned in a way resembling the sulfate groups in heparin or possibly chondroitin 6-sulfate such that they form ionic interactions with the basic residues in Apo B and Apo E. This type of interaction would explain the finding that heparin and heparan sulfate displace LDL from its cell receptor.

3. Growth Factors: The affinity of polypeptides for immobilized heparin has led to the purification, amino acid sequencing, and molecular cloning of several mitogenic, angiogenic growth factors, collectively known as heparin-binding growth factors [14, 15]. Two heparinbinding growth factors which are structurally and functionally related to each other are basic and acidic fibroblast growth factor (FGF). Acidic FGF was first identified in the brain by its ability to promote the proliferation of myoblasts in low-density culture. Basic FGF was detected by its effect on the proliferation and phenotype transformation of BALB/c 3T3 fibroblasts.

4. Lipolytic Enzymes: Intravenous heparin administration releases a number of macromolecules into the circulation, including alpha 2-macroglobulin; coagulation factors VIII, IX, and X; diamine oxidase; and the lipolytic enzymes, lipoprotein lipase (LPL) and hepatic triglyceride lipase (HTGL). Both LPL and HTGL play a key role in the metabolism and storage of fat by catalyzing the hydrolysis of triacylglycerols and phospholipids transported in circulating plasma lipoproteins. The free fatty acids produced by these enzymes are used as a fuel source in muscle or are resynthesized into triglycerides for storage in adipose tissue. In addition to anchoring LPL and HTGL to the cell surface, recent evidence suggests that GAGs also regulate the expression of these lypolytic enzymes.

5. Extracellular Matrix Proteins: The ECM plays a central role in the control of cell proliferation, differentiation, and migration by mediating cell adhesion and communication. The major components of the ECM include the cell adhesion proteins, fibronectin, vitronectin, thrombospondin, and laminin; the structural components, collagens and elastin; and the proteoglycans, including serglycin, proteoglycans I and II, versican, decorin, and possibly other GAG-associated proteins that have yet to be identified. The interaction of GAGs with fibronectin (FN) occurs on at least two different sites along the polypeptide chain. From the heparinbinding consensus sequences, Cardin and Weintraub [2] proposed that a GAG-binding domain in the carboxy terminus was contained in the sequence P-R-R-A-R-V. Direct evidence for this sequence being important for heparin binding has been provided by studies using site-directed mutagenesis of this region [15]; converting R-R to T-M in a carboxy-terminal fragment of FN reduced heparin binding 50-fold.

The interaction of vitronectin (VN) and GAGs may have an important physiological role in the regulation of the complement and coagulation systems. It is known that VN is present in serum and tissues and is located at the cell surface. On the basis of the finding that thrombin-AT (TAT) complexes bind to immobilized VN, whereas AT alone does not, it has been postulated that VN binds to a cryptic site on AT, possibly one exposed to the binding interaction with a GAG, thrombin, or a combination of both. VN neutralizes the heparin-induced inactivation of thrombin by AT, as do other heparin-binding proteins, such as Histidine-rich Glycoprotein, but it does not influence the inhibition of thrombin by AT in the absence of heparin. In addition to inhibiting coagulation, VN and its heparin binding domain (residues 347-359) inhibit hemolysis mediated by perforin, a pore forming protein from cytotoxic T-cells, and block the lytic activity of complement.

Laminin is a major ECM constituent of basement membrane, and its interaction with GAGs is thought to play a major role in cell adhesion.

6.Other GAG-Binding Proteins

a) Platelet Factor 4: PF4 is a heat stable heparin-binding protein that is released from alpha granules when platelets aggregate. The protein is released as a PF4-CSPG complex. A number of biological activities have been attributed to PF4, including inhibition of the anticoagulant activity of heparin and the ability to inhibit T suppressor cell activity and megakaryocytopoieses. PF4 is also a potent chemotactic agent for monocytes. The carboxylterminal region of PF4 has also been shown to be chemotactic for human monocytes and to inhibit megakaryocytopoieses in vitro. PF4 and heparin interact with a dissociation constant of 3 x 10-8 M. Approximately 4 PF4 molecules bind to one heparin molecule of 11kDa, suggesting that one heparin molecule interacts with a tetramer with an overall radius of PF4. PF4 does not distinguish between AT-binding and non-binding heparin molecules. When heparin is fragmented, as in ardeparin sodium, the number of oligosaccharides increases. Thus more PF4 is required to neutralize all of the heparin oligosaccharides, giving some fragments with an ATbinding site a chance to escape neutralization. A PF4-heparin complex is now widely accepted to be the major target antigen for the antibodies in heparin-induced thrombocytopenia (HIT). An ELISA developed by Amiral et al. [16] suggests that heparin administration induces the release of PF4 from endothelial cells and elevates plasma PF4 concentrations. The PF4 forms complexes with heparin which bind to the surface of platelets. The Fab portion of the anti-PF4- heparin antibodies then bind the platelet-bound antigenic complexes and induce platelet activation.

b) Viral Coat Proteins: Polyanions, including heparin have been shown to inhibit the growth of human immunodeficiency virus (HIV) and herpes simplex virus (HSV). Heparin binds to the gp120 envelope coat glycoprotein of HIV-1. A gp120 heparin-binding domain corresponds to amino acids 307-330 and is similar in sequence to the heparin-binding domain of vitronectin (residues 353-376). Cardin et al [17] synthesized this peptide and showed that it bound heparin. Heparin binding to the peptide was blocked by a monoclonal antibody specific for the 307-330 region of gp120, which also neutralizes HIV-1 infectivity. In addition, the heparin that bound to a column of immobilized peptide was a more potent inhibitor of HIV-1 replication than the unbound heparin fraction, suggesting an interaction of this heparin with the virus surface glycoprotein. The molecular interactions between this heparin-binding peptide of gp120 and a heparin octasaccharide has been examined by molecular dynamics simulation. Heparin was found to inhibit HIV replication at an IC50 of 7 µg/ml, which corresponds to 1.4 u/ml [18], a dose that could interfere with the coagulation process in vivo. The same study showed that fragmented heparins of lower molecular weight were less active against HIV than unfractionated heparin, and that antiviral activity appears to correlate with anticoagulant activity. Heparin is presumed to interfere with the interaction between gp120 of the virus and the CD4 receptor of the host cell and/or with the chemokine receptor such as CCR5/CXCR4 which mediate virus internalization of non-syncytial and syncytial inducing strains.

c) Prion Proteins: The accumulation in the brain of an abnormally protease-resistant isoform of the host protein PrP is specific to Creutzfeldt-Jakob disease (CJD) and related transmissible spongiform encephalopathies (TSE) and important in the pathogenesis of these neurodegenerative diseases that are characterized by rapid dementia onset, movement impairment and depression [19]. This includes bovine spongiform encephalopathy (BSE) also known as "Mad Cow Disease', and scrapie found in sheep. Scrapie or other forms of TSE are not known to exist in the hog population. Studies in scrapie infected cell cultures have indicated that the protease resistant PrP (PrP-res) is formed from an apparently normal, protease and phospholipase-sensitive PrP precursor (PrP-sen) [20]. It was shown that GAGs (1 - 100 ng/ml) have the capability to bind to PrP-sen [21] which inhibits the conversion to the infectious form, PrP-res, in cell culture. In mice, sulfated polysaccharides administered prior to scrapie infections, or for periods three month post infection, dramatically prolonged the onset of disease [22, 23, 24]. It is suggested that GAGs may block the association of PrP-sen with endogenous cell surface heparan sulfate-like proteoglycans involved in the template-assisted conversion of PrP-sen to PrP-res [25]. Recently, a unique CJD disease believed to be BSE, called new variant CJD (nvCJD), was identified [26].

References