Cell-surface heparan sulfate: an intercalated membrane proteoglycan

Two pools of heparan sulfate proteoglycans have been selectively solubilized from rat liver plasma membranes by successive incubations with heparin and detergent. The two populations of proteoglycans have similar polyanionic properties as indicated by identical elution positions on ion-exchange chromatography on DEAE-Sephacel but differ in buoyant density in CsCl when analyzed by density gradient centrifugation in the presence of 4 M guanidine. The detergent-extracted proteoglycan has a lower buoyant density (less than or equal to 1.40 g/ml) and is, as determined by gel chromatography, slightly larger than the heparin-released proteoglycan (buoyant density, greater than or equal to 1.55 g/ml). Furthermore, in contrast to the heparin-released proteoglycan, the detergent-extracted proteoglycan is able to bind detergent micelles, shows affinity for the hydrophobic gel octyl-Sepharose, and can be inserted into liposomes. We conclude that the detergent-extracted heparan sulfate represents a proteoglycan species that has its core protein rooted in the lipid bilayer of the plasma membrane.

Substrate adhesion of rat hepatocytes: a comparison of laminin and fibronectin as attachment proteins

In previous studies rat hepatocytes have been shown to adhere to substrates composed of collagen or fibronectin. In the present communication, the basement membrane protein laminin is reported to mediated the attachment and spreading of hepatocytes. The cell attachment-mediating activity of laminin was compared with that of fibronectin. The activity of fibronectin was heat sensitive, whereas laminin retained its activity after boiling. On the other hand, reduction and alkylation or periodate oxidation of the proteins affected only the cell attachment activity of laminin. Preincubation of cells with soluble fibronectin inhibited initial cell attachment to fibronectin but not to laminin substrates, and, reversely, soluble laminin selectively inhibited cell attachment to laminin. These results suggest that attachment of cells to substrates of the two proteins involves different cellular receptors recognizing distinct and nonidentical structures in the proteins.

Substrate adhesion of rat hepatocytes: mechanism of attachment to collagen substrates

Attachment of rat hepatocytes to collagen, which occurs without the aid of fibronectin, was found to be a time-dependent reaction characterized by an initial lag phase of 10-20 min before stable attachment bonds began to form. Increasing the density of molecules in the collagen substrates enhanced the rate of cell attachment. The hepatocytes attached essentially equally well to all the collagen types tested (types I, II, III, IV and V). The initial rate of cell attachment was more rapid to native collagen than to denatured collagen or alpha 1(I) chains, apparently indicating different affinities of the cells for these substrates. However, if cells were incubated for 60 min or more, efficient attachment occurred to the alpha 1(I) chain and to all cyanogen-bromide-treated peptides tested (alpha 1-CB2, alpha 1-CB3, alpha 1-CB4, alpha 1-CB5, alpha 1-CB6A, alpha 1-CB7, alpha 1-CB8, alpha 2-CB2, alpha 2-CB3 and alpha 2-CB4) but not to the aminopropeptide of type I procollagen. A low but significant degree of attachment also took place to substrates made of synthetic peptides with the collagen-like structures (Gly-Ala-Pro)n, (Gly-Pro-Pro)n and (Gly-Pro-Hyp)n, whereas no attachment was observed to polyproline. We suggest that the cell-binding sites in collagen have a simple structure and occur in multiple copies along the collagen molecule. Addition of collagen in solution inhibited initial cell attachment, an effect that persisted longer on substrates made of alpha 1(I) chain than on denatured collagen. The collected data are interpreted in terms of a model for cell-to-collagen adhesion where the formation of stable attachment bonds requires the binding of several low-affinity receptors, clustered at the site of adhesion, to collagen molecules in the substrate.

Isolation and characterization of heparin from human mastocytoma tissue

Polysaccharide was isolated from human spleen mastocytoma by proteolytic digestion, precipitation with cetylpyridinium chloride, digestion with chondroitinase ABC, and ion-exchange chromatography on DEAE-cellulose. The final product (0.7 mg per g of starting material, MW 8000) behaved like standard heparin on ion-exchange chromatography and on electrophoresis, and contained D-glucuronic acid, L-iduronic acid, D-glucosamine and sulfate in the proportions expected for heparin. Affinity chromatography on antithrombin-Sepharose separated a distinct high-affinity fraction (4-5% of the total material). Structural analysis of this fraction showed that about 10% of the D-glucosamine residues were N-acetylated, the remainder N-sulfated. The anticoagulant activity of the isolated heparin was 71 B.P. units per mg (whole-blood system), or 30 units per mg (anti-thrombin and chromogenic substrate). 205 and 10-15 units per mg (chromogenic assay) were found for high and low affinity fractions, respectively. These results demonstrate conclusively the occurrence of heparin in a human tissue.

Characterization of a platelet endoglycosidase degrading heparin-like polysaccharides

An endoglycosidase (heparitinase) acting on heparin and heparan sulfate was partially purified (approximately 300 times) from human platelets by affinity chromatography on heparan sulfate substituted Sepharose. Only heparin-like polysaccharides were degraded by the enzyme. The susceptibility of various biosynthetic heparin intermediates indicated that the platelet heparitinase had a requirement for sulfamino but not ester sulfate groups. No activity toward other uronic acid containing glycosaminoglycans could be demonstrated. Glucuronidic but not glucosaminidic linkages in heparin or heparan sulfate were attacked by the enzyme as shown by analysis of the reducing sugar moiety in oligosaccharide products. The anticoagulant activity of heparin, determined in an antithrombin III activation assay, was markedly reduced after treatment with the heparitinase. The enzyme was released from its storage site in platelets after induction of the platelet release reaction. The physiological function of platelet heparitinase is not known but may be to modify extracellular heparin-like polysaccharides in the vascular system.

Interaction of lipoprotein lipase with native and modified heparin-like polysaccharides

1. Lipoprotein lipase (EC 3.1.1.34), which was previously shown to bind to immobilized heparin, was now found to bind also to heparan sulphate and dermatan sulphate and to some extent to chondroitin sulphate. 2. The relative binding affinities were compared by determining (a) the concentration of NaCl required to release the enzyme from polysaccharide-substituted Sepharose; (b) the concentration of free polysaccharides required to displace the enzyme from immobilized polysaccharides; and (c) the total amounts of enzyme bound after saturation of immobilized polysaccharides. By each of these criteria heparin bound the enzyme most efficiently, followed by heparan sulphate and dermatan sulphate, which were more efficient than chondroitin sulphate. 3. Heparin fractions with high and low affinity for antithrombin, respectively, did not differ with regard to affinity for lipoprotein lipase. 4. Partially N-desulphated heparin (40-50% of N-unsubstituted glucosamine residues) was unable to displace lipoprotein lipase from immobilized heparin. This ability was restored by re-N-sulphation or by N-acetylation; the N-acetylated product was essentially devoid of anticoagulant activity. 5. Partial depolymerization of heparin led to a decrease in ability to displace lipoprotein lipase from heparin-Sepharose; however, even fragments of less than decasaccharide size showed definite enzyme-releasing activity. 6. Studies with hepatic lipase (purified from rat post-heparin plasma) gave results similar to those obtained with milk lipoprotein lipase. However, the interaction between the hepatic lipase and the glycosaminoglycans was weaker and was abolished at lower concentrations of NaCl. 7. The ability of the polysaccharides to release lipoprotein lipase to the circulating blood after intravenous injection into rats essentially conformed to their affinity for the enzyme as evaluated by the experiments in vitro.

Heparin enhances the rate of binding of fibronectin to collagen

125I-labelled fibronectin is shown to bind to both native and denatured collagen immobilized on Sephadex beads in reactions that exhibit different kinetics. The rates of both reactions were enhanced by the presence of heparin or highly sulphated dextran sulphate but not by other glycosaminoglycans or dextran sulphates having low sulphate contents.

Cell-surface heparan sulfate. Isolation and characterization of a proteoglycan from rat liver membranes

Solubilization of heparan sulfate proteoglycans from a rat liver membrane fraction was obtained by the use of the charged detergent deoxycholate or alternatively a combination of NaCl and the nonionic detergent Triton-X 100. Subsequently, proteoglycans solubilized from microsomal and plasma membrane fractions, respectively, were purified by a procedure involving gel chromatography, anion exchange chromatography, and density gradient centrifugation. The purified heparan sulfate proteoglycan had a molecular weight of about 75,000 as determined by sedimentation equilibrium analysis or gel chromatography. Molecular weights of 17,000 to 40,000 were obtained for the 125I-labeled core protein after removal of the heparan sulfate polysaccharide chains by different enzymatic and chemical methods. An average molecular weight of 14,000 was found for the polysaccharide chains released from the core protein by alkali treatment. The data are consistent with a proteoglycan structure containing four polysaccharide chains attached to the core protein. The amino acid composition of native and alkali-treated proteoglycan support the proposed proteoglycan model.

Structure of the antithrombin-binding site in heparin

Heparin preparations from pig intestinal mucosa and from bovine lung were separated by chromatography on antithrombin-Sepharose into a high-affinity fraction (with high anticoagulant activity) and a low-affinity fraction (with low anticoagulant). Antithrombin-binding heparin fragments (12-16 monosaccharide units) were prepared, either by digesting a high-affinity heparin-antithrombin complex with bacterial heparinase or by partial deaminative cleavage of the unfractionated polysaccharide with nitrous acid followed by affinity chromatography on immobilized antithrombin. Compositional analysis based on separation and identification of deamination products reduced with sodium boro[3H]hydride showed that nonsulfated L-iduronic acid occurred in larger amounts in high-affinity heparin than in low-affinity heparin; furthermore, this component was concentrated in the antithrombin-binding regions of the high-affinity heparin molecules, amounting to approximately one residue per binding site. It is suggested that nonsulfated L-iduronic acid is essential for the anticoagulant activity of heparin. The location of the non-sulfated uronic acid in the antithrombin-binding site was determined by periodate oxidation of antithrombin-binding fragments containing a terminal 2,5-anhydro-D-[1-3H]mannitol unit. Tentative structures for antithrombin-binding sequences in heparin are proposed, including some structural variants believed to be compatible with, but not required for, activity.

Isolation of the pericellular matrix of human fibroblast cultures

The pericellular matrix of human fibroblast cultures was isolated, using sequential extraction with sodium deoxycholate and hypotonic buffer in the presence of protease inhibitor. The matrix attached to the growth substratum had a "sackcloth-like" structure as seen by phase contrast, immunofluorescence, and scanning electron microscopy, and it had a vaguely filamentous ultrastructure similar to that seen in intact cell layers. The matrix consisted of hyaluronic acid and heparan sulfate as the major glycosaminoglycan components and fibronectin and procollagen as major polypeptides as shown by metabolic labeling, gel electrophoresis, immunofluorescence, and collagenase digestion. This pericellular matrix can be regarded as an in vitro equivalent of the loose connective tissue matrix.

Identification of N-sulphated disaccharide units in heparin-like polysaccharides

1. Preparations of heparin and heparan sulphate were degraded with HNO2. The resulting disaccharides were isolated by gel chromatography, reduced with either NaBH4 or NaB3H4 and were then fractionated into non-sulphated, monosulphated and disulphated species by ion-exchange chromatography or by paper electrophoresis. The non-sulphated disaccharides were separated into two, and the monosulphated disaccharides into three, components by paper chromatography. 2. The uronic acid moieties of the various non- and mono-sulphated disaccharides were identified by means of radioactive labels selectively introduced into uronic acid residues (3H and 14C in D-glucuronic acid, 14C only in L-iduronic acid units) during biosynthesis of the polysaccharide starting material. Labelled uronic acids were also identified by paper chromatography, after liberation from disaccharides by acid hydrolysis or by glucuronidase digestion. Similar procedures, applied to disaccharides treated with NaB3H4, indicated 2,5-anhydro-D-mannitol as reducing terminal unit. On the basis of these results, and the known positions and configurations of the glycosidic linkages in heparin, the two non-sulphated disaccharides were identified as 4-O-(beta-D-glucopyranosyluronic acid)-2,5-anhydro-D-mannitol and 4-O-(alpha-L-idopyranosyluronic acid)-2,5-anhydro-D-mannitol. 3. The three monosulphated [1-3H]anhydromannitol-labelled disaccharides were subjected to Smith degradation or to digestion with homogenates of human skin fibroblasts, and the products were analysed by paper electrophoresis. The results, along with the 1H n.m.r. spectra of the corresponding unlabelled disaccharides, permitted the allocation of O-sulphate groups to various positions in the disaccharides. These were thus identified as 4-O-(beta-D-glucopyranosyl-uronic acid)-2,5-anhydro-D-mannitol 6-sulphate, 4-O-(alpha-L-idopyranosyluronic acid)-2,5-anhydro-D-mannitol 6-sulphate and 4-O-(alpha-L-idopyranosyluronic acid 2-sulphate)-2,5-anhydro-D-mannitol. The last-mentioned disaccharide was found to be a poor substrate for the iduronate sulphatase of human skin fibroblasts, as compared with the disulphated species, 4-O-(alpha-L-idopyranosyluronic acid 2-sulphate)-2,5-anhydro-D-mannitol 6-sulphate. 4. The identified [1-3H]anhydromannitol-labelled disaccharides were used as reference standards in a study of the disaccharide composition of heparins and heparan sulphates. Low N-sulphate contents, most pronounced in the heparin sulphates, were associated with high ratios of mono-O-sulphated/di-O-sulphated (N-sulphated) disaccharide units, and in addition, with relatively large amounts of 2-sulphated L-iduronic acid residues bound to C-4 of N-sulpho-D-glucosamine units lacking O-sulphate substituents.

The molecular-weight-dependence of the anti-coagulant activity of heparin

It is proposed that the anti-coagulant activity of heparin is related to the probability of finding, in a random distribution of different disaccharides, a dodecasaccharide with the sequence required for binding to antithrombin. It is shown that this probability is a function of the degree of polymerization of heparin. The hypothesis has been been tested with a series of narrow-molecular-weight-range fractions ranging from 5,600 to 36,000. The fractions having mol.wts. below 18,000 (comprising 85% of the original preparation) followed the predicted probability relationship as expressed by the proportion of molecules capable of binding to antithrombin. The probability that any randomly chosen dodecasaccharide sequence in heparin should bind to antithrombin was calculated to 0.022. The fraction with mol.wt. 36,000 contained proteoglycan link-region fragments, which may explain the deviation of the high-molecular-weight fractions from the hypothetical relationship. The relationship between anti-coagulant activity and molecular weight cannot be explained solely on the basis of availability of binding sites for antithrombin. The activity of high-affinity heparin (i.e. molecules containing high-affinity binding sites for antithrombin), determined either by a whole-blood clotting procedure or by thrombin inactivation in the presence of antithrombin, thus remained dependent on molecular weight. Possible explanations of this finding are discussed. One explanation could be a requirement for binding of thrombin to the heparin chain adjacent to antithrombin.