Elements within the first 17 amino acids of human osteonectin are responsible for binding to type V collagen.

The region in human osteonectin (ON) responsible for binding to type V collagen has been identified as the first 17 NH2-terminal residues. This conclusion is based upon binding studies with deletion mutants of ON produced in Escherichia coli, in which parts of the first 17 amino acids have been removed. Wild-type ON from E. coli and mammalian cell-derived nonglycosylated ON bind identically to type V collagen and at least twice as effectively as mammalian cell-derived N-glycosylated ON. In previous studies, it was shown that N-glycosylation at residue 99 significantly reduces the capacity of ON to bind to type V collagen. Results reported in this communication demonstrate that the actual binding site on ON for type V collagen is distal from the site of N-glycosylation in terms of amino acid sequence but may be proximal in the folded, fully glycosylated, three-dimensional structure. Consistent with this conclusion is the ability of a synthetic peptide consisting of amino acids 1-17 to specifically inhibit the binding of ON to type V collagen.

Osteonectin (ON) 1 /SPARC/BM-40 (1-3), a secreted, singlechain, acidic, Ca 2ϩ -binding glycoprotein, is a major noncollagenous extracellular matrix protein in bone and dentine (1,4) as well as in many normal and neoplastic human soft tissues (5) and cultured cells (6,7). It is also synthesized, stored, and secreted by human blood platelets (8,9). Earlier studies have suggested several functions for ON in the extracellular matrix, including the regulation of bone mineralization (1,10), the control of cell shape (11), tissue remodeling or repair (12), cell migration (6), proliferation, and differentiation (12)(13)(14)(15)(16). Osteonectin from different sources binds differentially to different type collagens. Osteonectin from bovine bone binds to type I, III, and V collagen (1,15,17), whereas that from mouse parietal yolk sac cells binds only to type III and V collagen (18), and that from human platelets has no affinity for any of the three collagen types (17). The difference between bovine bone and human platelet ON binding to collagen has been attributed to differences in N-glycosylation (17). Osteoblast-and megakaryocyte-derived mRNA encoding ON are identical in size and restriction enzyme fragmentation patterns (19), lending fur-ther support to the hypothesis that differences in structure and collagen binding between bone and platelet-derived ON reside at the level of N-glycosylation. Osteonectin (BM-40) from mouse Engelbreth-Holm-Swarm tumor binds to type IV collagen but shows markedly reduced binding to type I, III, V, and VI collagen (20). The region of BM-40 binding to type IV collagen has been identified as the EF-hand and ␣-helical domains in the carboxyl-terminal half of BM-40 (21).
Mature human ON consists of 286 amino acids and contains two potential Asn-X-(Thr/Ser) N-gycosylation sites, located at positions 71 and 99 (22,23). Recently, we reported that the binding site in human ON for type V collagen resides in the amino-terminal half (amino acids 1-146) of ON and that the capacity for type V collagen binding is significantly reduced by N-glycosylation at residue 99 (24). Residue 71 appears to have no, or very little, attached carbohydrate (24). In the present study, employing a set of deletion mutants, the binding region in ON has been further localized to the first 17 amino acids of the mature protein and appears to involve a tertiary structure comprised of several amino acids.
Construction of Wild-type Truncated Human Osteonectin Expression Vector and Site-specific Deletion Mutagenesis-Polymerase chain reaction (PCR)-enabled mutagenesis (25) was used to construct a pUC19derived plasmid vector for the expression of wild-type amino-terminal half (amino acids 1-146) of human ON (tHON) in E. coli. Advantage was taken of the multiple cloning site (MCS) within the expressible ␤-galactosidase gene of pUC19 (26). Consequently, tHON and derived deletion mutants contain at their NH 2 terminii seven amino acids of ␤-galactosidase (NH 2 -(Met)-Thr-Met-Ile-Thr-Pro-Ser-Leu) and eight "zeno" amino acids (Thr-Gly-Arg-Arg-Phe-Thr-Thr-Ser-COOH) at their COOH terminii. Template for PCR-enabled mutagenesis was wild-type tHON (amino acids Ϫ17 to ϩ146) mammalian cell expression vector tHON/pD5, described elsewhere (24). Fig. 1A shows schematically the construction of wild-type tHON/pUC19 expression vector, and Fig. 1B shows the resulting recombinant protein products for wild-type and the six deletion mutants. Mutagenic primer B shown in Fig. 1A (5Ј-ccgca-GAATTCGGTCAGCTCAGA-3Ј) is composed of a five-nucleotide "clamp" followed by the inverse complement of amino acid codons 146 -141 of ON and includes a naturally occurring EcoRI digestion site (underlined). Mutagenic primers A consisted of the following: 5Ј-ca-gaaagcttgGCCCCTCAGCAAGAAGCCCTGC-3Ј for wild-type tHON (composed of a four-nucleotide clamp, seven nucleotides of the pUC19 MCS including a HindIII digestion site (underlined) and one nucleotide * This work is supported by National Institutes of Health Grants PO1-AG08777 and CO6-HL39475. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
PCR conditions were the same as described previously (25). After PCR, each product was digested with HindIII and EcoRI restriction enzymes, and the resulting fragment was resolved on a 1% agarose gel (TAE buffer), isolated by electroelution, and ligated into pUC19 from which the corresponding HindIII/EcoRI MCS fragment had been removed. The ligated mixture was then used to transform E. coli strain DH5␣ cells according to the supplier's directions and plated on 5-bromo-4-chloro-3-indoyl ␤-D-galactoside/isopropyl-1-thio-␤-D-galactopyranoside/ampicillin/LB agar plates. Isolated white colonies were grown in liquid culture, and plasmid DNA was isolated by the alkaline lysis/ Qiagen column procedure according to the supplier's directions. Resulting plasmids were analyzed by digestion with HindIII, EcoRI, and PvuII, and the region between the HindIII/EcoRI sites was sequenced (Applied Biosystems Inc., model 373 A DNA sequencer) to confirm the desired deletion and to exclude inadvertent mutations introduced from PCR. Sequencing reactions were performed with the TaqDye Deoxy Terminator cycle sequencing kit according to the supplier's instructions (Applied Biosystems Inc., Foster City, CA).
Expression and Purification of Wild-type tHON and Deletion Mutants from E. coli-Bacteria containing the correct mutant plasmid were grown in liquid LB media with 50 g of ampicillin/ml at 37°C in a shaker bath until the A 600 nm Ϸ 0.5. At this point isopropyl-1-thio-␤-Dgalactopyranoside (1 mM final concentration) was added to induce protein expression, and culture was continued for 6 h to achieve maximum expression. Cells were harvested by low speed centrifugation and lysed by sonication on ice (4 ϫ 30 s) in cold 20 mM Tris, 5 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10 mM benzamidine, pH 7.4. The sonicate was centrifuged at 30,000 ϫ g for 30 min to remove cell debris. Protamine sulfate in liquid form was slowly added with stirring to a final concentration of 0.1% to remove nucleic acids. After the addition of protamine sulfate, the solution was cleared by centrifugation at 30,000 ϫ g for 30 min (27). Following this modified step, the supernaturant containing ON was purified as described previously (24). The purified protein was analyzed by Western blotting and protein silver staining. Protein concentration was determined by Micro BCA protein assay (Pierce Inc. Rockford, IL) following the supplier's instructions and using bovine bone osteonectin as a standard.
Binding Assays-The binding of different forms of ON to type V collagen and monoclonal antibody IIIA 3 A 8 were performed by the ELISA method as described previously (24).
Synthesis and Purification of Peptides-A peptide corresponding to amino-terminal amino acids 1-17 of ON was synthesized on an Advanced Chemtech model 90 automatic peptide synthesizer using conventional Merrifield chemistry with t-butoxycarbonyl amino acid-derivatives and carbodiimide coupling. Following synthesis, the peptide was cleaved from the solid support resin with HF in trifluoroacetic acid and purified by HPLC reverse-phase chromatography. The peptide, migrating as a single, resolved peak by HPLC, was checked for purity and composition by conventional amino acid analysis and mass analysis on a PerSeptive Biosystems (Farmington, CT) Voyager "time-of-fight" mass spectrometer and shown to be at least 80 -85% pure. A control peptide of the same size and similar charge properties, SNNGNRRNYY-IAAEEISC, was synthesized and characterized by the above methods.

Construction of Wild-type tHON and Deletion Mutant Expression Vectors-cDNA sequencing and endonuclease digestion of wild-type and mutant vectors confirmed that the desired constructs had been accomplished and that no inadvertent mutations had been introduced by PCR amplification (data not shown).
Expression, Purification, and Properties of Wild-type tHON and Mutants-Each of the purified proteins showed a single band upon 12% SDS-PAGE under reducing conditions (Fig.  1C). Wild-type chimeric tHON migrates in a manner consistent with its predicted 161 amino acid size. As shown in Fig. 1C, each of the deletion mutants moves slightly faster than wildtype ON. Amino acid sequencing of purified forms of ON confirmed that the NH 2 terminii, including the deletion regions, were as expected (data not shown). Fig. 2 demonstrates that over a concentration range of 6 -100 nM ON, the affinity and capacity for monoclonal antibody IIIA 3 A 8 binding is equivalent for wild-type and mutant tHONs. These results validate the use of IIIA 3 A 8 to detect the different forms of ON in the binding assays described below.

Affinity of Antibody IIIA 3 A 8 for Wild-type tHON and Deletion Mutants-
Comparison of E. coli and Mammalian Cell-derived tHON Binding to Type V Collagen-In our previous report, it was shown that removal of N-linked carbohydrate from tHON by site-directed mutagenesis (Asn 99 3 Gln; N99Q) significantly increased the binding to type V collagen (24) . Fig. 3 compares the binding of E. coli and mammalian cell-derived wild-type tHONs and the N99Q mutant to type V collagen. The curves indicate that nonglycosylated wild-type tHON from E. coli has the same affinity for type V collagen as the nonglycosylated N99Q mutant and that both have enhanced binding compared with glycosylated wild-type tHON. These results confirm that the difference in type V collagen binding between the mammalian cell-derived wild-type and N99Q tHON is due to the lack of carbohydrate and not due to the amino acid substitution itself, since the presence of Asn at position 99 (E. coli material) versus Gln (N99Q mutant) has no apparent effect. These results also demonstrate that ON produced in E. coli binds equivalently to that from mammalian cells. Furthermore, the results demonstrate that the seven-amino acid NH 2 -terminal ␤-galactosidase and eight-amino acid COOH-terminal "zeno" extension peptides have no apparent effect on tHON binding to type V collagen.
Binding of the Deletion Mutants to Type V Collagen-Results shown in Fig. 4 indicate that the ability of wild-type tHON to bind type V collagen is significantly affected by deletions in the NH 2 -terminal region. Deletion of the first four (⌬4) or eight (⌬8) NH 2 -terminal residues of tHON reduces the binding capacity by 73 and 91%, respectively. Further deletion (⌬12 and ⌬17) showed only slight binding above background. These results suggest that the first four residues of ON play an important role in type V collagen binding. However, as shown in Fig.  4, the internal deletion mutants of residues 5-8 (⌬ 5-8 ) and 9 -12 (⌬ 9 -12 ) also exhibit severely reduced (if any) binding. Taken together, the mutant results clearly indicate that the NH 2 -terminal 17 amino acid region of ON is the major contributor to the binding interaction between ON and type V collagen; the binding site on ON is not just a unique amino acid residue or a few adjacent residues, but rather appears to involve an extended conformational interaction or distal binding residues within the first 17 amino acids.
Competitive Inhibition Assay of Amino Acids 1-17 Peptide Binding to Type V Collagen-As an independent confirmation of the ability of the amino acids 1-17 region of ON to bind to type V collagen, the effect on binding of a peptide representing amino acids 1-17 was studied. The results, presented in Fig. 5, indicate that the peptide competitively blocks the binding of tHON to type V collagen, with a K I Х 10 M (50% inhibition). In contrast, the control peptide over the same concentration range (5-100 M) had no effect. One possible explanation considered by us for the loss of binding by the deletion mutants was the necessary participation of an amino acid segment in the 1-17 region and a segment(s) in the distal 18 -146 region. To test the possibility of complimentation of binding, the effect of adding amino acids 1-17 peptide to ⌬ 17 mutant protein was also measured. As Fig. 5 indicates, the addition of the peptide did not facilitate the binding of the ⌬ 17 mutant ON. These results suggest the absence of co-participation by a second distal segment in type V collagen binding. DISCUSSION Earlier studies indicate that bovine bone and human bone ON bind to type V collagen (4,17). Osteonectin from mouse parietal yolk sac cells also binds to type V collagen (18). In addition, studies with a truncated form of human ON containing only the amino-terminal half of the protein effectively binds to type V collagen, demonstrating that the binding region of ON resides within amino acid residues 1-146 (24). Based upon differences in glycosylation patterns as well as lectin binding properties, it was proposed that the collagen binding specificity of bone and platelet ON is related to differences in glycosylation (16). In order to further understand the region of ON involved in collagen binding and the effects of specific N-glycosylation sites on biological activity, we have demonstrated that after removal of oligosaccharide chain structures from bovine bone and human platelet ON by N-glycanase, their ability to bind to type V collagen is increased to an equal level. In addition, the results of site-specific mutagenesis at each of the two potential Asn-X-Thr glycosylation sites (amino acids 71 and 99) in ON indicate that only glycosylation at residue 99 affects type V collagen binding activity (24). Further studies using deletion mutants reported here demonstrate that the binding site in ON for type V collagen is in the first 17 amino acids of the protein. This segment is distal from the N-glycosylation site in regard to the amino acid sequence, but due to protein folding, and the spacial umbrella of the carbohydrate, may be conformationally proximal to the site of N-linked carbohydrate. No information relating to the three-dimensional structure of ON is available to further interpret our results. A second important conclusion reached from this study involving a family of related deletion mutants is that it does not appear that a unique amino acid residue or short segment of amino acids is capable by itself of directing effective ON binding to type V collagen, but rather, several small segments and/or an overall conformational state in the first 17 residues is necessary.
As indicated in Fig. 3 and reported in our earlier studies (24), tHON achieves half-maximum binding at ϳ100 nM, representing approximately 100-fold greater affinity for type V collagen than the blocking peptide (K I Х 10 M). Differences of this order of magnitude are commonly observed for synthetic peptides versus the native protein and are thought to be largely due to the greater flexibility and conformational heterogeneity of the synthetic peptide. Despite the difference in affinity, the fact that complete inhibition can be accomplished by the peptide alone and the absence of binding in the presence of peptide plus ⌬ 17 ON mutant protein strongly suggest that the binding of ON to type V collagen is recapitulated by the peptide.
Type V collagen is particularly abundant in vascular tissue, primarily due to its synthesis in smooth muscle cells. Smooth muscle cells and their protein products, including possibly type V collagen, are believed to play an important role in the development of atherosclerotic plaque. In human atherosclerotic lesions, the ratio of type V collagen to other types is elevated relative to the normal situation (28,29). Additionally, two reports indicate that type V collagen can have a procoagulant effect (30,31). This might result from the disruption or destruction of vascular endothelial cells at sites of tissue injury or remodeling, and consequential exposure of type V collagen by underlying smooth muscle cells. Kelm et al. (32) have reported that osteonectin binds to plasminogen and enhances tissue plasminogen activator conversion of plasminogen to plasmin. Both plasmin and tissue plasminogen activator are important thrombolytic agents. Kelm and co-workers (32) also reported the mediation of plasminogen binding to type V collagen by bovine bone osteonectin (32). Consequently, osteonectin, by serving as a bridge between type V collagen and thrombolytic agents, may play an important role in hemostasis in the absence of functional endothelium. FIG. 4. Binding of deletion mutants to type V collagen. Samples were prepared as described under "Experimental Procedures" and assayed as described for Fig. 3. The absolute absorbance at 490 nm for wild-type tHON (lane 1) was 1.35, as shown in Fig. 3. Measured binding of all the protein species at concentrations indicated in Fig. 3 was equal or less than that at 1 M (data not shown).
FIG. 5. Competitive inhibition of ON binding to type V collagen by aa 1-17 peptide. Type V collagen, at 10 gm/ml in 50 mM NaHCO 3 (pH 9.7), was applied to plastic microtiter wells for 3 h. After washing and blocking steps, solutions of tHON and ⌬ 17 in TBS/Tween/ CaCl 2 buffer were applied to the wells and incubated with different concentrations of peptide for 18 h at 4°C. Final concentration of ON in each well is 0.125 M. After washing, ON-collagen complexes were detected by the ELISA assay as described under "Experimental Procedures." tHON ϩ 1-17 peptide (E); tHON ϩ control peptide (q); ⌬ 17 ϩ 1-17 peptide (Ç); 1-17aa peptide alone (å); and control peptide alone (Ⅺ). synthesis and purification of peptides, and for advice on the competitive peptide binding studies.