Identification and dynamics of a heparin-binding site in hepatocyte growth factor.

Hepatocyte growth factor (HGF) is a heparin-binding, multipotent growth factor that transduces a wide range of biological signals, including mitogenesis, motogenesis, and morphogenesis. Heparin or closely related heparan sulfate has profound effects on HGF signaling. A heparin-binding site in the N-terminal (N) domain of HGF was proposed on the basis of the clustering of surface positive charges [Zhou, H., Mazzulla, M. J., Kaufman, J. D., Stahl, S. J., Wingfield, P. T., Rubin, J. S., Bottaro, D. P., and Byrd, R. A. (1998) Structure 6, 109-116]. In the present study, we confirmed this binding site in a heparin titration experiment monitored by nuclear magnetic resonance spectroscopy, and we estimated the apparent dissociation constant (K(d)) of the heparin-protein complex by NMR and fluorescence techniques. The primary heparin-binding site is composed of Lys60, Lys62, and Arg73, with additional contributions from the adjacent Arg76, Lys78, and N-terminal basic residues. The K(d) of binding is in the micromolar range. A heparin disaccharide analogue, sucrose octasulfate, binds with similar affinity to the N domain and to a naturally occurring HGF isoform, NK1, at nearly the same region as in heparin binding. (15)N relaxation data indicate structural flexibility on a microsecond-to-millisecond time scale around the primary binding site in the N domain. This flexibility appears to be dramatically reduced by ligand binding. On the basis of the NK1 crystal structure, we propose a model in which heparin binds to the two primary binding sites and the N-terminal regions of the N domains and stabilizes an NK1 dimer.     

Identification and characterization of a conformational heparin-binding site involving two fibronectin type III modules of bovine tenascin-X

Tenascin-X is known as a heparin-binding molecule, but the localization of the heparin-binding site has not been investigated until now. We show here that, unlike tenascin-C, the recombinant fibrinogen-like domain of tenascin-X is not involved in heparin binding. On the other hand, the two contiguous fibronectin type III repeats b10 and b11 have a predicted positive charge at physiological pH, hence a set of recombinant proteins comprising these domains was tested for interaction with heparin. Using solid phase assays and affinity chromatography, we found that interaction with heparin was conformational and involved both domains 10 and 11. Construction of a three-dimensional model of domains 10 and 11 led us to predict exposed residues that were then submitted to site-directed mutagenesis. In this way, we identified the basic residues within each domain that are crucial for this interaction. Blocking experiments using antibodies against domain 10 were performed to test the efficiency of this site within intact tenascin-X. Binding was significantly reduced, arguing for the activity of a heparin-binding site involving domains 10 and 11 in the whole molecule. Finally, the biological significance of this site was tested by cell adhesion studies. Heparan sulfate cell surface receptors are able to interact with proteins bearing domains 10 and 11, suggesting that tenascin-X may activate different signals to regulate cell behavior.     

PWWP module of human hepatoma-derived growth factor forms a domain-swapped dimer with much higher affinity for heparin

Hepatoma-derived growth factor (hHDGF)-related proteins (HRPs) comprise a new growth factor family sharing a highly conserved and ordered N-terminal PWWP module (residues 1-100, previously referred to as a HATH domain) and a variable disordered C-terminal domain. We have shown that the PWWP module is responsible for heparin binding and have solved its structure in solution. Here, we show that under physiological conditions, both the PWWP module and hHDGF can form dimers. Surface plasmon resonance (SPR) studies revealed that the PWWP dimer binds to heparin with affinity that is two orders of magnitude higher (K(d)=13 nM) than that of the monomeric PWWP module (K(d)=1.2 microM). The monomer-dimer equilibrium properties and NMR structural data together suggest that the PWWP dimer is formed through a domain-swapping mechanism. The domain-swapped PWWP dimer structures were calculated on the basis of the NMR data. The results show that the two PWWP protomers exchange their N-terminal hairpin to form a domain-swapped dimer. The two monomers in a dimer are linked by the long flexible L2 loops, a feature supported by NMR relaxation data for the monomer and dimer. The enhanced heparin-binding affinity of the dimer can be rationalized in the framework of the dimer structure.     

The crystal structure of the heparin-binding reelin-N domain of f-spondin

The extracellular matrix protein F-spondin mediates axon guidance during neuronal development. Its N-terminal domain, termed the reelin-N domain, is conserved in F-spondins, reelins, and other extracellular matrix proteins. In this study, a recombinant human reelin-N domain has been expressed, purified, and shown to bind heparin. The crystal structure of the reelin-N domain resolved to 2.0 Å reveals a variant immunoglobulin-like fold and potential heparin-binding sites. Substantial conformational variations even in secondary structure are observed between the two chemically identical reelin-N domains in one crystallographic asymmetric unit. The variations may result from extensive, highly specific interactions across the interface of the two reelin-N domains. The calculated values of buried surface area and the interface's shape complementarity are consistent with the formation of a weak dimer. The homophilic asymmetric dimer can potentially offer advantages in binding to ligands such as glycosaminoglycans, which may, in turn, bridge the two reelin-N domains and stabilize the dimer.     

Identification of the heparin-binding region of snake venom vascular endothelial growth factor (VEGF-F) and its blocking of VEGF-A165

VEGF-A165 displays multiple effects through binding to KDR (VEGFR-2). Heparin/heparan sulfate-like molecules are known to greatly modulate their interaction. In fact, VEGF-A lacking a C-terminal heparin-binding region exhibits significantly reduced mitogenic activity. We recently found novel heparin-binding VEGFs in snake venom, designated VEGF-Fs, which specifically recognize KDR, rather than other VEGF receptors. VEGF-Fs virtually lack the C-terminal heparin-binding region when compared with other heparin-binding VEGF subtypes, despite their heparin-binding potential. The C-terminal region does not exhibit any significant homology with other known proteins or domains. In this study, we attempted to identify the heparin-binding region of VEGF-F using synthetic peptides. The C-terminal peptide of vammin (one of the VEGF-Fs, 19 residues) bound to heparin with similar affinity as native vammin. We then evaluated the effects of this peptide on the biological activity of VEGF-A165. The C-terminal peptide of VEGF-F exhibited specific blockage of VEGF-A165 activity both in vitro and in vivo. These observations demonstrate that the short C-terminal region of VEGF-F functions fully as the active heparin-binding domain and the corresponding peptide specifically blocks VEGF-A165, thus suggesting that the C-terminal heparin-binding region of VEGF-F recognizes similar heparin/heparan sulfate molecules as VEGF-A165. The present results will provide novel insight into VEGF-heparin interaction and may facilitate the design of new anti-VEGF drugs based on novel strategies.     

Orientation of heparin-binding sites in native vitronectin. Analyses of ligand binding to the primary glycosaminoglycan-binding site indicate that putative secondary sites are not functional

A primary heparin-binding site in vitronectin has been localized to a cluster of cationic residues near the C terminus of the protein. More recently, secondary binding sites have been proposed. In order to investigate whether the binding site originally identified on vitronectin functions as an exclusive and independent heparin-binding domain, solution binding methods have been used in combination with NMR and recombinant approaches to evaluate ligand binding to the primary site. Evaluation of the ionic strength dependence of heparin binding to vitronectin according to classical linkage theory indicates that a single ionic bond is prominent. It had been previously shown that chemical modification of vitronectin using an arginine-reactive probe results in a significant reduction in heparin binding (Gibson, A., Baburaj, K., Day, D. E., Verhamme, I. , Shore, J. D., and Peterson, C. B. (1997) J. Biol. Chem. 272, 5112-5121). The label has now been localized to arginine residues within the cyanogen bromide fragment-(341-380) that contains the primary heparin-binding site on vitronectin. One- and two-dimensional NMR on model peptides based on this primary heparin-binding site indicate that an arginine residue participates in the ionic interaction and that other nonionic interactions may be involved in forming a complex with heparin. A recombinant polypeptide corresponding to the C-terminal 129 amino acids of vitronectin exhibits heparin-binding affinity that is comparable to that of full-length vitronectin and is equally effective at neutralizing heparin anticoagulant activity. Results from this broad experimental approach argue that the behavior of the primary site is sufficient to account for the heparin binding activity of vitronectin and support an exposed orientation for the site in the structure of the native protein.     

Identification and kinetics analysis of a novel heparin-binding site (KEDK) in human tenascin-C

The interaction between tenascin-C (TN-C), a multi-subunit extracellular matrix protein, and heparin was examined using a surface plasmon resonance-based technique on a Biacore system. The aims of the present study were to examine the affinity of fibronectin type III repeats of TN-C fragments (TNIII) for heparin, to investigate the role of the TNIII4 domains in the binding of TN-C to heparin, and to delineate a sequence of amino acids within the TNIII4 domain, which mediates cooperative heparin binding. At a physiological salt concentration, and pH 7.4, TNIII3-5 binds to heparin with high affinity (K(D) = 30 nm). However, a major heparin-binding site in TNIII5 produces a modest affinity binding at a K(D) near 4 microm, and a second site in TNIII4 enhances the binding by several orders of magnitude, although it was far too weak to produce an observable binding of TNIII4 by itself. Moreover, mutagenesis of the KEDK sequence in the TNIII4 domain resulted in the significant reduction of heparin-binding affinity. In addition, residues in the KEDK sequences are conserved in TN-C throughout mammalian evolution. Thus the structure-based sequence alignment, mutagenesis, and sequence conservation data together reveal a KEDK sequence in TNIII4 suggestive of a minor heparin-binding site. Finally, we demonstrate that TNIII4 contains binding sites for heparin sulfate proteoglycan and enhances the heparin sulfate proteoglycan-dependent human gingival fibroblast adhesion to TNIII5, thus providing the biological significance of heparin-binding site of TNIII4. These results suggest that the heparin-binding sites may traverse TNIII4-5 and thus require KEDK in TNIII4 for optimal heparin-binding.     

Processing in the C-terminal domain of minicollagen XII removes a heparin-binding site

A minicollagen comprising the two C-terminal domains of collagen XII (COL1 and NC1) has been expressed in insect cells and characterized biochemically. An interaction with heparin is demonstrated, which depends on the correct folding of the molecule. After secretion, minicollagen XII is immediately processed to a form lacking heparin binding ability. Processed and unprocessed trimers differ only at the level of the eight or nine C-terminal residues but they reveal different structures as judged from rotary shadowing images. Similar processing is also observed in the medium of transfected human HeLa cells. These data show that a heparin-binding site is present in the C-terminal end of the chicken collagen XII sequence and strongly suggest that proteolytic processing in the NC1 domain can occur in vivo and regulate the interactive properties of collagen XII.     

Mutations in the heparin-binding domains of human basic fibroblast growth factor alter its biological activity

Eleven structural analogues of human basic fibroblast growth factor (bFGF) have been prepared by site-directed mutagenesis of a synthetic bFGF gene to examine the effect of amino acid substitutions in the three putative heparin-binding domains on FGF's biological activity. After expression in Escherichia coli, the mutant proteins were purified to homogeneity by use of heparin-Sepharose chromatography and analyzed for their ability to stimulate DNA synthesis in human foreskin fibroblasts. Recombinant human bFGF 1-146 and [Ala69,Ser87]bFGF, an analogue where two of the four cysteines had been replaced by alanine and serine, were equipotent to standard bovine basic fibroblast growth factor. Substitution of aspartic acid-19 by arginine in the first heparin-binding domain yielded a molecule that stimulated a higher total mitogenic response in fibroblasts as compared to bFGF. In addition, replacement of either arginine-107 in the second domain or glutamine-123 in the third domain with glutamic acid resulted in compounds that were 2 and 4 times more potent than bFGF. In contrast, substitution of arginine-107 with isoleucine reduced the activity of the molecule by 100-fold. Combination of domain substitutions to generate the [Glu107,123]bFGF and [Arg19,Lys123,126]bFGF mutants did not show any additivity of the mutations on biological activity. Alterations in the biological activity of the analogues was dependent on both the site of and the type of modification. Increased positive charge in the first domain and increased negative charge in the second and third domains enhanced biological potency. The altered activities of the derivatives appear to be due in part to changes in the affinity of the analogues for heparin. We conclude that changes in all three of the putative heparin-binding domains result in altered mitogenic activity and heparin interaction of basic fibroblast growth factor.     

Molecular modeling of the dimeric structure of human lipoprotein lipase and functional studies of the carboxyl-terminal domain.

Lipoprotein lipase (LPL) plays a key role in lipid metabolism. Molecular modeling of dimeric LPL was carried out using insight ii based upon the crystal structures of human, porcine, and horse pancreatic lipase. The dimeric model reveals a saddle-shaped structure and the key heparin-binding residues in the amino-terminal domain located on the top of this saddle. The models of two dimeric conformations - a closed, inactive form and an open, active form - differ with respect to how surface-loop positions affect substrate access to the catalytic site. In the closed form, the surface loop covers the catalytic site, which becomes inaccessible to solvent. Large conformational changes in the open form, especially in the loop and carboxyl-terminal domain, allow substrate access to the active site. To dissect the structure-function relationships of the LPL carboxyl-terminal domain, several residues predicted by the model structure to be essential for the functions of heparin binding and substrate recognition were mutagenized. Arg405 plays an important role in heparin binding in the active dimer. Lys413/Lys414 or Lys414 regulates heparin affinity in both monomeric and dimeric forms. To evaluate the prediction that LPL forms a homodimer in a 'head-to-tail' orientation, two inactive LPL mutants - a catalytic site mutant (S132T) and a substrate-recognition mutant (W390A/W393A/W394A) - were cotransfected into COS7 cells. Lipase activity could be recovered only when heterodimerization occurred in a head-to-tail orientation. After cotransfection, 50% of the wild-type lipase activity was recovered, indicating that lipase activity is determined by the interaction between the catalytic site on one subunit and the substrate-recognition site on the other.     

Molecular modeling of the collagen-like tail of asymmetric acetylcholinesterase

The asymmetric form of acetylcholinesterase comprises three catalytic tetramers attached to ColQ, a collagen-like tail responsible for the anchorage of the enzyme to the synaptic basal lamina. ColQ is composed of an N-terminal domain which interacts with the catalytic subunits of the enzyme, a central collagen-like domain and a C-terminal globular domain. In particular, the collagen-like domain of ColQ contains two heparin-binding domains which interact with heparan sulfate proteoglycans in the basal lamina. A three-dimensional model of the collagen-like domain of the tail of asymmetric acetylcholinesterase was constructed. The model presents an undulated shape that results from the presence of a substitution and an insertion in the Gly-X-Y repeating pattern, as well as from low imino-acid regions. Moreover, this model permits the analysis of interactions between the heparin-binding domains of ColQ and heparin, and could also prove useful in the prediction of interaction domains with other putative basal lamina receptors.     

Localization of three distinct heparin-binding domains of laminin by monoclonal antibodies

Monoclonal antibodies were utilized to localize novel heparin-binding domains of laminin. A solid-phase radioligand binding assay was designed such that [3H] heparin bound to laminin in a time- and concentration-dependent manner. Tritiated heparin binding to laminin was saturable and specific as determined by competition with unlabeled heparin, dextran sulfate, and dermatan sulfate. By Scatchard analysis, two distinct dissociation constants were calculated (Kd = 50 and 130 nM), suggesting the presence of at least two binding sites for heparin on laminin. Tritiated heparin bound to thrombin-resistant (600 kDa) and chymotrypsin-resistant (440 kDa) laminin fragments, both known to lack the terminal globular domain of the long arm. Sodium dodecyl sulfate-polyacrylamide gels of chymotrypsin- and thermolysin-digested laminin chromatographed on a heparin-Sepharose column showed multiple proteolytic fragments binding to the column. Monoclonal antibodies generated against laminin were tested for their ability to inhibit [3H]heparin binding to laminin. Four monoclonal antibodies significantly inhibited the binding of [3H]heparin to laminin in the range of 15-21% inhibition. Laminin-monoclonal antibody interactions examined by electron microscopy showed that one antibody reacted at the terminal globular domain of the long arm, domain Hep-1, while epitopes for two of these monoclonal antibodies were located on the lateral arms of laminin, domain Hep-2, and the fourth monoclonal antibody bound below the cross-region of laminin, domain Hep-3. When two monoclonal antibodies recognizing distinctly different regions of laminin were added concomitantly, the inhibition of [3H]heparin binding to laminin increased almost 2-fold. These results suggest that at least two novel heparin-binding domains of laminin may be located in domains distinct from the terminal globular domain of the long arm.     

Two different heparin-binding domains in the triple-helical domain of ColQ,the collagen tail subunit of synaptic acetylcholinesterase

ColQ, the collagen tail subunit of asymmetric acetylcholinesterase, is responsible for anchoring the enzyme at the vertebrate synaptic basal lamina by interacting with heparan sulfate proteoglycans. To get insights about this function, the interaction of ColQ with heparin was analyzed. For this, heparin affinity chromatography of the complete oligomeric enzyme carrying different mutations in ColQ was performed. Results demonstrate that only the two predicted heparin-binding domains present in the collagen domain of ColQ are responsible for heparin interaction. Despite their similarity in basic charge distribution, each heparin-binding domain had different affinity for heparin. This difference is not solely determined by the number or nature of the basic residues conforming each site, but rather depends critically on local structural features of the triple helix, which can be influenced even by distant regions within ColQ. Thus, ColQ possesses two heparin-binding domains with different properties that may have non-redundant functions. We hypothesize that these binding sites coordinate acetylcholinesterase positioning within the organized architecture of the neuromuscular junction basal lamina.     

Heparin-binding growth-associated molecule contains two heparin-binding beta -sheet domains that are homologous to the thrombospondin type I repeat

Heparin-binding growth-associated molecule (HB-GAM) is an extracellular matrix-associated protein implicated in the development and plasticity of neuronal connections of brain. Binding to cell surface heparan sulfate is indispensable for the biological activity of HB-GAM. In the present paper we have studied the structure of recombinant HB-GAM using heteronuclear NMR. These studies show that HB-GAM contains two beta-sheet domains connected by a flexible linker. Both of these domains contain three antiparallel beta-strands. In addition to this domain structure, HB-GAM contains the N- and C-terminal lysine-rich sequences that lack a detectable structure and appear to form random coils. Studies using CD and NMR spectroscopy suggest that HB-GAM undergoes a conformational change upon binding to heparin, and that the binding occurs primarily to the beta-sheet domains of the protein. Search of sequence data bases shows that the beta-sheet domains of HB-GAM are homologous to the thrombospondin type I repeat (TSR). Sequence comparisions show that the beta-sheet structures found previously in midkine, a protein homologous with HB-GAM, also correspond to the TSR motif. We suggest that the TSR sequence motif found in various extracellular proteins defines a beta-sheet structure similar to that found in HB-GAM and midkine. In addition to the apparent structural similarity, a similarity in biological functions is suggested by the occurrence of the TSR sequence motif in a wide variety of proteins that mediate cell-to-extracellular matrix and cell-to-cell interactions, in which the TSR domain mediates specific cell surface binding.     

Monomeric midkine induces tumor cell proliferation in the absence of cell-surface proteoglycan binding

Midkine (MK), a retinoic acid-inducible heparin-binding protein, is a mitogen which initiates a cascade of intracellular protein tyrosine phosphorylation mediated by the JAK/STAT pathway after binding to its high affinity p200(+)/MKR cell surface receptor in the G401 cell line [Ratovitski, E. A. (1998) J. Biol. Chem. 273, 3654-3660]. In this study, we determined the biophysical characteristics of purified recombinant murine MK and analyzed the requirements for ligand multimerization and cell surface proteoglycan binding for the G401 cell mitogenic activity of MK. Our studies indicate that the secreted form of MK (M = 13 kDa) exists in solution as an asymmetric monomer with a frictional coefficient of 1. 48 and a Stokes radius of 23.7 A. By constructing bead models of MK using the program AtoB and the program HYDRO to predict the hydrodynamic properties of each model, our data suggest that MK has a dumb-bell shape in solution composed of independent N- and C-terminal domains separated by an extended linker. This asymmetric MK monomer is a biologically active ligand with mitogenic activity on G401 cells in vitro. Neither heparin-induced formation of noncovalent MK multimers nor tissue transglutaminase II covalent multimerization of MK enhanced MK mitogenic activity in this system. Since neither heparin competition nor cell treatment with chondroitinase ABC or heparinase III abolished the mitogenic effects of MK on G401 cells, cell-surface proteoglycan binding by MK does not appear to be a requirement for its observed mitogenic effects. These results provide strong evidence that the MK-specific p200(+)/MKR has distinctive biochemical properties which distinguish it from the receptor tyrosine phosphatase cell-surface proteoglycan PTPzeta/RPTPbeta and support the hypothesis that the diverse biological effects of MK are mediated by multiple cell-specific signal transduction receptors.     

Model for the three-dimensional structure of vitronectin: predictions for the multi-domain protein from threading and docking.

The structure of vitronectin, an adhesive protein that circulates in high concentrations in human plasma, was predicted through a combination of computational methods and experimental approaches. Fold recognition and sequence-structure alignment were performed using the threading program PROSPECT for each of three structural domains, i.e., the N-terminal somatomedin B domain (residues 1-53), the central region that folds into a four-bladed beta-propeller domain (residues 131-342), and the C-terminal heparin-binding domain (residues 347-459). The atomic structure of each domain was generated using MODELLER, based on the alignment obtained from threading. Docking experiments between the central and C-terminal domains were conducted using the program GRAMM, with limits on the degrees of freedom from a known inter-domain disulfide bridge. The docked structure has a large inter-domain contact surface and defines a putative heparin-binding groove at the inter-domain interface. We also docked heparin together with the combined structure of the central and C-terminal domains, using GRAMM. The predictions from the threading and docking experiments are consistent with experimental data on purified plasma vitronectin pertaining to protease sensitivity, ligand-binding sites, and buried cysteines.     

Multiple binding sites in fibronectin and the staphylococcal fibronectin receptor.

The binding of fibronectin to Staphylococci exhibits the properties of a ligand-receptor interaction and has been proposed to mediate bacterial adherence to host tissues. To localize staphylococcal-binding sites in fibronectin, the protein was subjected to limited proteolysis and, of the generated fragments, Staphylococci appeared to preferentially bind to the N-terminal fragment. Different fibronectin fragments were isolated and tested for their ability to inhibit 125I-fibronectin binding to Staphylococci. The results indicate that only the N-terminal region effectively competed for fibronectin binding. However, when isolated fragments were adsorbed to microtiter wells, we found that two distinct domains, corresponding to the N-terminal fragment and to the heparin-binding peptide mapping close to the C-terminal end of fibronectin, promoted the attachment of both Staphylococcus aureus Newman and coagulase-negative strain of Staphylococcus capitis 651. These same domains were recognized by purified 125I-labeled staphylococcal receptor, either when immobilized on microtiter wells or probed after adsorption onto nitrocellulose membrane. The heparin-binding domain is comprised of type-III-homology repeats 14, 15 and 16. To determine which repeats participate in this interaction, we isolated and tested repeats type III14 and type III16. We found that the major staphylococcal binding site is located in repeat type III14. The staphylococcal receptor bound the N-terminal domain of fibronectin with a KD of 1.8 nM, whereas the dissociation constant of the receptor molecule for the internal heparin-binding domain was 10 nM. Since the fusion protein ZZ-FR, which contains the active sequences of fibronectin receptor (D1-D3) bound only to the N-terminus, it is reasonable to assume that the bacterial receptor may have additional binding sites outside the D domains, capable of interacting with the internal heparin-binding domain of fibronectin.     

Evidence that binding to the carboxyl-terminal heparin-binding domain (Hep II) dominates the interaction between plasma fibronectin and heparin

We assessed the participation of the three known heparin-binding domains of PFn (Hep I, Hep II, Hep III) in their interaction with heparin by making a quantitative comparison of the fluid-phase heparin affinities of PFn and PFn fragments under physiologic pH and ionic strength conditions. Using a fluorescence polarization binding assay that employed a PFn affinity-purified fluorescein-labeled heparin preparation, we found that greater than 98% of the total PFn heparin-binding sites exhibit a Kd in the 118-217 nM range. We also identified a minor (less than 2%) class of binding sites exhibiting very high affinity (Kd approximately 1 nM) in PFn and the carboxyl-terminal 190/170 and 150/136 kDa PFn fragments. This latter activity probably reflects multivalent inter- or intramolecular heparin-binding activity. Amino-terminal PFn fragments containing Hep I (72 and 29 kDa) exhibited low affinity for heparin under physiologic buffer conditions (Kd approximately 30,000 mM). PFn fragments (190/170 and 150/136 kDa) containing both the carboxyl-terminal Hep II and central Hep III domains retained most of the heparin-binding activity of native PFn (Kd = 278-492 nM). The isolated Hep II domain (33-kDa fragment) exhibited appreciable, but somewhat lower (2-5-fold), heparin affinity compared to the 190/170-kDa PFn fragment. Heparin binding to the 100-kDa PFn fragment containing Hep III was barely detectable (Kd greater than 30,000 nM). From these observations, we conclude that PFn contains only one major functional heparin-binding site per subunit, Hep II, that dominates the interaction between heparin and PFn.