| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 280, Issue 34, 30526-30537, August 26, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶¶**
From the
Wellcome Trust Centre for Cell-Matrix Research and ¶UK Centre for Tissue Engineering, Faculty of Life Sciences, University of Manchester, Manchester M13 9PT, United Kingdom, Cancer Research UK and University of Manchester, Department of Medical Oncology, Christie Hospital NHS Trust, Wilmslow Road, Manchester M20 4BX, United Kingdom, and ||Molecular Biotechnology, School of Molecular and Microbial Biosciences G08, University of Sydney, Sydney 2006, Australia
Received for publication, February 7, 2005 , and in revised form, May 17, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Elastin is synthesized as the soluble precursor tropoelastin (2). It has a molecular mass of 
70 kDa, and its alternating hydrophobic and cross-linking domains are critical for assembly and cross-linking by lysyl oxidases. Fibrillins, the principal structural molecules of microfibrils, are large glycoproteins (350 kDa) with multiple calcium-binding epidermal growth factor (EGF)1-like domain arrays interspersed with eight-cysteine (TB) motifs (3). Furin processing of the C-terminal region of fibrillin-1 may be necessary for extracellular fibrillin-1 deposition (4). Genetic linkage of fibrillin-1 to Marfan syndrome (5) and its abundance in developing and adult tissues (6) confirm that it is a major fibrillin isoform in elastic fibers. Several other proteins, including emilin-1 and fibulins-1 and -2, are present within the elastin core or at the interface between the elastin core and microfibrils (1). Fibulin-5, essential for elastic fiber formation, is present at the elastic fiber-cell interface (7, 8). Molecules that co-localize with microfibrils include microfibril-associated glycoproteins (MAGPs)-1 and -2 and LTBPs-1, -2, and -4 (1). Various proteoglycans and glycosaminoglycans (heparan sulfate, decorin, biglycan, and versican) may associate with microfibrils and elastic fibers (9-12).
In affinity chromatography studies, heparin has been shown to bind fibrillin-1 at the furin-processed C terminus (13). Heparin and heparan sulfate interacted in a calcium-independent manner to a large fibrillin-1 N-terminal fragment (Arg45-Thr450) and to the C-terminal half of fibrillin-1 (Asp1528-Arg2731), and in a calcium-dependent manner to a large central (Asp1028-Thr1486) fragment of fibrillin-1 (9). These heparan sulfate-binding sites had a selective specificity for sulfated L-iduronate-rich heparan sulfate. Addition of heparin or heparan sulfate to skin fibroblast cultures and treatment with 
-D-xylosides, which inhibit the attachment of glycosaminoglycans to core proteins, inhibited fibrillin-1 microfibril deposition (9). Sulfated heparan sulfate is also necessary for tropoelastin incorporation into the extracellular matrix (14, 15).
Heparan sulfate is an anionic linear polymer composed of 50-200 disaccharide units covalently attached to specific proteins, forming proteoglycans at the cell surface (syndecans and glypicans) and in the extracellular matrix (e.g. perlecan, agrin, and collagen XVIII) (16). Heparan sulfate chains are synthesized by enzymes that initially polymerize a chain of repeating GlcA and GlcNAc units, which are then variously modified by N- and O-sulfation and uronate epimerization (17, 18). These modifications occur in clusters, creating sulfation domains (S-domains) that form the principal recognition site for heparan sulfate-binding proteins (17, 18). Flanking the S-domains are regions of alternating N-sulfated and N-acetylated disaccharides, termed "transition zones," the presence of which distinguishes HS from heparin that is highly modified along its entire length. Interactions between proteins and S-domains can be dependent on a critical sequence of sulfate-modified sugars, e.g. fibroblast growth factor-2 (19). In other cases, saccharide requirements for ligand binding reflect the length and preference for C-6 sulfation over C-2 or N-sulfation. The C-terminal endostatin fragment of collagen XVIII has been proposed to bind to S-domains interrupted by one GlcNAc residue (20, 21) and with specific requirements for sulfation at C-6 of amino sugars (22). For other heparan sulfate-binding proteins, especially multimeric cytokines such as interleukin-8, the region of heparan sulfate that interacts with the ligand is more extended, encompassing typically two S-domains, including the intervening regions of lower sulfation (23).
|
In this study, we have used BIAcore technology to investigate fibrillin-1 interactions with glycosaminoglycans. We have mapped four high affinity heparan sulfate-binding sites on fibrillin-1 and defined their kinetics, established the binding efficacies of heparin fragments of defined lengths, and showed how heparin binding to fibrillin-1 affects interactions with tropoelastin and MAGP-1 but not the homotypic fibrillin-1 interactions. We have also demonstrated that hyaluronan and chondroitin sulfate do not interact significantly with fibrillin-1. These data contribute significantly to our understanding of how heparin, and by analogy its chemical analogue heparan sulfate, regulates microfibril and elastic fiber assembly.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Heparin and Heparin Saccharides—The heparin used was the sodium salt from porcine intestinal mucosa, with a molecular mass of 3000 Da (Sigma, catalog number H3400). The heparin saccharides were not obtained from a commercial source and were prepared by digestion of heparin with bacterial heparinase, as reported previously (35). Three defined sized heparin saccharides were used, dp12, dp16, and dp24, which contain 12, 16, and 24 monosaccharides, respectively. These fragments are chemical analogues of the sulfated domains (S-domains) of heparan sulfate; the domain structure of heparan sulfate is shown in Fig. 2. The chondroitin 6-sulfate sodium salt used was from shark cartilage (Fluka, UK, catalog number 27045). The biotinylated hyaluronan was purchased from Sigma (catalog number B1557).
Biotinylation of Heparin—Two methods of biotinylation were used in order to confirm the heparin saccharide interactions with fibrillin. The first approach was via oxidized cis-diol groups (Fig. 2) (36). Chondroitin sulfate, heparin, and heparin species dp12, dp16, and dp24, in 0.1 mM solutions, were made up in 0.1 M NaOAc, pH 5.5 (coupling buffer). To a volume of 100 µl was added the same volume of 20 mM sodium meta-periodate in coupling buffer. After 30 min on ice in the dark, 20 µl of 150 mM glycerol was added to stop the oxidation, for 5 min on ice in the dark. The samples were dialyzed into coupling buffer before the addition of biotin-LC-hydrazide (Pierce) at a final concentration of 0.5 mM. After incubation for 2 h at room temperature, the nonreacted biotin was removed by extensive dialysis into coupling buffer.
The second biotinylation approach involved coupling via the heparin reducing end in two stages (36, 37). First, reductive amination with ammonia and then biotin was coupled to the free amines (Fig. 2). Heparin species dp12, dp16, and dp24, (0.1 mg) were dissolved in 2 M NH4Cl, in a volume of 100 µl. 2 mg of NaCNBH3 was added, and the mixture was heated at 70 °C for 2 days. After cooling, the mixture was dialyzed extensively into Dulbecco's phosphate-buffered saline (Cambrex, UK). 10 µl of 3 mg/ml sulfo-NHS-LC-biotin (Pierce) was added and incubated overnight at 4 °C. The nonreacted biotin was then removed by further dialysis into 0.1 M NaOAc, pH 5.5.
BIAcore 3000 Binding Analysis—For kinetic binding studies of heparin, heparin oligosaccharides, chondroitin sulfate, and hyaluronan with fibrillin-1 by surface plasmon resonance, a BIAcore biosensor was used (BIAcore 3000, BIAcore AB, Sweden). Biotinylated heparin, heparin fragments, chondroitin sulfate, and hyaluronan oligosaccharides were immobilized onto commercially prepared SA sensor chips, which have preimmobilized streptavidin, to allow biotin capture. By using heparin concentrations of 1 µM, typically 150-200 RU of biotinylated heparin samples were immobilized, which was at a saturation level. Samples were applied to the sensor chip surface in 0.1 M NaOAc, pH 5.5. All subsequent binding experiments were performed in 10 mM HEPES, pH 7.4, 0.1 M NaCl, 1 mM CaCl2, and 0.005% surfactant P20 (designated HBS-Ca).
Heparin, heparin saccharides, hyaluronan, and chondroitin sulfate binding analyses were performed using a panel of fibrillin-1 protein fragments encompassing full-length fibrillin-1, as analytes, at 100 nM and a flow rate of 30 µl/min for 3 min. A high flow rate of 30 µl/min was used to minimize mass transport effects. After 2.5 min dissociation, the chip was regenerated using 50 mM NaOH and 1 M NaCl for 1 min. The chip was then stabilized for 20 min using HBS-Ca, before the next injection was carried out. The analyte was simultaneously passed over a blank flow cell, and this base line was subtracted from the experimental flow cell. None of the analytes was found to interact with the blank flow cell; however, this control allowed bulk shifts caused by subtle differences in solvents to be eliminated. The maximum relative response value for each injection was calculated using the binding assay result wizard (BIAcore control software 3.2).
|
2 values. 2 values are a standard statistical measure of the closeness of fit (mean square of the signal noise). Two alternative methods were used for PF1 binding to heparin. PF1 binding to heparin was calculated independently by using equilibrium analysis. The injection time was increased to 15 min and the dissociation time to 30 min. The equilibrium response was plotted against concentration, and nonlinear regression using the equation for one site binding was used to calculate KD values. To calculate association and dissociation constants for PF1, the observed rate constant kobs was calculated from the association phase using the 1:1 Langmuir association (kobs) model (BIAevaluation 4.1, BIAcore AB). A secondary plot of kobs against concentration (C) allows ka and kd to be calculated using linear regression, using the equation kobs = kaC + kd.
The binding of PF1 mutants R62C, T101A, and V449I was also analyzed against immobilized dp24, using the same methods. To investigate the kinetics of interactions between the PF1 mutants and MAGP-1, MAGP-1 was immobilized on the surface of a CM5 sensor chip via amine coupling, as described previously (29). PF1 and PF1 mutants R62C, T101A, and V449I were then injected at concentration ranges of 0.25-5 µg/ml at a flow rate 30 µl/min. Kinetic analysis was also performed on immobilized hyaluronan using the versican G1 domain (kindly provided by Dr. A. Day, MRC Immunochemistry Unit, Oxford, UK). Versican was injected at concentrations ranging from 1 to 5 µg/ml at a flow rate 30 µl/min. For a potential positive control for chondroitin 6-sulfate, binding of fibroblast growth factor-1 and -2 was tested, but no significant binding was detected. However, chondroitin ABC lyase lowered the response level of the chondroitin sulfate immobilized flow cell, indicating cleavage may have taken place.
Inhibition Assays Using BIAcore—To investigate the effects of MAGP-1 on binding of PF1 to dp24, kinetic analysis was performed using 1 µg/ml PF1 preincubated with 0.4-1200 nM MAGP-1 in HBS-Ca. For MAGP-1 only, 0.4-1200 nM was separately passed over the sensor chip, as a control. These responses were then subtracted from the corresponding PF1/MAGP-1 data. Analytes were injected over the sensor chip in the same manner as described for the kinetic studies. An inhibition curve was then plotted using the response value of each normalized curve, at the end of the association period. To investigate the effects of dp24 on the interaction of PF1 to MAGP-1, MAGP-1 was immobilized, using amine coupling, as described previously (29). Kinetic analysis was carried out using 1 and 5 µg/ml of PF1 preincubated with increasing concentrations of un-biotinylated dp24 0.4-4000 nM; dp24 0.4-4000 nM was separately passed over the sensor chip and then subtracted from the PF1/dp24 plots. An inhibition curve was plotted, as described above.
To investigate the effects of dp24 on the interaction of PF7 to tropoelastin, tropoelastin was immobilized, using amine coupling as described previously (29). Kinetic analysis was carried out using 1 and 5 µg/ml of PF7 preincubated with increasing concentrations of un-biotinylated dp24 0.4-4000 nM or with dp24 only (0.4-4000 nM). An inhibition curve was plotted, as described above.
Calcium dependence for fibrillin-1/heparin interactions was determined by kinetic analysis, monitoring immobilized biotinylated dp24 binding to 5 µg/ml PF1, PF7, or PF11, in the presence of increasing concentrations of EDTA (0.01-10 mM). EDTA only sensorgrams were also plotted and subtracted from the FBN-1/EDTA sensorgrams. An inhibition curve was plotted, as described above.
Inhibition Using Solid Phase Binding Assays—Competition binding assays were also conducted using both nonbiotinylated and biotinylated soluble proteins; biotinylation was carried out as described previously (29). Flat-bottomed microtiter plates were coated with PF1 and PF13, at 100 nM in 50 mM Tris, pH 7.4, 0.1 M NaCl, 1 mM CaCl2 (TBSa/CaCl2), overnight at 4 °C. Nonspecific binding sites were then blocked with TBSa/CaCl2 containing 4% BSA at room temperature for at least 2 h. The plates were washed three times with TBSa, CaCl2, 0.1% BSA and incubated with either 100 nM of biotinylated PF1 or PF13, which had been preincubated with heparin oligosaccharide dp24 (4-12000 nM) in TBSa/CaCl2, overnight at 4 °C. After a further three washes, plates were incubated with 1:200 dilution of extravidin peroxidase conjugate at room temperature for 15 min. Bound protein was quantified after four more washes by a colorimetric assay using 40 mM 2,2'-azino-bis(3-ethyl-benzthiazoline-6-sulfonic acid) solution (Sigma) for 10-20 min at room temperature. Plates were read at a wavelength of 405. Control wells with the first nonbiotinylated soluble ligand not added were incubated in TBSa/CaCl2 only. For both methods, wells were then washed four times, and the color was developed. Any nonspecific fibrillin-1 binding was detected by blocking wells before incubation with BSA only. All assays were performed in triplicate and repeated at least twice to confirm observed results. We used biotinylation as it is a widely used and sensitive molecular detection system.
|
Modeling of Fibrillin-1 TB5/Heparin Interaction—The cbEGF-TB4-cbEGF structure was taken from the Protein Data Bank (code 1UZJ [PDB] ) (38); the file was edited to remove the first cbEGF domain, and the remaining TB4-cbEGF domains were used as the negative (non-heparan sulfate binding) control. The TB4-cbEGF structure was also used as a starting model to generate a TB5-cbEGF homology model. Modeling was performed in Quanta2000 (Accelrys) on a Silicon Graphics O2 R12,000 using the protein modeling modules. Partial charges were assigned in the ADT program (AutoDock tools) (39).
The structures of two heparin pentasaccharides that differ in the conformation of the iduronate residues (1C4 or 2S0) were generated from the Protein Data Bank (code 1HPN [PDB] ) (40). Partial charges were assigned to atoms according to the PIM force field (41) and allowed torsionable bonds were assigned in ADT.
Docking of the pentasaccharides was performed using the AutoDock 3.0 program (39). The pentasaccharides were placed arbitrarily at two different positions on either side of the TB-cbEGF structure, 10 Å from the protein surface. Grids of probe atom interaction energies and electrostatic potential were generated with a spacing of 0.375 Å, using the AutoGrid program within the AutoDock suite. For each pentasaccharide, one simulation of 100 Lamarckian Genetic Algorithm runs was performed using 200 individuals and an energy evaluation number of 3 x 106. Each run produced one binding solution, and cluster analysis was performed at the end of each simulation. Solutions that were within 2 Å root mean square deviation of each other belonged to the same cluster, and clusters were ranked according to energy. Docking runs were performed with either the TB5-cbEGF model or the TB4-cbEGF structure.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
5000 response units (RU) for dp24). Internal fragment PF7 (encoded by exons 24-30) exhibited a response of over 200 RU that did not greatly depreciate after 2.5 min of dissociation. Because overlapping fibrillin-1 fragments PF5 and PF8 did not bind heparin, this binding site must be within the four cbEGF-like domains encoded by exons 26-29. Fragments PF2, PF12, and PF14 did not interact with heparin. Of the four overlapping fibrillin-1 fragments, PF9, PF10, PF11, and PF14, all except PF14 bound polymeric heparin and dp24. Thus, this heparin-binding site is located in the overlap between PF9 and PF11 (TB5 and the downstream cbEGF-like domain are encoded by exons 41-43). The C-terminal fragment PF13 also interacted with both heparin and dp24 (100 RU). These data show that there are four heparin-binding sites on fibrillin-1 and further localize the first three of these sites.
|
Binding Affinities of Polymeric Heparin and Defined Heparin Saccharides for Fibrillin-1—The association (ka) and dissociation (kd) rate constants of the molecular interactions between polymeric heparin and fibrillin-1 protein fragments (PF1, PF7, overlapping PF9/PF10/PF11/PF14, and PF13), and their dissociation constants (KD) were determined by surface plasmon resonance (Fig. 4 and Table I). Heparin strongly bound the first three fibrillin-1 sites on fragments PF1, PF7, and PF9/PF10/PF11, all with low nanomolar KD values, whereas the interaction of the C-terminal fragment PF13 had lower binding affinity.
|
To establish whether heparin oligosaccharide size influences binding to fibrillin-1, experiments were conducted using heparin saccharides dp12, dp16, and dp24, all of which were biotinylated by two comparable methods (cis-diol oxidation and reducing end) (36). Kinetic data were obtained, using the appropriate models outlined above, for the binding of PF1, PF7, and PF11 to dp12, dp16, and dp24. PF11 was chosen from the four overlapping fragments (PF9/PF10/PF11/PF14) around TB5, for kinetic analysis of heparin saccharide binding. The kinetic data are summarized in Table II, and Figs. 5 and 6. Only PF1 bound to dp12. Affinity to the three fibrillin-1 fragments increased with increasing saccharide chain length. PF1 had the highest affinity and PF7 the lowest affinity for dp24.
|
Effects of Calcium on Heparin/Fibrillin-1 Interactions—A previous study reported that 10 mM EDTA influences heparin binding to a large central fragment of fibrillin-1 but not to N- or C-terminal fragments (9). We have investigated the calcium dependence of the interactions between heparin and PF1, PF7, and PF11. These fibrillin-1 fragments (1 µg/ml) were incubated with increasing concentrations of EDTA (0.01-10 mM), prior to interaction with immobilized dp24. After subtraction of EDTA-only sensorgrams, the binding responses in all cases increased with increasing concentrations of EDTA up to 1 mM but then sharply decreased up to 10 mM EDTA (Fig. 7C). Reduced binding at EDTA concentrations that exceed buffer Ca2+ concentration indicates that these interactions between heparin and fibrillin-1 are all directly or indirectly calcium-dependent. Fragment PF1 contains 2 cbEGF-like domains; PF7 contains 6 cbEGF-like domains, and PF11 contains 10 cbEGF-like domains.
Impact of Heparin on Fibrillin-1 Assembly and Interactions with MAGP-1 and Tropoelastin
Effects of Heparin dp24 Binding to PF1 on N- and C-terminal Fibrillin-1 Interactions—To investigate whether homotypic fibrillin-1 interactions were affected by heparin, solid phase inhibition assays were performed. In a recent study, we used solid phase binding assays to show that PF1 and PF13 both interact strongly in a homotypic manner and with each other (25). Here heparin dp24 (4-12000 nM) was incubated with PF1 or PF13 (100 nM), before addition to either PF1 or PF13 pre-bound to the wells. Increasing concentrations of dp24 had no effect on the interaction of PF1 with PF13, or vice versa, and there was also no effect on the N-terminal PF1/PF1 interaction (data not shown). Thus, heparin dp24 and, by analogy, the S-domains of heparan sulfate in vivo do not inhibit head-to-tail fibrillin-1 assembly.
Effects of Heparin dp24 Binding to Fibrillin-1 PF1 on MAGP-1 Binding—We conducted inhibition assays to investigate whether MAGP-1 and heparin compete for binding to PF1. MAGP-1 binds close to the fibrillin-1 N terminus (28, 29). We investigated first whether MAGP-1 could inhibit the interaction of PF1 to dp24. Biotinylated DP24 was immobilized on the sensor chip, and MAGP-1 was initially passed over dp24, at concentrations ranging from 0.4 to 1200 nM, but no interaction was seen even at high concentrations (Fig. 7A, inset). PF1 (1 µg/ml, 20 nM) was then preincubated with MAGP-1 (0.4-1200 nM concentration range) and passed over the dp24 sensor chip surface. The binding response between PF1 and dp24 decreased with increasing concentrations of MAGP-1. The inhibition curve is shown in Fig. 7A. The EC50 value was calculated using GraphPad Prism 2.0 (nonlinear regression, sigmoidal dose response) and was found to be 57.1 nM. Similar experiments were conducted to determine whether heparin dp24 could inhibit the interaction of PF1 and immobilized MAGP-1. dp24 was found not to interact with MAGP-1 at concentrations up to 4000 nM. dp24 (0.4-4000 nM) was then preincubated with PF1 at 1 µg/ml (20 nM), prior to analysis of binding to MAGP-1. Increasing concentrations of dp24 inhibited the response of PF1 with MAGP-1 (Fig. 7A), and the EC50 value was calculated to be 59.9 nM. These data show that the heparin- and MAGP-1-binding sites on PF1 are adjacent or overlapping and that heparin strongly regulates MAGP-1 interactions with N-terminal fibrillin-1.
|
|
|
|
Influence of Heparin on RGD-mediated Cell Adhesion to PF9/PF11
Effects of Heparin dp24 on RGD-mediated Cell Adhesion to Fibrillin-1—Cell attachment assays were performed to study the effects of heparin on RGD-mediated cell adhesion. Cell binding to fibrillin-1 fragments PF14 (contains RGD but no heparin binding), PF9 (contains RGD and binds heparin), and PF10 (no RGD and binds heparin) was not affected by the presence of 125 nM or 1.25 µM heparin sulfate dp24 (data not shown).
|
All clusters were visualized using Quanta2000, and hydrogen bonds were calculated. The clusters with 5 and 4 members docked to the same site on TB5-cbEGF and bound in the interface between TB5 and the following cbEGF (site 1). The residues forming the pentasaccharide-binding site were Arg1697, Asn1724, Arg1727, and Arg1752 from the TB5 domain and Ser1788 and Arg1790 from the cbEGF domain (Fig. 9A). The clusters with 10 and 12 members bound toward the N-terminal region of the TB5 domain (site 2). The residues forming the pentasaccharide-binding site were Arg1691, Arg1692, Asn1713, Asn1730, Lys1731, and Tyr1760 all from the TB5 domain (Fig. 9B). For both heparin-binding sites, the surface charge distribution was viewed using Grasp (42) (Fig. 9, C and D). The N-terminal site (site 2) is distinct from the TB5-cbEGF interface site (site 1) and is actually on the opposite face of the TB5 domain (Fig. 9E).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The predominant disaccharide unit (80% of total disaccharides) of heparin used in this study was the trisulfated unit GlcNS,6S-IdoA,2S-; this unit is found exclusively in the S-domains of heparan sulfate, although it is very rare to find S-domains composed entirely of trisulfated disaccharide sequences (18). Fibrillin-1 fragment affinity for heparin increased with saccharide chain length. Comparison of two biotinylation methods confirmed similar profiles of heparin saccharide binding to four fibrillin-1 fragments, and thus the specificity of these interactions.
Binding of heparin to the N-terminal fibrillin-1 fragment PF1 exhibited high affinity and rapid binding kinetics. The importance of protein conformation in this interaction is suggested because a specific Marfan N-terminal mutation (T101A, within the first EGF-like domain) markedly reduces the affinity of PF1 for heparin despite not involving a positively charged residue. The distinctive kinetics may depend, in part, on protein folding involving residues within or adjacent to this domain to create the glycosaminoglycan-binding site. The eight Arg and Lys residues that occur in the mutated domain and its flanking domains are not clustered, and because the structure of PF1 is not known, the exact site of heparin binding is unclear, although it is likely to be conformational rather than linear. Because the T101A mutation does not involve a charged residue, it is possible that Thr101 is critical for the correct folding of the EGF-like domain encoded by exon 3. In addition, the N-terminal region of fibrillin-1 has a tendency to form disulfide-bonded aggregates that could be important in lateral assembly (44); heparin binding may be enhanced by such multimerization.
|
The third heparin-binding site was localized using overlapping fibrillin-1 fragments (PF9, PF10, PF11, and PF14) to the domains encoded by exons 41-43 (residues 1689-1807; TB5 and following cbEGF-like domain). This site is two domains downstream from the RGD-containing TB4 motif; however, bound heparin did not disrupt RGD-dependent cell attachment. Modeling revealed two possible heparin-docking sites; site-directed mutagenesis is now warranted to determine the usage of these sites.
The C-terminal HS-binding site was previously localized to the last 17 residues of mature processed fibrillin-1, which ends with the furin cleavage sequence RKRR (13). We have shown that the presence or absence of the post-furin cleavage sequence has no effect on the ability of heparin to bind this fragment, but heparan sulfate may, in vivo, regulate the furin cleavage event.
Analysis of the kinetics of heparin binding to fibrillin-1 revealed that the N-terminal and central sites all bound heparin very strongly with low nanomolar KD values, so these interactions are likely to be physiologically very significant. The very different kinetic parameters between the N-terminal PF1 fragment and the other peptides suggest different physiological roles. The rapid association of heparin with PF1 suggests that this is likely to be a more dynamic site, whereas interactions with PF7 and PF11 are "lazy" in that heparin both loads on and dissociates slowly. The C-terminal interaction is also likely to play an important role in fibrillin-1 biology. Implications of these kinetic studies for assembly and function are further discussed below.
A previous study indicated that 5 mM EDTA (heparin affinity chromatography) or 10 mM EDTA (solid phase assays) did not significantly affect heparin binding to the fibrillin-1 N terminus, although it did influence an internal fibrillin-1 binding region that includes the PF7 sequence. Here BIAcore analysis has revealed that increasing EDTA concentrations (1-10 mM) have complex effects but correlate with progressive reduction in heparin binding to N-terminal PF1, as well as the two internal sites identified here (PF7 and PF11). Thus, we conclude that calcium plays a role, directly or indirectly through conformation changes, in all three interactions.
The N terminus of fibrillin-1 is highly interactive. We and others have shown that it strongly binds to itself and to the fibrillin-1 C terminus (24, 25) and to MAGP-1 (28, 29). Here we have shown that heparin does not inhibit its ability to bind homotypically or with the C terminus, so heparin is unlikely to regulate linear fibrillin-1 assembly. Because heparin and MAGP-1 compete in a 1:1 manner for binding to the fibrillin1 N-terminal fragment PF1, their binding sites must be adjacent or overlapping. However, the Marfan mutation T101A, which reduces heparin affinity for PF1, does not affect MAGP-1 binding, so MAGP-1 and heparin-binding sites are not identical. The ability of tropoelastin to bind fibrillin-1 (PF7) was also compromised by heparin; however, previous binding assays using overlapping fibrillin-1 fragments indicated that this tropoelastin-binding site was within the domains encoded by exons 24 and 25 (29). Thus, it is likely that tropoelastin and heparan sulfate sites are adjacent but not overlapping. These studies have shown that heparan sulfate may profoundly influence microfibril and elastic fiber assembly.
In vivo, heparan sulfate occurs in two main sites, as syndecan and glypican cell surface receptors (46), and as perlecan in basement membranes. Fibrillin-1 interactions with heparan sulfate may be important at both sites. Interactions with cell surface receptors may profoundly influence microfibril and elastic fiber assembly. Fibrillin-1 binds integrin receptors 
51 and v3 (31, 47, 48), whereas tropoelastin binds integrin v3 and the elastin-binding protein (49, 50). It is possible that syndecan heparan sulfate chains may interact with newly secreted fibrillin-1 and influence molecular alignment and multimerization. Heparan sulfate proteoglycans with several heparan sulfate chains in close proximity on the core proteins might act as a template for assembly of fibrillin monomers using the N-domain as a docking site with fast association/dissociation from heparan sulfate chains. Previous studies have revealed that heparan sulfate, and sulfation, also play a key role in elastic fiber formation (14, 15). Heparan sulfate can interact strongly with tropoelastin, which contains numerous charged Lys residues, thereby influencing its potential to interact with fibrillin-1 and to form lysyl-derived cross-links. A recurring theme is the colocalization of fibrillin-1 microfibrils with basement membranes. Our previous structural studies of microfibrils allowed us to map the alignment of fibrillin-1 within mature beaded microfibrils (32). The two central fibrillin-1 heparan sulfate-binding sites map within the "interbead" region of microfibrils and are likely to be available for interactions with heparan sulfate chains. However, N- and C-terminal binding sites may be within the bead structure and unavailable for interactions.
In summary, this study has provided important new insights into the molecular basis of heparin/heparan sulfate interactions with fibrillin-1, shown how these glycosaminoglycans regulate fibrillin-1 interactions with MAGP-1 and tropoelastin, and demonstrate that they do not influence RGD-dependent cell attachment to fibrillin-1.
|
FOOTNOTES |
|---|
** Supported by the Royal Society. To whom correspondence should be addressed: Faculty of Life Sciences, Michael Smith Bldg., Oxford Rd., Manchester M13 9PT, UK. Tel.: 44-161-275-5739; Fax: 44-161-275-5082; E-mail: cay.kielty{at}manchester.ac.uk.
1 The abbreviations used are: EGF, epidermal growth factor-like domain; cbEGF, calcium-binding EGF; HS, heparan sulfate; PF, protein fragment; BSA, bovine serum albumin; RU, response units; MES, 4-morpholineethanesulfonic acid; dp, degree of polymerization.
2 Marfan syndrome mutation data base (www.umd.be).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), PF1 (1 µg/ml) was incubated with MAGP-1 (0.4-1200 nM), before addition to immobilized dp24. Inhibition of PF1 with MAGP-1 by dp24, PF1 (1 µg/ml) (
), was incubated with dp24 (0.4-4000 nM) before addition to immobilized MAGP-1. The response difference (Resp. Diff.) was normalized to a percentage of the response of no inhibitor added. B, inhibition of fibrillin-1 PF7 with tropoelastin by dp24. PF7 (1 µg/ml) (
