Heparin-binding Properties Research Articles

Analysis of Putative Heparin-binding Domains of Fibroblast Growth Factor-1: USING SITE-DIRECTED MUTAGENESIS AND PEPTIDE ANALOGUES

The contribution of individual basic amino acids within three putative "consensus sequences" for heparin binding of fibroblast growth factor-1 have been examined by site-directed mutagenesis. The results indicate that a significant reduction in the apparent affinity of fibroblast growth factor-1 for heparin is only observed when basic residues in one of the three regions are mutated. Mutation in the other regions are without affect on heparin binding. The heparin binding properties of synthetic peptides based on the three "consensus sequences" paralleled the mutagenesis results. That is, synthetic peptides corresponding to regions of the protein that were affected by mutagenesis with respect to heparin binding exhibited a relatively high affinity for immobilized heparin, whereas those corresponding to regions of similar charge density that were unaffected by mutagenesis did not. In addition, amino acid substitution of a nonbasic residue in the heparin-binding peptide could abolish its heparin binding capacity. The heparin-binding peptide could antagonize the mitogenic activity of FGF-1, probably because of the heparin dependence of this activity. Together these data demonstrate that the heparin binding properties of fibroblast growth factor-1 are dictated by structural features more complex than clusters of basic amino acids. The results of these and other studies indicate that consensus motifs for heparin-binding require further definition. More importantly, the results provide a basis for the design of peptide-based inhibitors of FGF-1.

Cellular receptor structures for pseudorabies virus are blocked by antithrombin III.

Pseudorabies virus (PrV), an alphaherpesvirus of swine, uses cellular heparan sulfate residues as a receptor for attachment. Interaction of the virus with its receptor is mediated by the envelope glycoprotein C (PrV-gC), a protein with heparin-binding properties. We have previously shown that a region of this protein shows structural similarities to the high-affinity heparin-binding site of the serum protease-inhibitor antithrombin III (ATII). In this publication, we describe the effect of ATIII on interaction of PrV with its cellular receptor. ATIII bound specifically to heparan sulfate residues on the surface of herpesvirus-permissive RK13 cells. Binding of ATIII to RK13 cells interfered with adsorption of radioactively labelled PrV to these cells. Enzymatic treatment using heparinase I (E.C. 4.2.2.7) removed the receptor for PrV as well as the receptor for ATIII. Since amino acids 130-137 of the high affinity heparin-binding site of ATIII show structural similarities to amino acids 134-141 of PrV-gC, both sequences were synthesized as synthetic peptides. Although interaction of the peptide derived from ATIII with heparin was significantly stronger, both peptides interacted specifically with heparin in assays in vitro. These results suggest that PrV and ATIII interact with the same structure on the cellular surface.

Expression and Characterization of Murine Osteoblast-Specific Factor 2 (OSF-2) in a Baculovirus Expression System

Osteoblast-specific factor 2 (OSF-2) is a ∼90-kDa protein selectively expressed in bone. OSF-2 cDNA was recently isolated from mouse and human cDNA libraries and shows limited sequence homology with fasciclin I, a cell adhesion protein expressed in insect nerve cells, Here we describe the expression of recombinant murine OSF-2 (rmOSF-2) in a baculovirus/insect cell system, Western blotting analysis employing polyclonal antiserum raised against a C-terminal synthetic OSF-2 peptide detected a protein of ∼90-kDa as early as 2 days after infection of Sf9 cells with the recombinant virus, Tunicamycin treatment of infected cells resulted in a mobility shift of OSF-2 (∼90-kDa band) on Western blots, N-Glycanase digestion resulted in the same mobility shift of OSF-2, indicating that rmOSF-2 expressed in insect cells is N-glycosylated, However, OSF-2 was insensitive to endoglycosidase H digestion while a major fraction of this protein had affinity for concanavalin A. Finally, it was demonstrated that rmOSF-2 was able to bind to heparin. This finding suggests that OSF-2 might be associated with the bone extracellular matrix after secretion by osteo blasts and participate in cell adhesion and/or cell communication. The establishment of the baculovirus expression system with a high productivity of recombinant OSF-2 (around 40 μg/ml at maximum) and its heparin binding properties should allow us to obtain large amounts of rmOSF-2.

Engineering of fibroblast growth factor: alteration of receptor binding specificity.

A five-residue loop structure in basic fibroblast growth factor (FGF-2) which extends from amino acid residue 118 to residue 122 was replaced, by cassette mutagenesis, with the corresponding seven-residue loop structure from the structural homologue acidic fibroblast growth factor (FGF-1) or the corresponding five-residue loop from interleukin-1 beta to give FGF-2LA and FGF-2LI, respectively. The mutants were expressed in Escherichia coli and purified to homogeneity, and their heparin and receptor binding and biological properties were examined. The ability of FGF-2LA to induce endothelial cell proliferation was the same as that of FGF-2. Affinities of the mutants to heparin and to cells that express FGF receptor-1 (FGFR-1) were identical to those of the wild-type protein. The role of the loop structure in FGF-1 and FGF-2 was elucidated by using soluble FGF receptor systems, which display distinct ligand binding specificities. Thus, FGF-2LA bound, with the same affinity as FGF-1 and FGF-2, to FGFR-1 and FGFR-2, whereas only FGF-1 and the FGF-1 loop-containing mutant, FGF-2LA, bound to the keratinocyte growth factor receptor. A change in receptor binding specificity was not observed with the FGF-2LI engineered mutant. That the binding specificity of FGF-2 was dramatically altered by transfer of a loop structure from FGF-1 to resemble the binding profile of the donor protein provides strong evidence that this motif is a receptor binding specificity determinant of fibroblast growth factors.

Secondary structure of neutrophil-activating peptide-2 determined by 1H-nuclear magnetic resonance spectroscopy

Neutrophil-activating protein-2 (NAP-2) is a 72 residue protein demonstrating a range of proinflammatory activities. The solution structure of monomeric NAP-2 has been investigated by two-dimensional 1H-n.m.r. spectroscopy. Sequence-specific proton resonance assignments have been made and secondary structural elements have been identified on the basis of nuclear Overhauser data, coupling constants and amide hydrogen/deuteron exchange. The NAP-2 monomer consists of a triple-stranded anti-parallel beta-sheet arranged in a 'Greek key' and a C-terminal helix (residues 59-70) and is very similar to that found in the n.m.r. solution conformation of dimeric interleukin-8 and the crystal structure of tetrameric bovine platelet factor-4. Results are discussed in terms of heparin binding and neutrophil-activation properties of NAP-2.

An Unusual Heparin-Binding Peptide from the Carboxy-Terminal hep-2 Region of Fibronectin

A synthetic 22 residue peptide, N22W, with sequence NVSPPRRARVTDATETTITISW, derived from the amino terminus of type III module 13 in the carboxyterminal hep-2 domain of fibronectin, was found to exhibit unusual heparin binding properties. Titration of fluoresceinamine-labeled heparin (FA-heparin) with N22W at 25°C and pH 7.4 in 0.02 M Tris buffer containing 0.15 M NaCl (TBS) produced a cooperative sigmoidal increase in fluorescence polarization anisotropy with half-saturation near 2.5 μM. The increase in anisotropy was even greater than that produced by the much larger 30-kDa hep-2 fragment of fibronectin and saturation was achieved at lower concentration. Simply deleting the C-terminal Trp from the peptide abolished its heparin-binding activity as did deletion of residues TETTITIS or mutation of the RR doublet to SS. Further analysis suggested that peptide-peptide interactions mediated by the carboxy-terminal region of N22W play an important role in its binding to heparin. A branched tetrameric peptide containing four copies of N21S caused a nearly hyperbolic increase in anisotropy of FA-heparin with an apparent K d of 0.3 μM in TBS, 10-fold lower than that of the monomer or of the parent domain from which the peptide was derived. The results illustrate that peptide-peptide interactions can lead to stronger binding by allowing multiple points of contact with the negatively charged polysaccharide.

Diminished heparin binding of a basic fibroblast growth factor mutant is associated with reduced receptor binding, mitogenesis, plasminogen activator induction, and in vitro angiogenesis.

Using modeling of heparin-fibroblast growth factor interactions, we replaced four basic residues of basic fibroblast growth factor (FGF-2) with neutral glutamine residues by site-specific mutagenesis to give the mutants K128Q, K138Q, K128Q-K138Q, R129Q, K134Q, and R129Q-K134Q. The FGF mutants were characterized for their receptor and heparin binding affinities, mitogenic and cell proliferation activities, and their ability to induce plasminogen activator (PA) production and in vitro angiogenesis by cultured endothelial cells. Heparin binding properties and biological activities of the three mutants involving R129 and K134 remained essentially unchanged; however, significant changes for three mutants involving K128 and K138 were found. The KD values for heparin binding for K128Q and K138Q mutants were increased about 10-fold, and that for the K128Q-K138Q double mutant was increased by about 100-fold. The mutant K128Q-K138Q required a 10-fold higher concentration of heparin to promote binding to heparan sulfate proteoglycan (HSPG)-deficient CHO cells transfected with fibroblast growth factor receptor-1 (FGFR1) or to induce DNA synthesis in HSPG-deficient myeloid cells transfected with FGFR1. Binding affinities of the mutants to cell surface receptors on BHK-21 cells, however, were similar to that of wild-type FGF-2. In endothelial cell proliferation assays the activities of K128Q and K128Q-K138Q were about 10-fold lower than that of the wild-type protein, whereas the K138Q mutant exhibited wild-type activity. In addition, the K128Q-K138Q mutant displayed a markedly lowered capacity to induce PA activity in cultured endothelial cells and to form capillary-like structures in an in vitro angiogenesis model.(ABSTRACT TRUNCATED AT 250 WORDS)

Evidence that the heparin-binding consensus sequence of vitronectin is recognized by Staphylococcus aureus.

Binding of heparin-binding form of vitronectin to Staphylococcus aureus was inhibited completely by heparin or by the same form of vitronectin. The binding was inhibited only to about 50% by the non-heparin-binding form of vitronectin, indicating an apparent involvement of the heparin-binding properties in the interaction between vitronectin and S. aureus. This was supported by experiments in which a synthetic peptide (Ala347-Arg361, comprising heparin-binding consensus sequences) was found to partly inhibit bacterial adherence to immobilized vitronectin. A bacterial cell surface protein could bind to the quinquedecapeptide, but not to the highly charged peptides consisting entirely of arginine or lysine, immobilized on microtiter plates and the binding could be competitively inhibited by an excess of soluble peptide. Direct binding of radiolabeled peptide to bacterial cells was also demonstrated, which was rapid, saturable, and pH-dependent. Furtherly a bacterial surface protein having molecular mass of 60 kDa was isolated by affinity chromatography on a quinquedecapeptide-HiTrap-NHS column. Our data suggest that the heparin-binding properties of vitronectin play a role in bacterial recognition.

Inhibition of rat mast cell protease 1 by vitronectin

Rat mast cell protease 1 (RMCP-1) is a chymotrypsin-like serine protease specifically expressed by connective tissue-type mast cells. The enzyme is stored in the secretary granules in a macromolecular complex with heparin proteoglycan. In the present investigation it was shown that RMCP-1 is inhibited by vitronectin (VN), an RGD-containing adhesive glycoprotein with heparin-binding properties. RMCP-1 that had been separated from heparin proteoglycan was less susceptible to inhibition than RMCP-1 present in complex with heparin proteoglycan. Pre-incubation of VN with purified heparin partially blocked the RMCP-1 inhibiting activity of VN. Plasma VN had negligible RMCP-1-inhibiting activity. However, heat treatment of plasma VN, which is known to expose the heparin-binding domain, induced RMCP-1-inhibiting activity. Affinity chromatography on immobilized VN showed that RMCP-1 bound with high affinity to VN. The binding of RMCP-1 to VN was not heparin-dependent since free RMCP-1 bound with equal affinity to the immobilized VN as RMCP-1 present in complex with heparin. The inhibition of RMCP-1 by VN was shown to be reversible.

Lipoprotein lipase domain function.

Human lipoprotein lipase (LPL) monomer consists of two domains, a larger NH2-terminal domain that contains catalytic residues and a smaller COOH-terminal domain that modulates substrate specificity and is a major determinant of heparin binding. Analyses of NH2-terminal domain function were performed after site-directed mutagenesis of the putative active-site serine residue, while COOH-terminal domain function was assessed following reaction with a monoclonal antibody. The native enzyme and mutant LPL in which serine 132 was replaced with alanine, cysteine, or glycine were transiently expressed in COS-7 cells. Mutant proteins were synthesized and secreted at levels comparable to native LPL; however, none of the mutants retained enzymatic activity. The mutant with alanine replacing serine 132 was purified and shown to be inactive with both esterase and lipase substrates; however, binding to a 1,2-didodecanoyl-sn-glycero-3-phosphatidylcholine monolayer was comparable to native LPL. These results are consistent with a catalytic, and not a lipid binding, role for serine 132. To investigate the function of the smaller COOH-terminal domain, LPL lipolytic and esterolytic activities as well as heparin binding properties were determined after reaction with a monoclonal antibody specific for this domain. Lipolytic activity was inhibited by the monoclonal antibody, whereas esterolytic activity was only marginally affected, indicating that the LPL COOH-terminal domain is required for lipolysis, perhaps by promoting interaction with insoluble substrates. Also, the affinity of antibody-reacted LPL for heparin was not significantly different from that of LPL alone, suggesting that (i) the heparin-binding site is physically distinct from the COOH-terminal domain region required for lipolysis and (ii) binding of antibody did not cause dimer dissociation. A model is proposed for the two LPL domains fulfilling different roles in the lipolytic process.

Heparin modulates the interaction of VEGF165 with soluble and cell associated flk-1 receptors.

The 165-amino acid form of vascular endothelial growth factor (VEGF165) is a mitogen for vascular endothelial cells and a potent angiogenic factor. Expression of a chimeric receptor containing the extracellular domain of the flk-1 receptor fused to the transmembrane and intracellular domains of the human c-fms receptor in NIH-3T3 cells, resulted in the appearance of high affinity binding sites for 125I-VEGF165 on transfected cells. The binding of 125I-VEGF165 to the flk-1/fms chimeric receptor of the transfected cells as well as the VEGF165-induced autophosphorylation of the chimeric receptors were inhibited in the presence of low concentrations of heparin (1-10 micrograms/ml). In contrast, similar concentrations of heparin potentiated the binding of 125I-VEGF165 to the endogenous VEGF receptors of the transfected cells, indicating that to some extent, the effect of heparin on 125I-VEGF165 binding is receptor type-dependent. A soluble fusion protein containing the extracellular domain of flk-1 fused to alkaline phosphatase (flk-1/SEAP) was used to study the effects of heparin on the binding of 125I-VEGF165 to flk-1 in a cell-free environment. The fusion protein specifically inhibited VEGF165-induced proliferation of vascular endothelial cells, but bound 125I-VEGF165 inefficiently in the absence of heparin. Addition of low concentrations of heparin or heparan sulfate (0.1-1 microgram/ml) resulted in a strong potentiation of 125I-VEGF165 binding, whereas higher heparin or heparan sulfate concentrations inhibited the binding. The effect of heparin on the binding of 125I-VEGF165 to flk-1/SEAP could not be mimicked by desulfated heparin or by chondroitin sulfate. Both bFGF and aFGF inhibited the binding when low concentrations of heparin were added to the binding reaction. However, higher concentrations of heparin abolished the inhibition, indicating that the inhibition is probably caused by competition for available heparin. Taken as a whole, these results indicate that heparin-like molecules regulate the binding of VEGF165 to its receptors in complex ways which depend on the heparin binding properties of VEGF165, on the specific VEGF receptor type involved, and on the amount and composition of heparin-like molecules that are present on the cell surface of VEGF receptor containing cells.

Protection of transforming growth factor β activity by heparin and fucoidan

The transforming growth factor-beta (TGF-beta) family of proteins exert diverse and potent effects on proliferation, differentiation, and extracellular matrix synthesis. However, relatively little is known about the stability or processing of endogenous TGF-beta activity in vitro or in vivo. Our previous work indicated that 1) TGF-beta 1 has strong heparin-binding properties that were not previously recognized because of neutralization by iodination, and 2) heparin, and certain other polyanions, could block the binding of TGF-beta 1 to alpha 2-macroglobulin (alpha 2-M). The present studies investigated the influence of heparin-like molecules on the stability of the TGF-beta 1 signal in the pericellular environment. The results indicate that heparin and fucoidan, a naturally occurring sulfated L-fucose polymer, suppress the formation of an initial non-covalent interaction between 125I-TGF-beta 1 and activated alpha 2-M. Electrophoresis of 125I-TGF-beta 1 showed that fucoidan protects TGF-beta 1 from proteolytic degradation by plasmin and trypsin. While plasmin caused little, if any, activation of latent TGF-beta derived from vascular smooth muscle cells (SMC), plasmin degraded acid-activated TGF-beta, and purified TGF-beta 1, and this degradation was inhibited by fucoidan. In vitro, heparin and fucoidan tripled the half-life of 125I-TGF-beta 1 and doubled the amount of cell-associated 125I-TGF-beta 1. Consistent with this protective effect, heparin- and fucoidan-treated SMC demonstrated elevated levels of active, but not latent, TGF-beta activity.

A Novel Heparin-Binding Protein, HBp15, Is Identified as Mammalian Ribosomal Protein L22

A 15kDa-protein (HBp15) was purified from mouse submandibular gland and bovine brain by virtue of its heparin-binding property. The amino acid sequences of mouse and bovine HBp15 showed a high degree of homology to a sea urchin protein encoded by gene called "development specific protein 217." Using reverse transcription-polymerase chain reaction methods, cDNA clones for HBp15 were isolated from submandibular gland mRNA of mouse, human and pig, and sequenced. Database search of HBp15 showed that HBp15 also resembles yeast ribosomal protein YL31 in addition to the 217 protein. Using specific antibodies against HBp15, HBp15 was identified as mammalian ribosomal protein L22, for which no sequence information is available.

Functional domains of recombinant bactericidal/permeability increasing protein (rBPI23).

The 23-kDa recombinant amino-terminal bactericidal/permeability increasing protein fragment (rBPI23) has all of the antibacterial and antiendotoxin properties of the holoprotein. In the current studies, we have identified multiple active domains within rBPI23 with chemical and proteolytic cleavage fragments and with synthetic overlapping peptides. We also demonstrate a novel, high affinity heparin binding property for rBPI23, in addition to its established bactericidal and lipopolysaccharide binding properties. Cleavage fragments and synthetic, overlapping peptides of rBPI23 were analyzed for inhibition of the lipopolysaccharide-induced Limulus amebocyte lysate reaction, for bactericidal activity, and for heparin binding. Three separate, active domains were identified in amino acid regions 17-45, 65-99, and 142-169. A single synthetic peptide (85-99) was bactericidal. These results indicate that rBPI23 is comprised of three separate functional domains which contribute to the high affinity interaction of rBPI23 with Gram-negative bacteria. The individual activity of each domain and the cooperative interaction among domains provide the basis for developing rBPI23 analogues with increased biologic efficacy.

A peptide-model for the heparin-binding property of pseudorabies virus glycoprotein III

The pseudorabies virus glycoprotein III (PrV-gIII) has been identified previously as the major viral component binding to a heparin-like receptor on the surface of target cells. The amino acid sequence of gIII contains three regions corresponding to consensus sequences for heparin binding. A synthetic peptide corresponding to amino acids 134 to 141 of PrV-gIII bound heparin in a dot blot assay. In contrast, a synthetic peptide derived from amino acids 290-299 of PrV-gIII did not bind heparin. We therefore conclude that the region containing amino acid 134-141 is involved in binding to the heparin-like cellular receptor.

Spleen-derived growth factor, SDGF-3, is identified as keratinocyte growth factor (KGF)

A heparin-binding mitogen rat for rat hepatocytes was partially purified from bovine spleen by a combination of heparin-affinity, cation-exchange and gel-filtration chromatography. Besides stimulating rat hepatocytes, this factor, which was designated spleen-derived growth factor-3 (SDGF-3), exhibited mitogenic activity for mouse epidermal keratinocytes but not mouse fibroblasts. Its apparent epithelial specificity and heparin-binding properties corresponded to those of keratinocyte growth factor (KGF). These findings, together with the fact that the mitogenic activity of SDGF-3 was abolished by a neutralizing monoclonal antibody specific for KGF, identify this bovine spleen-derived hepatocyte mitogen as KGF.

Functional characterization of a chimeric lipase genetically engineered from human lipoprotein lipase and human hepatic lipase

Lipoprotein lipase (LPL) and hepatic lipase (HL) mediate the hydrolysis of triglycerides and phospholipids present in circulating lipoprotein particles and are essential for normal lipid metabolism. Both enzymes have a similar primary amino acid structure and share requirements for intact catalytic, lipid binding, and heparin binding domains. However, LPL and HL exhibit different substrate specificities and cofactor requirements. In order to characterize the functional domains necessary for LPL activity, a chimeric lipase consisting of the amino-terminal 314 amino acids of human LPL and the carboxyl-terminal 146 amino acids of human HL was synthesized by joining the cDNA of both lipases at the 5'-end of exon 7. Northern blot hybridization and Western blot analyses revealed the size of the chimera mRNA and protein to be approximately 1.5 kb and 55 kDa, respectively. The chimeric enzyme hydrolyzed both long chain and short chain fatty acid triacylglycerols and had catalytic properties that were similar to lipoprotein lipase. Thus, apolipoprotein (apo)C-II was required for maximal lipase activity, and high salt concentration abolished the ability of the chimera to hydrolyze triolein even in the presence of apoC-II. A monospecific anti-HL polyclonal antibody interacting with the C-terminal HL-derived domain of the chimeric enzyme abolished the enzyme's ability to hydrolyze triglyceride emulsion but not tributyrin substrates. Analysis of the heparin binding properties of the chimeric enzyme using heparin-Sepharose affinity chromatography revealed an elution pattern which was intermediate between that of lipoprotein and hepatic lipase. In summary, we have characterized the functional properties of an LPL-HL chimeric enzyme.(ABSTRACT TRUNCATED AT 250 WORDS)

Structural and functional specificity of FGF receptors

Fibroblast growth factors (FGFs) represent a group of polypeptide mitogens eliciting a wide variety of responses depending on the target cell type. The knowledge of the cell surface receptors mediating the effects of FGFs has recently expanded remarkably. Perhaps not surprisingly, the complexity of the FGF family and FGF induced responses is reflected in the diversity and redundancy of the FGF receptors. The molecular cloning of the signal transducing receptors for fibroblast growth factors has revealed a tyrosine kinase gene family with at least four members. Differential splicing and polyadenylation creates further diversity in the FGF receptor system. These numerous receptor forms have both distinct and redundant properties. We are only now beginning to understand how the different receptors are activated by the various FGFs and how they are expressed by various cells and tissues. FGF binding to the tyrosine kinase receptors needs the assistance of heparan sulphate side chains of proteoglycans present at the cell surface and in the extracellular matrix. As several other growth factors share the heparin binding property of FGFs, the dual receptor system for FGFs might be an example of a more widely used principle.

Expression of biologically active recombinant keratinocyte growth factor. Structure/function analysis of amino-terminal truncation mutants.

Keratinocyte growth factor (KGF) is a newly identified member of the fibroblast growth factor (FGF) family (FGF-7). KGF is expressed by stromal fibroblasts and acts on epithelial cells in a paracrine mode. To facilitate structure/function studies, we utilized the T7 prokaryotic expression system to synthesize this growth factor. Recombinant KGF (rKGF) was mitogenic with a specific activity around 10-fold higher than native KGF. By in vitro mutagenesis, we generated a series of KGF mutants with sequential deletions of the amino-terminal domain, the most divergent region among different FGF members. Mutant proteins, produced in bacteria, were tested for their ability to bind heparin, bind and activate the KGF receptor, and induce DNA synthesis. Heparin binding properties were preserved with deletion of up to 28 amino-terminal residues of the mature KGF but lost by the deletion of an additional 10 residues. Biological activity of mutants with deletions of up to 10 residues was comparable to that of rKGF. However, deletion of 29 residues resulted in significantly reduced ability to stimulate KGF receptor tyrosine-kinase activity and DNA synthesis, although this mutant bound the receptor at high affinity. These characteristics of a partial agonist may be useful in the development of competitive antagonists of KGF action.

Structure, composition and heparin binding properties of a human cytomegalovirus glycoprotein complex designated gC-II.

The structure and heparin binding properties of a family of human cytomegalovirus (HCMV) disulphide-linked glycoprotein complexes designated gC-II were analysed. gC-II complexes contain two groups of glycoproteins designated Group 1 and Group 2. These glycoproteins were separated from each other by short exposure of virions to a reducing agent. This showed that the disulphide bonds between these glycoproteins were on the external surface of the virion. Although these glycoproteins were no longer associated they were not released from the virion, suggesting that they were transmembrane glycoproteins. Approximately 75 to 90% of the gC-II complexes and 18% of the complexes containing the HCMV gB glycoprotein obtained from the virion envelope bound immobilized heparin. When virions were incubated with [3H]heparin, gC-II complexes bound more heparin than gB complexes, by approximately threefold. These data showed that gC-II complexes had a greater heparin-binding capacity. After treatment of virions with a reducing agent the affinity of gC-II glycoproteins for heparin was greatly reduced whereas the affinity of gB glycoproteins was only slightly reduced. Thus, higher order structure was important for heparin binding by gC-II complexes but not by those of gB. Relative to gC-II Group 1 glycoproteins, a greater portion of gC-II Group 2 glycoproteins still bound to heparin after reduction, suggesting that Group 2 glycoproteins may be the important heparin binding component of the gC-II complexes. Both gB and gC-II complexes were eluted from immobilized heparin with soluble heparin or 0.65 M-NaCl suggesting that both formed ionic bonds with heparin. Chondroitin sulphate was not effective at eluting HCMV envelope glycoproteins from immobilized heparin. Thus, the structure of the glucosaminoglycan backbone is important to the binding of HCMV glycoproteins to heparin.