Expression of ECM proteins fibulin-1 and -2 in acute and chronic liver disease and in cultured rat liver cells.

Fibulin-2 has previously been considered as a marker to distinguish rat liver myofibroblasts from hepatic stellate cells. The function of other fibulins in acute or chronic liver damage has not yet been investigated. The aim of this study has been to evaluate the expression of fibulin-1 and -2 in models of rat liver injury and in human liver cirrhosis. Their cellular sources have also been investigated. In normal rat liver, fibulin-1 and -2 were both mainly present in the portal field. Fibulin-1-coding transcripts were detected in total RNA of normal rat liver, whereas fibulin-2 mRNA was only detected by sensitive, real-time quantitative polymerase chain reaction. In acute liver injury, the expression of fibulin-1 was significantly increased (17.23-fold after 48 h), whereas that of fibulin-2 was not modified. The expression of both fibulin-1 and -2 was increased in experimental rat liver cirrhosis (19.16- and 26.47-fold, respectively). At the cellular level, fibulin-1 was detectable in hepatocytes, "activated" hepatic stellate cells, and liver myofibroblasts (2.71-, 122.65-, and 469.48-fold over the expression in normal rat liver), whereas fibulin-2 was restricted to liver myofibroblasts and was regulated by transforming growth factor beta-1 (TGF-beta1) in 2-day-old hepatocyte cultures and in liver myofibroblasts. Thus, fibulin-1 and -2 respond differentially to single and repeated damaging noxae, and their expression is differently present in liver cells. Expression of the fibulin-2 gene is regulated by TGF-beta1 in liver myofibroblasts.

Compositional differences between infant and adult human corneal basement membranes.

PURPOSE: Adult human corneal epithelial basement membrane (EBM) and Descemet's membrane (DM) components exhibit heterogeneous distribution. The purpose of the study was to identify changes of these components during postnatal corneal development. METHODS: Thirty healthy adult corneas and 10 corneas from 12-day- to 3-year-old children were studied by immunofluorescence with antibodies against BM components. RESULTS: Type IV collagen composition of infant corneal central EBM over Bowman's layer changed from alpha1-alpha2 to alpha3-alpha4 chains after 3 years of life; in the adult, alpha1-alpha2 chains were retained only in the limbal BM. Laminin alpha2 and beta2 chains were present in the adult limbal BM where epithelial stem cells are located. By 3 years of age, beta2 chain appeared in the limbal BM. In all corneas, limbal BM contained laminin gamma3 chain. In the infant DM, type IV collagen alpha1-alpha6 chains, perlecan, nidogen-1, nidogen-2, and netrin-4 were found on both faces, but they remained only on the endothelial face of the adult DM. The stromal face of the infant but not the adult DM was positive for tenascin-C, fibrillin-1, SPARC, and laminin-332. Type VIII collagen shifted from the endothelial face of infant DM to its stromal face in the adult. Matrilin-4 largely disappeared after the age of 3 years. CONCLUSIONS: The distribution of laminin gamma3 chain, nidogen-2, netrin-4, matrilin-2, and matrilin-4 is described in the cornea for the first time. The observed differences between adult and infant corneal BMs may relate to changes in their mechanical strength, corneal cell adhesion and differentiation in the process of postnatal corneal maturation.

Gene structure and functional analysis of the mouse nidogen-2 gene: nidogen-2 is not essential for basement membrane formation in mice.

Nidogens are highly conserved proteins in vertebrates and invertebrates and are found in almost all basement membranes. According to the classical hypothesis of basement membrane organization, nidogens connect the laminin and collagen IV networks, so stabilizing the basement membrane, and integrate other proteins. In mammals two nidogen proteins, nidogen-1 and nidogen-2, have been discovered. Nidogen-2 is typically enriched in endothelial basement membranes, whereas nidogen-1 shows broader localization in most basement membranes. Surprisingly, analysis of nidogen-1 gene knockout mice presented evidence that nidogen-1 is not essential for basement membrane formation and may be compensated for by nidogen-2. In order to assess the structure and in vivo function of the nidogen-2 gene in mice, we cloned the gene and determined its structure and chromosomal location. Next we analyzed mice carrying an insertional mutation in the nidogen-2 gene that was generated by the secretory gene trap approach. Our molecular and biochemical characterization identified the mutation as a phenotypic null allele. Nidogen-2-deficient mice show no overt abnormalities and are fertile, and basement membranes appear normal by ultrastructural analysis and immunostaining. Nidogen-2 deficiency does not lead to hemorrhages in mice as one may have expected. Our results show that nidogen-2 is not essential for basement membrane formation or maintenance.

Opposing roles of integrin alpha6Abeta1 and dystroglycan in laminin-mediated extracellular signal-regulated kinase activation.

Laminin-integrin interactions can in some settings activate the extracellular signal-regulated kinases (ERKs) but the control mechanisms are poorly understood. Herein, we studied ERK activation in response to two laminins isoforms (-1 and -10/11) in two epithelial cell lines. Both cell lines expressed beta1-containing integrins and dystroglycan but lacked integrin alpha6beta4. Antibody perturbation assays showed that both cell lines bound to laminin-10/11 via the alpha3beta1and alpha6beta1 integrins. Although laminin-10/11 was a stronger adhesion complex than laminin-1 for both cell lines, both laminins activated ERK in only one of the two cell lines. The ERK activation was mediated by integrin alpha6beta1 and not by alpha3beta1 or dystroglycan. Instead, we found that dystroglycan-binding domains of both laminin-1 and -10/11 suppressed integrin alpha6beta1-mediated ERK activation. Moreover, the responding cell line expressed the two integrin alpha6 splice variants, alpha6A and alpha6B, whereas the nonresponding cell line expressed only alpha6B. Furthermore, ERK activation was seen in cells transfected with the integrin alpha6A subunit, but not in alpha6B-transfected cells. We conclude that laminin-1 and -10/11 share the ability to induce ERK activation, that this is regulated by integrin alpha6Abeta1, and suggest a novel role for dystroglycan-binding laminin domains as suppressors of this activation.

Endostatin regulates endothelial cell adhesion and cytoskeletal organization.

Endostatin, an endogenous angiogenesis inhibitor, attenuates endothelial cell migration through an unknown mechanism. We show that endostatin induced tyrosine phosphorylation of focal adhesion kinase and paxillin, and promoted formation of focal adhesions and actin stress fibers, similar to fibroblast growth factor-2 (FGF-2). In cells cotreated with endostatin and FGF-2, focal adhesions and actin stress fibers were decreased, indicating that endostatin disturbs cell-matrix adhesion. Reduced tyrosine phosphorylation and cytoplasmic relocalization of beta-catenin in cells treated with FGF-2 and endostatin indicates that loosening of cell-cell adhesion is also disturbed by endostatin. These data provide a molecular basis both for the lack of effect of endostatin on the normal, quiescent vasculature, and its antagonistic effects on stimulated tumor vessels.

The minimal active domain of endostatin is a heparin-binding motif that mediates inhibition of tumor vascularization.

Endostatin constitutes the COOH-terminal 20,000 Da proteolytic fragment of collagen XVIII and has been shown to possess antiangiogenic and antitumorigenic properties. In the present study, we have investigated the role of the heparin-binding sites in the in vivo mechanism of action of endostatin. The majority of the heparin binding is mediated by arginines 155/158/184/270 in endostatin, but there is also a minor site constituted by arginines 193/194. Using endostatin mutants lacking either of these two sites, we show that inhibition of fibroblast growth factor-2-induced angiogenesis in the chicken chorioallantoic membrane requires both heparin-binding sites. In contrast, inhibition of vascular endothelial growth factor-A-induced chorioallantoic membrane angiogenesis by endostatin was only dependent on the minor heparin-binding site (R193/194). These arginines were also required for endostatin to inhibit fibroblast growth factor-2- and vascular endothelial growth factor-A-induced chemotaxis of primary endothelial cells. Moreover, we show that a synthetic peptide corresponding to amino acids 180-199 of human endostatin (which covers the minor heparin-binding site) inhibits endothelial cell chemotaxis and reduces tumor vascularization in vivo. Substitution of arginine residues 193/194 for alanine attenuates the antiangiogenic effects of the peptide. These data show an essential role for heparin binding in the antiangiogenic action of endostatin.

Disruption of perlecan binding and matrix assembly by post-translational or genetic disruption of dystroglycan function.

Dystroglycan is a cell-surface matrix receptor that requires LARGE-dependent glycosylation for laminin binding. Although the interaction of dystroglycan with laminin has been well characterized, less is known about the role of dystroglycan glycosylation in the binding and assembly of perlecan. We report reduced perlecan-binding activity and mislocalization of perlecan in the LARGE-deficient Large(myd) mouse. Cell-surface ligand clustering assays show that laminin polymerization promotes perlecan assembly. Solid-phase binding assays provide evidence for the first time of a trimolecular complex formation of dystroglycan, laminin and perlecan. These data suggest functional disruption of the trimolecular complex in glycosylation-deficient muscular dystrophy.

Distinct requirements for heparin and alpha-dystroglycan binding revealed by structure-based mutagenesis of the laminin alpha2 LG4-LG5 domain pair.

Laminin-2 (alpha2beta1gamma1) is found in basement membranes surrounding muscle and peripheral nerve cells. Several types of cellular receptors bind to the laminin G-like (LG) domains at the C terminus of the alpha2 chain, the interaction with alpha-dystroglycan (alpha-DG) being particularly important in muscle. We have used site-directed mutagenesis and in vitro binding assays to map the binding sites on the laminin alpha2 chain LG4-LG5 domain pair for alpha-DG, heparin and sulfatides. Calcium-dependent alpha-DG recognition requires the calcium ion in LG4, but not the one in LG5, as well as basic residues in both LG domains. Heparin and sulfatides also bind to basic residues in both LG domains, but there is little overlap in the binding sites for alpha-DG and heparin/sulfatides. The results should prove useful for the molecular dissection of laminin-receptor interactions in vivo.

Homoassociation of VE-cadherin follows a mechanism common to "classical" cadherins.

Vascular endothelial cadherin (VE-cadherin/cadherin5) is specifically expressed in adherens junctions of endothelial cells and exerts important functions in cell-cell adhesion as well as signal transduction. To analyze the mechanism of VE-cadherin homoassociation, the ectodomains CAD1-5 were connected by linker sequences to the N terminus of the coiled-coil domain of cartilage matrix protein (CMP). The chimera VECADCMP were expressed in mammalian cells. The trimeric coiled-coil domain leads to high intrinsic domain concentrations and multivalency promoting self-association. Ca(2+)-dependent homophilic association of VECADCMP was detected in solid phase assays and cross-linking experiments. A striking analogy to homoassociation of type I ("classical") cadherins like E, N or P-cadherin was observed when interactions in VECADCMP and between these trimeric proteins were analyzed by electron microscopy. Ca(2+)-dependent ring-like and double ring-like arrangements suggest interactions between domains 1 and 2 of the ectodomains, which may be correlated with lateral and adhesive contacts in the adhesion process. Association to complexes composed of two VECADCMP molecules was also demonstrated by chemical cross-linking. No indication for an antiparallel association of VECAD ectodomains to hexameric complexes as proposed by Legrand et al. was found. Instead the data suggest that homoassociation of VE-cadherin follows the conserved mechanism of type I cadherins.

Chondrogenic activity of the heparan sulfate proteoglycan perlecan maps to the N-terminal domain I.

C3H10T1/2 cells differentiate along a chondrogenic pathway when plated onto the extracellular matrix (ECM) protein perlecan (Pln). To identify the region(s) within the large Pln molecule that provides a differentiation signal, recombinant Pln-sequence-based polypeptides representing distinct structural domains were assayed for their ability to promote chondrogenesis in C3H10T1/2 cells. Five distinct domains, along with structural variations, were tested. The N-terminal domain I was tested in two forms (IA and IB) that contain only heparan sulfate (HS) chains or both HS and chondroitin sulfate (CS) chains, respectively. A mutant form of domain I lacking attachment sites for both HS and CS (Pln I(mut)) was tested also. Other constructs consecutively designated Pln domains II, III(A-C), IV(A,B), and V(A,B) were used to complete the structure-function analysis. Cells plated onto Pln IA or Pln IB but no other domain rapidly assembled into cellular aggregates of 40-120 microm on average. Aggregate formation was dependent on the presence of glycosaminoglycan (GAG) chains, because Pln I-based polypeptides lacking GAG chains either by enzymatic removal or mutation of HS/CS attachment sites were inactive. Aggregates formed on GAG-bearing Pln IA stained with Alcian Blue and were recognized by antibodies to collagen type II and aggrecan but were not recognized by an antibody to collagen type X, a marker of chondrocyte hypertrophy. Collectively, these studies indicate that the GAG-bearing domain I of Pln provides a sufficient signal to trigger C3H10T1/2 cells to enter a chondrogenic differentiation pathway. Thus, this matrix proteoglycan (PG) found at sites of cartilage formation in vivo is likely to enhance early stage differentiation induced by soluble chondrogenic factors.

Targeted inactivation of murine laminin gamma2-chain gene recapitulates human junctional epidermolysis bullosa.

Junctional forms of epidermolysis bullosa (JEB) are associated with mutations in six distinct genes expressed in the cutaneous basement membrane zone; these include LAMA3, LAMB3, and LAMC2, which encode laminin 5 subunit polypeptides, the alpha3-, beta3-, and gamma2-chains, respectively. Here we generated a mouse model for JEB by inactivating the laminin gamma2-chain gene by targeted frameshift deletion of exon 8 in Lamc2. Heterozygous mice were phenotypically normal, whereas the majority of Lamc2-/- mice showed blistering phenotype on days 1 to 2 and died within 5 days of birth. The Lamc2-/- mice demonstrated absent expression of laminin gamma2-chain on the basement membrane zone as well as attenuated expression of alpha3- and beta3-chains of laminin. Transmission electron microscopy revealed rudimentary, poorly developed hemidesmosomes. The epidermis of the Lamc2-/- mice revealed induced apoptosis in the basal cells of the blistered skin, suggesting that cell-matrix adhesion provided by laminin 5 plays a role in cell survival in vivo. Cultured Lamc2-/- keratinocytes demonstrated slightly positive staining with gamma2-chain-specific antibodies, which could be explained by the presence of a transcript with partial restoration of the reading frame owing to alternative splicing in vitro. These cells proliferated in different matrices and attached to type IV collagen and Matrigel as efficiently as the wild-type keratinocytes, whereas their attachment on plastic and laminin was significantly weaker. In summary, Lamc2-/- mouse recapitulates human JEB and provides novel insight into the role of laminin 5 in keratinocyte biology.

Evidence of nidogen-2 compensation for nidogen-1 deficiency in transgenic mice.

Previous studies have shown that inhibition of nidogen-laminin binding interferes with basement membrane stabilization in various mouse organ cultures while no overt phenotype has been observed following inactivation of the nidogen-1 gene in mice. We have now used recombinant mouse nidogen-1 and nidogen-2 in order to evaluate a possible compensation between the two isoforms in the knock-out mice. Essentially, a comparable in vitro binding of nidogens-1 and -2 to the same laminin gamma1 chain structure and to several other basement membrane proteins has been revealed. Quantitative radioimmuno-assays have demonstrated high concentrations of nidogen-1 exceeding those of laminin gamma1 and nidogen-2 by factors of 5 and 20-50, respectively, in tissue extracts of wild-type mice. A three- to sevenfold increase in nidogen-2 was observed in heart and muscle of mice with nidogen-1 deficiency and confirmed by a similar increase in the intensity of immunogold staining of these tissues. However, a few of the tissues from mice with the gene knock-out still contained some nidogen-1-like immunoreactivity (1% of wild-type). Furthermore, both nidogen isoforms showed a similar distribution in various organs during embryonic development which, however, as shown previously, changed in some adult tissues. The data support the nidogen-2 compensation hypothesis to explain the limited phenotype observed following elimination of the nidogen-1 gene.

Endostatin reduces vascularization, blood flow, and growth in a rat gliosarcoma.

Endostatin, the 20-kDa C-terminal fragment of collagen XVIII, has previously been shown to inhibit growth and induce regression of different experimental tumors in rodents. In this study, we show that recombinant murine and human endostatin, produced in 293 EBNA cells and yeast, respectively, inhibit ectotopic as well as orthotopic growing BT4Cn gliosarcomas in BD-IX rats. In rats in which s.c. gliomas were grown for a total of 29 days, systemic treatment with recombinant murine endostatin induced about 50% reduction of intratumoral blood flow and tumor size after only 10 days of therapy. In contrast, the blood flow to irrelevant organs was unaffected by endostatin, indicating its specificity of action. Tumors were not observed to increase in size or regrow after cessation of therapy. Furthermore, endostatin-treated rats with i.c. tumors had significantly longer survival time than did untreated controls. In the treated rats, endostatin therapy resulted in a reduced tumor blood vessel volume and an increased tumor cell density with an increased apoptotic index within a given tumor volume, as verified by flow cytometry and by staining with deoxynucleotidyltransferase-mediated dUTP nick-end labeling. This work verifies the general anti-angiogenic and antitumor effects of endostatin and indicates that the protein may also be considered as a treatment strategy for malignant brain tumors.

Laminin-121--recombinant expression and interactions with integrins.

Laminin-121, previously referred as to laminin-3, was expressed recombinantly in human embryonic kidney (HEK) 293 cells by triple transfection of full-length cDNAs encoding mouse laminin α1, ß2 and γ1 chains. The recombinant laminin-121 was purified using Heparin-Sepharose followed by molecular sieve chromatography and shown to be correctly folded by electron microscopy and circular dichroism (CD). The CD spectra of recombinant laminin-121 were very similar to those of laminin-111 isolated from Engelbreth-Holm-Swarm tumor (EHS-laminin) but its T(m) value was smaller than EHS-laminin and recombinant lamnin-111 suggesting that the replacement of the ß chain reduced the stability of the coiled-coil structure of laminin-121. Its binding to integrins was compared with EHS-laminin, laminin-3A32 purified from murine epidermal cell line and recombinantly expressed laminins-111, -211 and -221. Laminin-121 showed the highest affinity to α6ß1 and α7ß1 integrins and furthermore, laminin-121 most effectively supported neurite outgrowth. Together, this suggests that the ß2 laminins have higher affinity for integrins than the ß1 laminins.

Mapping of SPARC/BM-40/osteonectin-binding sites on fibrillar collagens.

The 33-kDa matrix protein SPARC (BM-40, osteonectin) binds several collagen types with moderate affinity. The collagen-binding site resides in helix alphaA of the extracellular calcium-binding domain of SPARC and is partially masked by helix alphaC. Previously, we found that the removal of helix alphaC caused a 10-fold increase in the affinity of SPARC for collagen, and we identified amino acids crucial for binding by site-directed mutagenesis. In this study, we used rotary shadowing, CNBr peptides, and synthetic peptides to map binding sites of SPARC onto collagens I, II, and III. Rotary shadowing and electron microscopy of SPARC-collagen complexes identified a major binding site approximately 180 nm from the C terminus of collagen. SPARC binding was also detected with lower frequency near the matrix metalloproteinase cleavage site. These data fit well with our analysis of SPARC binding to CNBr peptides, denaturation of which abolished binding, indicating triple-helical conformation of collagen to be essential. SPARC binding was substantially decreased in two of seven alpha2(I) mutant procollagen I samples and after N-acetylation of Lys/Hyl side chains in wild-type collagen. Synthetic peptides of collagen III were used to locate the binding sites, and we found SPARC binding activity in a synthetic triple-helical peptide containing the sequence GPOGPSGPRGQOGVMGFOGPKGNDGAO (where O indicates 4-hydroxyproline), with affinity for SPARC comparable with that of procollagen III. This sequence is conserved among alpha chains of collagens I, II, III, and V. In vitro collagen fibrillogenesis was delayed in the presence of SPARC, suggesting that SPARC might modulate collagen fibril assembly in vivo.

A comparative analysis of the fibulin protein family. Biochemical characterization, binding interactions, and tissue localization.

Fibulins are a family of five extracellular matrix proteins characterized by tandem arrays of epidermal growth factor-like domains and a C-terminal fibulin-type module. They are widely distributed and often associated with vasculature and elastic tissues. In this study, we expressed the three more recently identified family members, fibulin-3, fibulin-4, and fibulin-5, as recombinant proteins in mammalian cells. The purified proteins showed short rod structures of approximately 20 nm with a globule at one end, after rotary shadowing and electron microscopy. Two forms of mouse fibulin-3 were purified, and the O-glycan profiles of the larger form were characterized. Polyclonal antibodies raised against the purified proteins did not show any cross-reactivity with other family members and were used to assess the levels and localization of the fibulins in mouse tissues. Their binding interactions, cell adhesive properties, and tissue localization were analyzed in parallel with the previously characterized fibulin-1 and -2. Binding to tropoelastin was strong for fibulin-2 and -5, moderate for fibulin-4 and -1, and relatively weak for fibulin-3. Fibulin-4, but not fibulin-3 and -5, exhibited distinct interactions with collagen IV and nidogen-2 and moderate binding to the endostatin domain from collagen XV. Cell adhesive activities were not observed for all fibulins, except mouse fibulin-2, with various cell lines tested. All five fibulins were found in perichondrium and various regions of the lungs. Immunoelectron microscopy localized fibulin-4 and -5 to fibrillin microfibrils at distinct locations. Our studies suggest there are unique and redundant functions shared by these structurally related proteins.

Structural basis for Gas6-Axl signalling.

Receptor tyrosine kinases of the Axl family are activated by the vitamin K-dependent protein Gas6. Axl signalling plays important roles in cancer, spermatogenesis, immunity, and platelet function. The crystal structure at 3.3 A resolution of a minimal human Gas6/Axl complex reveals an assembly of 2:2 stoichiometry, in which the two immunoglobulin-like domains of the Axl ectodomain are crosslinked by the first laminin G-like domain of Gas6, with no direct Axl/Axl or Gas6/Gas6 contacts. There are two distinct Gas6/Axl contacts of very different size, both featuring interactions between edge beta-strands. Structure-based mutagenesis, protein binding assays and receptor activation experiments demonstrate that both the major and minor Gas6 binding sites are required for productive transmembrane signalling. Gas6-mediated Axl dimerisation is likely to occur in two steps, with a high-affinity 1:1 Gas6/Axl complex forming first. Only the minor Gas6 binding site is highly conserved in the other Axl family receptors, Sky/Tyro3 and Mer. Specificity at the major contact is suggested to result from the segregation of charged and apolar residues to opposite faces of the newly formed beta-sheet.

WARP is a novel multimeric component of the chondrocyte pericellular matrix that interacts with perlecan.

WARP is a novel member of the von Willebrand factor A domain superfamily of extracellular matrix proteins that is expressed by chondrocytes. WARP is restricted to the presumptive articular cartilage zone prior to joint cavitation and to the articular cartilage and fibrocartilaginous elements in the joint, spine, and sternum during mouse embryonic development. In mature articular cartilage, WARP is highly specific for the chondrocyte pericellular microenvironment and co-localizes with perlecan, a prominent component of the chondrocyte pericellular region. WARP is present in the guanidine-soluble fraction of cartilage matrix extracts as a disulfide-bonded multimer, indicating that WARP is a strongly interacting component of the cartilage matrix. To investigate how WARP is integrated with the pericellular environment, we studied WARP binding to mouse perlecan using solid phase and surface plasmon resonance analysis. WARP interacts with domain III-2 of the perlecan core protein and the heparan sulfate chains of the perlecan domain I with K(D) values in the low nanomolar range. We conclude that WARP forms macromolecular structures that interact with perlecan to contribute to the assembly and/or maintenance of "permanent" cartilage structures during development and in mature cartilages.

Endostatin influences endothelial morphology via the activated ERK1/2-kinase endothelial morphology and signal transduction.

Endostatin, the proteolytic fragment of collagen XVIII, is known to be a potent inhibitor of angiogenesis. However, to date, only limited knowledge exists with regard to the effects of endostatin on vessel morphology and the underlying signaling pathway. The aim of the present work was therefore to determine the impact of endostatin and its collagen XV analogue restin on vessel development during wound healing and embryonic angio- and vasculogenesis. Time lapse experiments and electron microscopy demonstrate similar morphological changes evoked by endostatin and the ERK1/2-kinase inhibitor PD98059. Furthermore, we show that ERK1/2 phosphorylation, a crucial signaling event in vascular morphogenesis, is regulated by endostatin via the protein phosphatase 2A PP2A. These findings provide new insight into a key signaling pathway of vascular remodeling evoked by a matrix-derived factor.

Chondroitin sulfate perlecan enhances collagen fibril formation. Implications for perlecan chondrodysplasias.

Inactivation of the perlecan gene leads to perinatal lethal chondrodysplasia. The similarity to the phenotypes of the Col2A1 knock-out and the disproportionate micromelia mutation suggests perlecan involvement in cartilage collagen matrix assembly. We now present a mechanism for the defect in collagen type II fibril assembly by perlecan-null chondrocytes. Cartilage perlecan is a heparin sulfate or a mixed heparan sulfate/chondroitin sulfate proteoglycan. The latter form binds collagen and accelerates fibril formation in vitro, with more defined fibril morphology and increased fibril diameters produced in the presence of perlecan. Interestingly, the enhancement of collagen fibril formation is independent on the core protein and is mimicked by chondroitin sulfate E but neither by chondroitin sulfate D nor dextran sulfate. Furthermore, perlecan chondroitin sulfate contains the 4,6-disulfated disaccharides typical for chondroitin sulfate E. Indeed, purified glycosaminoglycans from perlecan-enriched fractions of cartilage extracts contain elevated levels of 4,6-disulfated chondroitin sulfate disaccharides and enhance collagen fibril formation. The effect on collagen assembly is proportional to the content of the 4,6-disulfated disaccharide in the different cartilage extracts, with growth plate cartilage glycosaminoglycan being the most efficient enhancer. These findings demonstrate a role for perlecan chondroitin sulfate side chains in cartilage extracellular matrix assembly and provide an explanation for the perlecan-null chondrodysplasia.