Molecular and structural mapping of collagen fibril interactions.

The fibrous collagens form the structural basis of all mammalian connective tissues, including the vasculature, dermis, bones, tendons, cartilage, and those tissues that support organs such as the heart, kidneys, liver, and lungs. The helical structure of collagen has been extensively studied but in addition to its helical character, its molecular packing arrangement (in its aggregated or fibrillar form) and the presence of specific amino acid sequences govern collagen's in vivo functions. Collagen's molecular packing arrangement helps control cellular communication, attachment and movement, and conveys its tissue-specific biomechanical properties. Recent progress in understanding collagen's molecular packing, fibrillar structure, domain organization, and extracellular matrix (ECM) interactions in light of X-ray fiber diffraction data provides significant new insights into how the ECM is organized and functions. In this review, the hierarchy of fibrillar collagen structure is discussed in the context of how this organization affects ECM-"ligand" interactions, with specific attention to collagenolysis, integrins, fibronection, glycoprotein VI receptor (GPVI), and proteoglycans (PG). Understanding the complex structure of collagen and its attached ligands should provide new insights into tissue growth, development, regeneration, and disease.

Collagen fibril surface displays a constellation of sites capable of promoting fibril assembly, stability, and hemostasis.

Fibrillar collagens form the structural basis of organs and tissues including the vasculature, bone, and tendon. They are also dynamic, organizational scaffolds that present binding and recognition sites for ligands, cells, and platelets. We interpret recently published X-ray diffraction findings and use atomic force microscopy data to illustrate the significance of new insights into the functional organization of the collagen fibril. These data indicate that collagen's most crucial functional domains localize primarily to the overlap region, comprising a constellation of sites we call the "master control region." Moreover, the collagen's most exposed aspect contains its most stable part-the C-terminal region that controls collagen assembly, cross-linking, and blood clotting. Hidden beneath the fibril surface exists a constellation of "cryptic" sequences poised to promote hemostasis and cell-collagen interactions in tissue injury and regeneration. These findings begin to address several important, and previously unresolved, questions: How functional domains are organized in the fibril, which domains are accessible, and which require proteolysis or structural trauma to become exposed? Here we speculate as to how collagen fibrillar organization impacts molecular processes relating to tissue growth, development, and repair.

Type I collagen and collagen mimetics as angiogenesis promoting superpolymers.

Angiogenesis, the development of blood vessels from the pre-existing vasculature, is a key component of embryogenesis and tissue regeneration. Angiogenesis also drives pathologies such as tumor growth and metastasis, and hemangioma development in newborns. On the other hand, promotion of angiogenesis is needed in tissues with vascular insufficiencies, and in bioengineering, to endow tissue substitutes with appropriate microvasculatures. Therefore, much research has focused on defining mechanisms of angiogenesis, and identifying pro- and anti-angiogenic molecules. Type I collagen, the most abundant protein in humans, potently stimulates angiogenesis in vitro and in vivo. Crucial to its angiogenic activity appears to be ligation and possibly clustering of endothelial cell (EC) surface alpha 1 beta 1/alpha 2 beta 1 integrin receptors by the GFPGER(502-507) sequence of the collagen fibril. However, additional aspects of collagen structure and function that may modulate its angiogenic properties are discussed. Moreover, type I collagen and fibrin, another angiogenic polymer, share several structural features. These observations suggest strategies for creating "angiogenic superpolymers", including: modifying type I collagen to influence its biological half-life, immunogenicity, and integrin binding capacity; genetically engineering fibrillar collagens to include additional integrin binding sites or angiogenic determinants, and remove unnecessary or deleterious sequences without compromising fibril integrity; and exploring the suitability of poly(ortho ester), PEG-lysine copolymer, tubulin, and cholesteric cuticle as collagen mimetics, and suggesting means of modifying them to display ideal angiogenic properties. The collagenous and collagen mimetic angiogenic superpolymers described here may someday prove useful for many applications in tissue engineering and human medicine.

Insulin-like growth factor (IGF) binding protein-3 regulation of IGF-I is altered in an acidic extracellular environment.

While extracellular acidification within solid tumors is well-documented, how reduced pH impacts regulation of insulin-like growth factor-I (IGF-I) has not been studied extensively. Because IGF-I receptor binding is affected by IGF binding proteins (IGFBPs), we examined how pH impacted IGFBP-3 regulation of IGF-I. IGF-I binding in the absence of IGFBP-3 was diminished at reduced pH. Addition of IGFBP-3 reduced IGF-I cell binding at pH 7.4 but increased surface association at pH 5.8. This increase in IGF-I binding at pH 5.8 corresponded with an increase in IGFBP-3 cell association. This, however, was not due to an increase in affinity of IGFBP-3 for heparin at reduced pH although both heparinase III treatment and heparin addition reduced IGFBP-3 enhancement of IGF-I binding. An increase in IGF-I binding to IGFBP-3, though, was seen at reduced pH using a cell-free assay. We hypothesize that the enhanced binding of IGF-I at pH 5.8 is facilitated by increased association of IGFBP-3 at this pH and that the resulting cell associated IGF-I is IGFBP-3 and not IGF-IR bound. Increased internalization and nuclear association of IGF-I at pH 5.8 in the presence of IGFBP-3 was evident, yet cell proliferation was reduced by IGFBP-3 at both pH 5.8 and 7.4 indicating that IGFBP-3-cell associated IGF-I does not signal the cell to proliferate and that the resulting transfer of bound IGF-I from IGF-IR to IGFBP-3 results in diminished proliferation. Solution binding of IGF-I by IGFBP-3 is one means by which IGF-I-induced proliferation is inhibited. Our work suggests that an alternative pathway exists by which IGF-I and IGFBP-3 both associate with the cell surface and that this association inhibits IGF-I-induced proliferation.

Novel design of peptides to reverse the anticoagulant activities of heparin and other glycosaminoglycans.

Patients undergoing anticoagulation with unfractionated heparin, low molecular weight heparin, or danaparoid may experience excess bleeding which requires reversal of the anticoagulant agent. Protamine is at present the only agent available for reversal of unfractionated heparin. Protamine is not effective in patients who have received low molecular weight heparin or danaparoid. We have developed a series of peptides based on consensus heparin binding sequences (Verrecchio et al., J Biol Chem 2000; 275: 7701-7707) that are capable of neutralizing the anti-thrombin activity of unfractionated heparin in vitro, the antifactor Xa activity of unfractionated heparin, Enoxaparin (Lovenox) and danaparoid (Orgaran) in vitro and the anti-Factor Xa activity of Enoxaparin in vivo in rats. These peptides may serve as alternatives for Protamine reversal of UFH and may be useful for neutralization of enoxaparin and danaparoid in humans.

Synthesis, secretion, and subcellular localization of serglycin proteoglycan in human endothelial cells.

The serglycin proteoglycan is best known as a hematopoietic cell granule proteoglycan. It has been found that serglycin is synthesized by endothelial cells, is localized to cytoplasmic vesicles, and is constitutively secreted. Serglycin messenger RNA in human umbilical vein endothelial cells (HUVECs) and cultured human aortic endothelial cells was detected by reverse transcription-polymerase chain reaction. (35)S-sulfate-labeled secreted and intracellular proteoglycans were analyzed. It was found that 85% of the proteoglycans synthesized during culture were secreted. A core protein of the appropriate size for serglycin was detected by analysis of the chondroitinase-digested (35)S-sulfate-labeled HUVEC proteoglycans. This was the major core protein of the secreted chondroitin sulfate proteoglycans. Recombinant serglycin core protein was used to generate an antibody in chickens. A core protein identified by Western blotting of chondroitinase digests of HUVEC proteoglycans corresponded to the major (35)S-sulfate- labeled core protein. Identical results were obtained with 2 hematopoietic cell lines. Cyto-immunofluorescence showed cytoplasmic vesicular and perinuclear labeling in hematopoietic cells and HUVECs. The serglycin-containing vesicles in HUVECs are distinct from the Weibel-Palade bodies, which contain von Willebrand factor. Confocal microscopy showed that tissue plasminogen activator was distributed similarly to serglycin. Serglycin may be important for the function of these vesicles and, once secreted, for the modulation of the activity of their constituents.

Myofibroblast involvement in glycosaminoglycan synthesis and lipid retention during coronary repair.

Myofibroblasts of adventitial origin have been linked to neointimal formation and remodeling after coronary injury. Accordingly, the goal of this study was to examine whether myofibroblasts contribute to focal accumulation of glycosaminoglycans (GAGs) and lipids during coronary repair. GAG synthesis was assessed by ex vivo labeling of balloon-injured porcine coronary arteries with (14)C-glucosamine. The synthesis of total GAGs transiently increased at 8 days in the normolipemic model (a 2.2-fold increase over baseline, p < 0.05). The majority of newly synthesized GAGs were sensitive to chondroitin ABC lyase (chondroitin/dermatan sulfate GAGs). Versican was localized to myofibroblast-rich regions in the adventitia and neointima [positive for alpha-smooth muscle (SM) actin, negative for h-caldesmon and SM myosin heavy chain]. In contrast, the adjacent SM-rich media showed no increase in versican expression. The association between injury-induced GAG accumulation and lipid retention was examined at 2 weeks after coronary injury in the hyperlipemic model. Lipid (Oil Red O) accumulated in the neointima and adventitia, but not in the adjacent media. Coronary repair under hyperlipemic conditions was associated with macrophage infiltration (19 +/- 5 vs. 3 +/- 2% of neointimal cells in normolipemic animals, p < 0.001) and increased neointimal formation (1.8 +/- 0.5 vs. 1.0 +/- 0.3 mm(2) in normolipemic animals, p < 0.01). In conclusion, this study demonstrated a transient increase in GAG synthesis following coronary injury. Chondroitin sulfate proteoglycans (e.g., versican) were rapidly synthesized by activated adventitial and neointimal cells which could contribute to early lipid retention in injured vessels.

Decorin binds near the C terminus of type I collagen.

Decorin belongs to a family of small leucine-rich proteoglycans that are directly involved in the control of matrix organization and cell growth. Genetic evidence indicates that decorin is required for the proper assembly of collagenous matrices. Here, we sought to establish the precise binding site of decorin on type I collagen. Using rotary shadowing electron microscopy and photoaffinity labeling, we mapped the binding site of decorin protein core to a narrow region near the C terminus of type I collagen. This region is located within the cyanogen bromide peptide fragment alpha1(I) CB6 and is approximately 25 nm from the C terminus, in a zone that coincides with the c(1) band of the collagen fibril d-period. This location is very close to one of the major intermolecular cross-linking sites of collagen heterotrimers. Thus, decorin protein core possesses a unique binding specificity that could potentially regulate collagen fibril stability.

Altered dermatan sulfate structure and reduced heparin cofactor II-stimulating activity of biglycan and decorin from human atherosclerotic plaque.

Biglycan and decorin are small dermatan sulfate-containing proteoglycans in the extracellular matrix of the artery wall. The dermatan sulfate chains are known to stimulate thrombin inhibition by heparin cofactor II (HCII), a plasma proteinase inhibitor that has been detected within the artery wall. The purpose of this study was to analyze the HCII-stimulatory activity of biglycan and decorin isolated from normal human aorta and atherosclerotic lesions type II through VI and to correlate activity with dermatan sulfate chain composition and structure. Biglycan and decorin from plaque exhibited a 24-75% and 38-79% loss of activity, respectively, in thrombin-HCII inhibition assays relative to proteoglycan from normal aorta. A significant negative linear relationship was observed between lesion severity and HCII stimulatory activity (r = 0.79, biglycan; r = 0.63, decorin; p < 0.05). Biglycan, but not decorin, from atherosclerotic plaque contained significantly reduced amounts of iduronic acid and disulfated disaccharides DeltaDi-2,4S and DeltaDi-4,6S relative to proteoglycan from normal artery. Affinity coelectrophoresis analysis of a subset of samples demonstrated that increased interaction of proteoglycan with HCII in agarose gels paralleled increased activity in thrombin-HCII inhibition assays. In conclusion, both biglycan and decorin from atherosclerotic plaque possessed reduced activity with HCII, but only biglycan demonstrated a correlation between activity and specific glycosaminoglycan structural features. Loss of the ability of biglycan and decorin in atherosclerotic lesions to regulate thrombin activity through HCII may be critical in the progression of the disease.

Design of peptides with high affinities for heparin and endothelial cell proteoglycans.

Proteoglycan-binding peptides were designed based on consensus sequences in heparin-binding proteins: XBBXBX and XBBBXXBX, where X and B are hydropathic and basic residues, respectively. Initial peptide constructs included (AKKARA)(n) and (ARKKAAKA)(n) (n = 1-6). Affinity coelectrophoresis revealed that low M(r) peptides (600-1,300) had no affinities for low M(r) heparin, but higher M(r) peptides (2,000-3,500) exhibited significant affinities (K(d) congruent with 50-150 nM), which increased with peptide M(r). Affinity was strongest when sequence arrays were contiguous and alanines and arginines occupied hydropathic and basic positions, but inclusion of prolines was disruptive. A peptide including a single consensus sequence of the serglycin proteoglycan core protein bound heparin strongly (K(d) congruent with 200 nM), likely owing to dimerization through cysteine-cysteine linkages. Circular dichroism showed that high affinity heparin-binding peptides converted from a charged coil to an alpha-helix upon heparin addition, whereas weak heparin-binding peptides did not. Higher M(r) peptides exhibited high affinities for total endothelial cell proteoglycans (K(d) congruent with 300 nM), and approximately 4-fold weaker affinities for their free glycosaminoglycan chains. Thus, peptides including concatamers of heparin-binding consensus sequences may exhibit strong affinities for heparin and proteoglycans. Such peptides may be applicable in promoting cell-substratum adhesion or in the design of drugs targeted to proteoglycan-containing cell surfaces and extracellular matrices.

Heparin sensitive and resistant vascular smooth muscle cells: biology and role in restenosis.

Vascular smooth muscle cells (VSMC)s are characterized by their acute growth inhibition by heparin and heparan sulfates; however, recently the isolation of VSMCs which display greatly diminished sensitivity to the antiproliferative action of heparin have been reported. These heparin resistant (HR) VSMCs have been derived through multiple passage of normal rat VSMCs in culture media containing high heparin doses, by transformation of VSMCs with oncogene-containing vectors, or have been isolated from vascular tissues of spontaneously hypertensive rats, healthy humans, or humans with restenosis where their presence is not limited to sites of injury. Initial characterizations of HR VSMCs are reviewed, and here we propose a definition of HR VSMCs. To date the mechanisms underlying heparin insensitivity remain elusive. Further study of HR VSMCs may expand our understanding of cell growth regulation by heparin, establish whether HR VSMCs contribute to the reported failure of heparin to combat restenosis in humans, and identify cellular mechanisms driving certain vascular proliferative diseases.

Defining the domains of type I collagen involved in heparin- binding and endothelial tube formation.

Cell surface heparan sulfate proteoglycan (HSPG) interactions with type I collagen may be a ubiquitous cell adhesion mechanism. However, the HSPG binding sites on type I collagen are unknown. Previously we mapped heparin binding to the vicinity of the type I collagen N terminus by electron microscopy. The present study has identified type I collagen sequences used for heparin binding and endothelial cell-collagen interactions. Using affinity coelectrophoresis, we found heparin to bind as follows: to type I collagen with high affinity (Kd approximately 150 nM); triple-helical peptides (THPs) including the basic N-terminal sequence alpha1(I)87-92, KGHRGF, with intermediate affinities (Kd approximately 2 microM); and THPs including other collagenous sequences, or single-stranded sequences, negligibly (Kd >> 10 microM). Thus, heparin-type I collagen binding likely relies on an N-terminal basic triple-helical domain represented once within each monomer, and at multiple sites within fibrils. We next defined the features of type I collagen necessary for angiogenesis in a system in which type I collagen and heparin rapidly induce endothelial tube formation in vitro. When peptides, denatured or monomeric type I collagen, or type V collagen was substituted for type I collagen, no tubes formed. However, when peptides and type I collagen were tested together, only the most heparin-avid THPs inhibited tube formation, likely by influencing cell interactions with collagen-heparin complexes. Thus, induction of endothelial tube morphogenesis by type I collagen may depend upon its triple-helical and fibrillar conformations and on the N-terminal heparin-binding site identified here.

Decreased serglycin proteoglycan size is associated with the platelet alpha granule storage defect in Wistar Furth hereditary macrothrombocytopenic rats. Serglycin binding affinity to type I collagen is unaltered.

The Wistar Furth (WF) rat has a hereditary defect in platelet formation that resembles gray platelet syndrome of man with a large mean platelet volume and platelet alpha granule deficiency. The alpha granule abnormality is suggestive of a defect in granule packaging and/or stability. Proteoglycans are hypothesized to play a role in granule packaging. Therefore, we have analyzed the structure of the platelet proteoglycan, serglycin, in platelets of WF and normal Wistar rats. Normal and Wistar Furth rats were injected with 35S-sulfate to label platelet proteoglycans via synthesis by the megakaryocytes, and platelets were isolated 3 days later. We found that WF rat platelets have only one-third of the normal proteoglycan mass per unit platelet volume, and the proteoglycans are smaller in hydrodynamic size with shorter glycosaminoglycan chains than those of Wistar rats. However, WF rat platelet proteoglycans showed no defect in binding to collagen on affinity coelectrophoresis gels. We conclude that the structure of WF rat platelet proteoglycans is abnormal, and speculate that this abnormality may contribute to abnormal packaging of the alpha granule contents. Leakage of alpha granule contents into the marrow by platelets and megakaryocytes could perturb the marrow matrix, and promote the development of myelofibrosis noted in gray platelet syndrome.

Chondrogenic potential of chick embryonic calvaria: II. Matrix calcium may repress cartilage differentiation.

Chick embryos cultured in the absence of their eggshell are rendered severely calcium-deficient, and develop a cartilage-like phenotype in the calvarium, a normally osteogenic tissue. In the preceding paper (Jacenko and Tuan [1995] Dev. Dyn. 202:13-26), experiments using organ cultured calvaria from day-12 normal and shell-less embryos showed that depletion of calcium alone may be responsible in promoting chondrogenic differentiation in calvaria. Here these findings were extended using an in vivo calvarial grafting technique, such that the extent of calvarial matrix calcification was a function of the calcium status of both the graft and the host. In these calvarial grafts, undermineralized regions again were shown to support chondrogenesis. To identify possible mechanisms which promote chondrogenesis in the calvaria, cells were enzymatically dissociated from the calvaria and cultured in media with varied levels of soluble calcium, under conditions which should modulate cell-to-cell interactions, including monolayer, micromass, agarose gels, and suspension cultures. Soluble calcium had no effect on calvarial cell differentiation, whereas conditions which enhanced cell-cell interactions, e.g., suspension culture, elicited cartilage expression. Based on these findings, we propose that the calcified matrix of the calvarium is repressive to chondrogenesis during normal development, but that the lack of mineral in a calcium-deficient calvarium creates a microenvironment permissive for cell-to-cell interactions which lead to chondrogenic differentiation.

Interactions of syndecan-1 and heparin with human collagens.

Glycosaminoglycan (GAG)-collagen interactions play important roles in cell adhesion and extracellular matrix assembly; however, the chemical bases for these interactions are not fully understood. We have used affinity co-electrophoresis (ACE) (Lee, M.K. and Lander, A.D., Proc. Natl. Acad. Sci, USA, 88, 2768-2772, 1991) to study the binding of the heparan sulphate proteoglycan syndecan-1 and heparin to human collagens. [35S]Syndecan-1 [from normal murine mammary gland (NMuMG) epithelial cells] and low-M(r) (approximately 6 kDa) [125I]heparin were subjected to electrophoresis through agarose gel lanes containing human collagens at various concentrations, and binding affinities were measured from shifts in migration of the labelled materials. Results demonstrate that the affinities of each collagen for syndecan-1 and low-M(r) heparin were similar, and followed the order: type V >> type IV approximately type III approximately type I > type VI >> type II, and ranged in Kd from approximately 10(-8) to approximately 3 x 10(-6) M. These data suggest that syndecan-1 and heparin may contain similar collagen-binding determinants. It was also found that the same heparin subpopulation was selectively bound with high affinity by each of the collagens. The published amino acid sequences of the six collagens were examined for what are thought to be heparin-binding consensus sequences (Cardin, A.D. and Weintraub, H.J.R., Arteriosclerosis, 9, 21-32, 1989). The presence of such sequences did not correlate with affinity for heparin or syndecan-1, and collagens I, II and III lacked such sequences entirely. The data suggest that collagens may use novel types of binding sites to interact with GAGs.

Mapping the heparin-binding sites on type I collagen monomers and fibrils.

The glycosaminoglycan chains of cell surface heparan sulfate proteoglycans are believed to regulate cell adhesion, proliferation, and extracellular matrix assembly, through their interactions with heparin-binding proteins (for review see Ruoslahti, E. 1988. Annu. Rev. Cell Biol. 4:229-255; and Bernfield, M., R. Kokenyesi, M. Kato, M. T. Hinkes, J. Spring, R. L. Gallo, and E. J. Lose. 1992. Annu. Rev. Cell Biol. 8:365-393). Heparin-binding sites on many extracellular matrix proteins have been described; however, the heparin-binding site on type I collagen, a ubiquitous heparin-binding protein of the extracellular matrix, remains undescribed. Here we used heparin, a structural and functional analogue of heparan sulfate, as a probe to study the nature of the heparan sulfate proteoglycan-binding site on type I collagen. We used affinity coelectrophoresis to study the binding of heparin to various forms of type I collagen, and electron microscopy to visualize the site(s) of interaction of heparin with type I collagen monomers and fibrils. Using affinity coelectrophoresis it was found that heparin has similar affinities for both procollagen and collagen fibrils (Kd's approximately 60-80 nM), suggesting that functionally similar heparin-binding sites exist in type I collagen independent of its aggregation state. Complexes of heparin-albumin-gold particles and procollagen were visualized by rotary shadowing and electron microscopy, and a preferred site of heparin binding was observed near the NH2 terminus of procollagen. Native or reconstituted type I collagen fibrils showed one region of significant heparin-gold binding within each 67-nm period, present near the division between the overlap and gap zones, within the "a" bands region. According to an accepted model of collagen fibril structure, our data are consistent with the presence of a single preferred heparin-binding site near the NH2 terminus of the collagen monomer. Correlating these data with known type I collagen sequences, we suggest that the heparin-binding site in type I collagen may consist of a highly basic triple helical domain, including several amino acids known sometimes to function as disaccharide acceptor sites. We propose that the heparin-binding site of type I collagen may play a key role in cell adhesion and migration within connective tissues, or in the cell-directed assembly or restructuring of the collagenous extracellular matrix.

Cisplatin-induced peptic ulcers, vagotomy, adrenal and calcium modulation.

Cytochemical and autoradiographic studies in Wistar rats [Crl:(WI)BR] show that cisplatin treatment (9 mg/kg) inhibits the release of acetylcholine from the axonal endings of the stomach smooth muscle resulting in bloating of the stomach and ulceration. Cisplatin also induces corticosteroid release from the adrenal gland stimulating peptic ulceration. Vagotomy helps ameliorate the effect but not eliminate it. Calcium supplementation restores normal neuromuscular function to gastric smooth muscle, thereby eliminating the gastro-intestinal toxicity due to cisplatin.

Specificity in the interactions of extracellular matrix proteins with subpopulations of the glycosaminoglycan heparin.

Many extracellular matrix glycoproteins--including laminin, fibronectin, thrombospondin, type I collagen, and other collagens--bind the glycosaminoglycan heparin, yet little is known about the functional significance of these interactions. It is also not known if heparin-binding extracellular matrix proteins recognize distinct structural elements in heparin, nor whether all extracellular matrix proteins recognize the same or different aspects of heparin structure. If extracellular matrix proteins each recognize distinct features of heparin, such specificity could be of importance in vivo, where structurally distinct heparan sulfate species occur. To investigate specificity in the binding between extracellular matrix proteins and heparin, the method of affinity coelectrophoresis (ACE) was used [Lee, M. K., & Lander, A. D. (1991) Proc. Natl. Acad. Sci. U.S.A. 88, 2768-2772]. Low M(r) (approximately 6 kDa) 125I-heparin was fractionated by electrophoresis through agarose gel lanes containing extracellular matrix proteins at various concentrations; from heparin migration patterns, binding affinities were calculated. The results indicate that fibronectin, type I collagen, and laminin--but not thrombospondin--each fractionate heparin into subpopulations that differ substantially in binding affinity. From ACE gels containing either fibronectin, type I collagen, or laminin, fractions of heparin were isolated that represent the 25% of molecules most strongly bound and the 25% least strongly bound by each of these proteins. Subsequent ACE analysis of these six fractions showed that (1) for each of fibronectin, type I collagen, and laminin, strongly- and weakly-binding heparin subfractions differ approximately 5-30-fold in Kd; (2) heparin that binds strongly to any one of fibronectin, type I collagen, or laminin also binds strongly to the other two; (3) heparin that binds weakly to any one of fibronectin, type I collagen, or laminin, also binds weakly to the other two; (4) heparin subfractions that differ greatly in affinity for fibronectin, type I collagen, and laminin show little difference in Kd for thrombospondin or for the heparin-binding growth factor basic fibroblast growth factor (bFGF); (5) neither heterogeneity in molecular charge [as measured by diethylaminoethyl (DEAE) chromatography] nor size nor the presence or absence of antithrombin III recognition sequences can account for the selective binding of heparin subpopulations to fibronectin, type I collagen, and laminin. These results suggest that structural elements within heparin can confer preferential binding to extracellular matrix proteins. Sensitivity of some, but not all, extracellular matrix proteins to these structural features suggests that similar features, if present in heparan sulfates or other glycosaminoglycans, may be physiologically relevant in vivo.