Bone Morphogenetic Protein-1/Tolloid-related Metalloproteinases ...

6 downloads 4828 Views 314KB Size Report
Jun 14, 2004 - processing of NH2-terminal globular domains, or in some cases. C-propeptides, of minor ... to free the TGF-ß-like morphogens BMP-2 and BMP-4 from ..... alone (No enzyme), or in the presence of BMP-1 (BMP1), mTLD (TLD),.
THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2004 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 279, No. 40, Issue of October 1, pp. 41626 –41633, 2004 Printed in U.S.A.

Bone Morphogenetic Protein-1/Tolloid-related Metalloproteinases Process Osteoglycin and Enhance Its Ability to Regulate Collagen Fibrillogenesis* Received for publication, June 14, 2004, and in revised form, July 28, 2004 Published, JBC Papers in Press, July 29, 2004, DOI 10.1074/jbc.M406630200

Gaoxiang Ge‡§, Neung-Seon Seo§¶, Xiaowen Liang ¶, Delana R. Hopkins储, Magnus Ho¨o¨k ¶**, and Daniel S. Greenspan‡储‡‡ From the ‡Department of Pathology and Laboratory Medicine and 储Program in Molecular and Cellular Pharmacology, University of Wisconsin, Madison, Wisconsin 53706 and the ¶Center for Extracellular Matrix Biology, Albert B. Alkek Institute of Biosciences and Technology, Texas A&M University Health Science Center, Houston, Texas 77030

The mammalian bone morphogenetic protein-1 (BMP1)/Tolloid-related metalloproteinases play key roles in regulating formation of the extracellular matrix (ECM) via biosynthetic processing of various precursor proteins into mature functional enzymes, structural proteins, and proteins involved in initiating the mineralization of hard tissue ECMs. They also have been shown to activate several members of the transforming growth factor-␤ superfamily, and may serve to coordinate such activation with formation of the ECM in morphogenetic events. Osteoglycin (OGN), a small leucine-rich proteoglycan with unclear functions, is found in cornea, bone, and other tissues, and appears to undergo proteolytic processing in vivo. Here we have successfully generated recombinant OGN and have employed it to demonstrate that a pro-form of OGN is processed to varying extents by all four mammalian BMP-1/Tolloid-like proteinases, to generate a 27-kDa species that corresponds to the major form of OGN found in cornea. Moreover, whereas wild-type mouse embryo fibroblasts (MEFs) produce primarily the processed, mature form of OGN, MEFs homozygous null for genes encoding three of the four mammalian BMP-1/Tolloid-related proteinases produce only unprocessed pro-OGN. Thus, all detectable pro-OGN processing activity in MEFs is accounted for by products of these genes. We also demonstrate that both proand mature OGN can regulate type I collagen fibrillogenesis, and that processing of the prodomain by BMP-1 potentiates the ability of OGN to modulate the formation of collagen fibrils.

Bone morphogenetic protein-1 (BMP-1)1 is the prototype of a class of structurally similar metalloproteinases that play various morphogenetic roles in a broad spectrum of species (1, 2).

* This work was supported by National Institutes of Health Grants AR47746, GM63471 (to D. S. G.), and AR44415 (to M. H.). 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. § Both authors contributed equally to this work. ** To whom correspondence may be addressed. Tel.: 713-677-7552; Fax: 713-677-7576; E-mail: [email protected]. ‡‡ To whom correspondence may be addressed. Tel.: 608-262-4676; Fax: 608-262-6691; E-mail: [email protected]. 1 The abbreviations used are: BMP, bone morphogenetic protein; mTLD, mammalian Tolloid; mTLL, mammalian Tolloid-like; ECM, extracellular matrix; SLRP, small leucine-rich proteoglycan; OGN, osteoglycin; MEF, mouse embryo fibroblast; MALDI-TOF, matrix-assisted laser desorption ionization time-of-flight; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid.

There are four mammalian members of this family: BMP-1, mammalian Tolloid (mTLD), and mammalian Tolloid-like 1 and 2 (mTLL-1 and mTLL-2) (3–5). BMP-1 and mTLD are encoded by alternatively spliced mRNAs of the same gene (3), whereas mTLL-1 and -2 are genetically distinct (4, 5). These proteinases play key roles in regulating formation of mammalian extracellular matrix (ECM), via biosynthetic processing of precursor proteins to form mature, functional matrix components. In the case of collagen fibers, this includes processing of the C-propeptides of procollagens I–III to yield the major fibrous components of ECM (5–9); proteolytic activation of the enzyme lysyl oxidase (10), which is necessary to the formation of covalent cross-links in collagen and elastic fibers (11); and processing of NH2-terminal globular domains, or in some cases C-propeptides, of minor fibrillar procollagen V and XI chains (12–14) to yield type V and XI monomers. Such monomers are incorporated into collagen types I and II fibrils, respectively, and appear to control the shapes and diameters of the resultant heterotypic collagen fibrils (15–18). Members of the same small group of proteinases also process precursors for laminin 5 (19, 20) and type VII collagen (21), both of which are involved in securing epithelia to underlying basement membrane-like structures (22); and for dentin matrix protein-1 (23), which is involved in initiating mineralization in the ECM of bones and teeth (24). The mammalian BMP-1-related proteinases are all capable of activating the TGF-␤-like protein growth differentiation factor 8 (also known as myostatin), by freeing it from a noncovalent latent complex with its cleaved prodomain (25). Similarly, BMP-1 and mTLL-1, but not mTLD or mTLL-2, are able to free the TGF-␤-like morphogens BMP-2 and BMP-4 from latent complexes with the extracellular antagonist chordin (5, 9). Thus, BMP-1-like proteinases may orchestrate formation of the ECM with signaling by various TGF-␤-like proteins in morphogenetic and homeostatic events. The small leucine-rich proteoglycans (SLRPs) constitute a small family of secreted proteoglycans/glycoproteins with structurally related core proteins, with 11 members in mammals (26 –30). SLRP core proteins contain a central domain consisting of 6 –12 tandem leucine-rich repeats, and flanking NH2- and COOH-terminal domains containing characteristic cysteine clusters. The SLRPs can be divided into three subclasses, based on the intron/exon organization of cognate genes, spacing of the four cysteine residues in NH2-terminal regions, and the number of leucine-rich repeats in the central domains. Class I comprises the chondroitin/dermatan sulfate chain-containing proteoglycans decorin and biglycan, and the glycoprotein asporin; class II comprises the keratan sulfate chain-con-

41626

This paper is available on line at http://www.jbc.org

BMP-1/Tolloid-like Proteinases Cleave Osteoglycin taining proteoglycans fibromodulin, lumican, keratocan, and osteoadherin, and the glycoprotein PRELP; and class III consists of osteoglycin (also known as mimecan) and epiphycan, both of which carry a glycosaminoglycan chain, and the glycoprotein opticin. Whereas evidence is lacking for processing of precursors to mature forms of class II SLRPs, biglycan is cleaved from a precursor to the mature protein by BMP-1related proteinases, in vitro and in vivo (31). Because biglycan appears to be a positive regulator of bone growth (32), involvement of the BMP-1-related proteinases in biglycan biosynthesis conforms to other roles of these proteinases in formation of bone, through formation and mineralization of ECM and through activation of TGF-␤-like BMPs. Interestingly, the two class III SLRPs, osteoglycin (OGN) (33) and epiphycan (34) appear to be proteolytically processed in vivo and the cleavage sites have features resembling those of the cleavage sites of known substrates of BMP-1-like proteinases. Epiphycan, also known as PG-Lb, is localized to the ECM of growth plate chondrocytes and may play structural and/or instructional roles in the integrity and development of this tissue (34, 35). OGN is one of the three major keratan sulfatecontaining proteoglycans in cornea (33), and may play a role in regulating collagen fibrillogenesis (36). Here we extend the known range of substrates in general and of SLRP substrates in particular of the BMP-1-like proteinases, by demonstrating that such proteinases play roles in biosynthetic processing of the class III SLRP OGN. Moreover, cleavage by these proteinases is shown to affect the ability of OGN to regulate the fibrillogenesis of type I collagen. EXPERIMENTAL PROCEDURES

Pro-OGN Expression Vector—An 897-bp cDNA fragment encoding full-length pro-OGN was amplified by PCR using an EST clone, image number 5067073 (ATCC), as a template, Pfu DNA polymerase (Stratagene) and two primers, 5⬘-CGGGATCCATGGAGACTGTGCACTCTACATTTCTC-3⬘ and 5⬘-GGAATTCCTTAGAAGTATGACCCTATGGGTAATC-3⬘. Use of these primers created a BamHI site at the 5⬘ end and an EcoRI site at the 3⬘ end of the fragment (restriction site sequences are in boldface), which was inserted between BamHI and EcoRI sites of pBluescript II SK(⫹) (Stratagene). The full-length pro-OGN cDNA was then used as a template for PCR amplification of full-length pro-OGN sequences to which a His6 tag was fused via a thrombin cleavage site to the NH2 terminus. Primers were 5⬘-CGCGGATCCTCACCATCATCACCATCACCTGGTTCCGCGTGGATCTGCACCACAGTCGCAGCTGGAC-3⬘ and 5⬘-GGAATTCCTTAGAAGTATGACCCTATGGGTAATC-3⬘. These added a BamHI site to the 5⬘ end and an EcoRI site to the 3⬘ end (restriction site sequences are in boldface) of the amplified fragment, which was inserted between BamHI and EcoRI sites of pAcGP67-A (BD Pharmingen). Predicted sequences of the resulting expression vector, pOGN, were confirmed by DNA sequence analysis. Oligonucleotide primers were synthesized by IDT, Inc. Insect Cell Culture—Spodoptera frugiperda 9 (Sf9) cells (obtained from Dr. Mary Estes, Baylor College of Medicine, Houston, TX) were grown and maintained on Grace’s insect cell medium (Invitrogen) containing 10% (v/v) fetal bovine serum (Gemini Bio-products), lactalbumin hydrolysate (3.33 mg/liter), yeastolate (3.33 mg/liter), and L-glutamine. Spinner cultures were started at an initial density of 5 ⫻ 105 cells/ml with stirring at 80 –90 rpm. The spinner cultures were subcultured when the cell density reached ⬃3 ⫻ 106 cells/ml. Expression and Purification of Recombinant Pro-OGN—Monolayers of exponentially growing Sf9 insect cells (2 ⫻ 106 cells in a 60-mm plate) were co-transfected with 5 ␮g of transfer vector pOGN and 0.5 ␮g of linearized BaculoGold virus DNA (BD Pharmingen) in transfection buffer (BD Pharmingen). Recombinant baculoviruses, vOGN, were selected and purified three times by plaque assay. Positive colonies were confirmed by purification of secreted pro-OGN and identification via Western blot using affinity purified polyclonal rabbit antibodies raised against a synthetic peptide derived from the COOH-terminal region of murine OGN (amino acid 255–298). Positive colonies were used to propagate high titer stocks (108–109 plaque forming units/ml), which were stored at 4 °C. For a scale-up of protein production, Sf9 insect cells were maintained in spinner flasks (Bellco Glass, Inc.) and the recombinant vOGN virus

41627

stocks were used at a multiplicity of infection of 10 to infect Sf9 insect cells at a density of 3 ⫻ 106 cells/ml. At 1 h post-infection, recombinant virus-infected cells were harvested and resuspended in serum-free medium (BD Pharmingen). At 96 h post-infection conditioned medium was collected, centrifuged, and filtered using a 0.2-␮m membrane filter. The supernatant was then applied to the Labscale TFF system (Millipore, Co.) with the Pellicon XL module (Biomax 10) for concentration of the sample volume and buffer change to 20 mM Tris, pH 8.0. The sample was then applied to a 5-ml High Trap metal chelating column charged with Ni2⫹ (Amersham Biosciences), which was washed with 30 mM imidazole buffer containing 20 mM Tris, pH 8.0, 0.5 M NaCl, and 0.02% CHAPS, after which bound protein was eluted from the resin with the same buffer containing 100 mM imidazole. Eluted fractions were analyzed by SDS-PAGE, and the peak fractions were pooled and concentrated by Centricon-10 spin columns (Amicon). The mass of purified protein was determined by MALDI-TOF mass spectrometry (Tufts University, Protein chemistry facility). Prior to analysis the proteins were dialyzed into 10 mM NH4HCO3. SDS-PAGE and Western Blotting—SDS-PAGE was performed with 10 (Figs. 1 and 2), 15 (Figs. 4 and 5), and 12% (Fig. 6) polyacrylamide gels with 4% stacking gels and the Laemmli buffer system (37). For Western blots, proteins separated on SDS-PAGE were transferred to nitrocellulose membranes and probed with affinity purified polyclonal rabbit antibodies raised against a synthetic peptide derived from the COOH-terminal region of murine OGN (amino acids 255–298). Horseradish peroxidase-conjugated goat anti-rabbit affinity purified immunoglobulin G (heavy and light chains) (Bio-Rad) was used as the secondary antibody, and SuperSignal chemiluminescent substrate (Pierce) was used for signal detection. Prestained broad-range marker proteins (Bio-Rad) were used as molecular mass standards. A monoclonal mouse antibody (anti-Penta His, Qiagen) and horseradish peroxidase-conjugated goat anti-mouse affinity purified immunoglobulin G (Bio-Rad) were used as first and secondary antibodies, respectively, for detecting the recombinant pro-OGN His tag. Peptide N-glycosidase F Digestion—Purified recombinant pro-OGN was denatured with 0.5% SDS, 1% 2-mercaptoethanol for 10 min at 100 °C and was then treated with 25 units of peptide N-glycosidase F (New England Biolabs) in 50 mM sodium phosphate, pH 7.5, 1% Nonidet P-40 at 37 °C for 16 h. After incubation, samples were immediately subjected to SDS-PAGE analysis. Thrombin Cleavage and Purification of Non-tagged Pro-OGN—Purified recombinant pro-OGN was treated with 1 unit of biotinylated thrombin (Novagen) in 20 mM Tris-HCl, pH 8.4, 150 mM NaCl, 0.25 mM CaCl2 at room temperature for 16 h. After incubation, samples were sequentially subjected to streptavidin-agarose column and Ni2⫹charged metal chelating column chromatography to remove the biotinylated thrombin and the cleaved His tag, respectively. Cleavage of the His6 tag was confirmed by SDS-PAGE and Western blot analysis, using anti-His antibody. Prior to use in experiments non-tagged proteins were dialyzed into phosphate-buffered saline. Collagen Fibrillogenesis Assay—Fibrillogenesis assays were performed as described previously (38). Briefly, at 4 °C, stock solutions (1–2 mg ml⫺1) of bovine dermal type I collagen (VITROGEN, Palo Alto, CA) were neutralized with 10⫻ phosphate-buffered saline and brought to a concentration of 0.5 mg ml⫺1 with 1⫻ phosphate-buffered saline, pH 7.2. Recombinant protein was added to aliquots of the mixture at 1– 60 ␮g ml⫺1, samples were transferred to a 96-well plate in a SpectraMax Plus 384 Microplate Spectrophotometer (Molecular Devices Co.), and assays were performed at 37 °C. In Vitro Enzyme Assays—500 ng of recombinant pro-OGN was incubated alone or in combination with 15 ng of recombinant BMP-1, mTLD, mTLL-1, or mTLL-2, containing COOH-terminal FLAG epitopes, in 20 ␮l total volume of 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM CaCl2, and incubated 20 h at 37 °C. Reactions were stopped by addition of 5⫻ SDS-PAGE sample buffer (37) containing 1% 2-mercaptoethanol, and boiling of samples for 5 min. 50 ng of proteins per sample were subjected to SDS-PAGE and subsequent Western blot analysis, as described above. Amino Acid Sequence Analysis—2 ␮g of purified recombinant proOGN was incubated with 50 ng of BMP-1 at 37 °C for 20 h, and the reaction was quenched and prepared for SDS-PAGE as above. Products were resolved by SDS-PAGE on a 12% polyacrylamide gel and electrotransferred to Sequi-Blot polyvinylidene difluoride membrane (BioRad). Proteins were revealed with 0.025% Coomassie Brilliant Blue R-250, and NH2-terminal amino acid sequences were determined by automated Edman degradation at the Harvard University Microchemistry Facility using a PerkinElmer/Applied Biosystems Division Procise 494 HT Protein Sequencing System.

41628

BMP-1/Tolloid-like Proteinases Cleave Osteoglycin

FIG. 1. Expression and purification of recombinant OGN. A, schematic representation of prepro-OGN. Shown are signal sequence (sig), prodomain (pro), glycosaminoglycan (GAG chain) attachment site, six leucine-rich repeats (numbered boxes), and an N-linked oligosaccharide attachment site. B, pro-OGN expression construct comprising a GP67 signal sequence (GP67 sig), and sequences encoding six consecutive histidine residues (His x 6), a thrombin protease recognition sequence (Th), and the murine pro-OGN core protein, all under the control of the polyhedrin gene promoter (polyhedrin). C, recombinant pro-OGN eluted from a nickel-chelating column (lane 2) was analyzed by SDS-PAGE and staining with Coomassie Blue. Molecular weight standards are indicated (lane 1). D, purified recombinant OGN was detected by Western blot using an anti-OGN antibody (lane 3). Lanes 1 and 2 represent similarly treated samples derived from uninfected Sf9 cells or Sf9 cells infected with the wild-type virus AcMNPV, respectively. Mouse Embryo Fibroblasts (MEFs)—MEFs were isolated from 13.5day post-conception embryos as previously described (8). Cells were maintained in growth medium consisting of Dulbecco’s modified Eagle’s medium, 1 mM L-glutamine, 10 IU/ml penicillin/streptomycin, and 10% fetal bovine serum, and were immortalized by routine serial passage. To detect endogenous OGN protein in MEFs, wild-type or Bmp1/Tll1 doubly homozygous null cells at 80% confluence were washed twice with phosphate-buffered saline, and incubated in serum-free Dulbecco’s modified Eagle’s medium for 15 min at 37 °C. Cells were then treated with or without 2 ng/ml TGF-␤1 (R&D Systems) in serum-free Dulbecco’s modified Eagle’s medium for 48 h. Proteins in conditioned media were ethanol precipitated and resuspended in 0.5% SDS, 1% 2-mercaptoethanol. Deglycosylation of precipitated media proteins with ␤134 galactosidase was performed according to the manufacturer’s protocol (Sigma). RESULTS

Expression, Purification, and Characterization of Recombinant OGN—The murine full-length pre-OGN protein contains

298 amino acid residues that include a signal peptide, six cysteine residues; four within the NH2- and two within the COOH-terminal domains, respectively; six leucine-rich repeats in the central domain, and a single potential N-glycosylation site at amino acid residue 258 (Fig. 1A). To generate recombinant OGN using a baculovirus expression system, a transfer vector, pOGN, was constructed using a cDNA encoding the core protein of murine OGN, such that a fusion protein would be produced corresponding to the acidic glycoprotein gp67 signal sequence, followed by six consecutive histidine residues, a thrombin cleavage site, and the full OGN coding sequence (Fig. 1B). Recombinant baculovirus vOGN was then generated by homologous recombination between transfer vector pOGN and linearized baculoGold virus DNA. Sf9 cells infected with vOGN were incubated in serum-free medium and 96 h later culture medium was harvested and recombinant His-tagged protein

BMP-1/Tolloid-like Proteinases Cleave Osteoglycin

FIG. 2. Mass spectrometry and peptide N-glycosidase F digestion of recombinant OGN. A, the mass of recombinant OGN, dialyzed overnight against 10 mM NH4HCO3 at 4 °C, was determined by MALDITOF mass spectrometry. B, recombinant OGN was analyzed by SDSPAGE and Coomassie Blue staining without further treatment (lane 2) or after incubation with peptide N-glycosidase F (lane 3). N-Glycosidase F enzyme appears as a band of higher mobility than OGN in lanes 3 and 4, (marked with an asterisk).

was purified by a single passage over a nickel chelating column. The concentration of purified samples was estimated using the theoretical molar extinction coefficient (39) and analyzed by SDS-PAGE and staining with Coomassie Blue (Fig. 1C). Recombinant protein was predominantly eluted from the nickel column at 100 mM imidazole and no contaminating protein bands were visualized in this fraction via SDS-PAGE, indicating a high degree of purity for the sample. The yield of recombinant OGN was ⬃5 mg/3.5 ⫻ 109 cells per 96 h. Identity of the recombinant protein was confirmed by Western blot analysis (Fig. 1D) using a polyclonal antiserum specific for mouse OGN, with detection of a single band of recombinant OGN (Fig. 1D, lane 3), and no signal detected from samples derived from uninfected control cells (Fig. 1D, lane 1) or from control cells infected with wild-type baculovirus AcNMPV (Fig. 1D, lane 2). Recombinant OGN appeared on SDS-PAGE, under reducing conditions as a single ⬃39-kDa band (Figs. 1C and 2), whereas the expected mass of murine OGN is 33,245 Da, based on the primary amino acid sequence. MALDI-TOF MS yielded a peak with a mass of 34,557 Da for recombinant OGN (Fig. 2A). One possible explanation for the mass difference between MALDITOF MS results and the expected mass might have been glycosylation of recombinant OGN with the N-linked high mannose oligosaccharides characteristic of insect cells. To test this possibility, recombinant OGN was incubated with peptide Nglycosidase F, which cleaves between the innermost N-acetylglucosamine and asparagine residues of N-linked high mannose oligosaccharides. Upon enzyme digestion, OGN protein underwent a shift to a higher mobility form (Fig. 2B, lanes 2 and 3). Thus, the results of mass spectrometry and peptide N-glycosidase F treatment are both consistent with the conclusion that the recombinant OGN produced in the insect cell system for our studies is secreted as a glycoprotein with high mannose N-linked carbohydrate and few, if any, additional post-translational modifications.

41629

FIG. 3. BMP-1 as a candidate enzyme for the proteolytic removal of the NH2-terminal prodomain of OGN. Alignment of the sites at which BMP-1 has previously been shown to cleave the Cpropeptides of the types I-III, pro-␣2(V), and pro-␣1(VII) procollagen chains (6, 13, 21), chordin (5), prolysyl oxidase (10), dentin matrix protein-1 (23), myostatin (25), and biglycan (31) with the murine OGN BMP-1 cleavage site, determined by NH2-terminal amino acid sequencing of the 27-kDa fragment produced by BMP-1 cleavage of recombinant murine pro-OGN, and with the in vivo cleavage site that produces the mature 27-kDa OGN extracted from bovine cornea (33). Aspartate residues conserved at the P1⬘ positions of the various scissile bones are in boldface, as are methionines and residues with aromatic side chains, previously noted NH2-terminal to scissile bonds in the majority of previously identified substrates of BMP-1-like proteinases.

OGN as a Candidate Substrate of BMP-1 and Related Metalloproteinases—The ⬃25-kDa KSPG form of OGN that predominates in adult bovine cornea is a cleavage product resulting from proteolytic processing at the site shown in Fig. 3, which removes a third of the full OGN translation product (33). The corresponding sequences of murine OGN are aligned with the bovine sequences in Fig. 3. Interestingly, the site at which bovine OGN is cleaved in vivo, and aligned murine OGN sequences, resemble sites at which BMP-1 and related mammalian metalloproteinases mTLD, mTLL-1, and mTLL-2 cleave a number of proteins previously demonstrated to be in vivo substrates of these proteinases (Fig. 3). This observation, plus the fact that the class I SLRP biglycan is one of the previously identified substrates of the BMP-1-like proteinases, made OGN a reasonable candidate as a possible substrate of the same small family of proteinases. BMP-1 Processes the N-prodomain of Pro-OGN at the Physiologically Relevant Site—To determine whether OGN might be processed by BMP-1-like proteinases, purified recombinant OGN was incubated with purified recombinant BMP-1 for 20 h at 37 °C. Analysis by SDS-PAGE and staining with Coomassie Blue showed that OGN (⬃39 kDa) was cleaved by BMP-1 to produce fragments of ⬃27 and ⬃17 kDa (Fig. 4). To determine which other of the mammalian BMP-1-related proteinases might cleave OGN, all four (BMP-1, mTLD, mTLL-1, and mTLL-2) were incubated with OGN, and processing was assayed by Western blot analysis, using antibodies directed against the NH2-terminal His tag, and antibodies directed against sequences in the mature portion of OGN (Fig. 5). As can be seen, BMP-1, mTLD, and mTLL-1 all efficiently processed OGN under the conditions of the in vitro assay, whereas mTLL-2 had somewhat lesser levels of OGN processing activity. As can be seen in the Western blots of Fig. 5, antibodies directed against sequences in the mature portion of OGN recognized the 27-kDa cleavage product. However, they did not recognize the 17-kDa cleavage product. This suggested that the

41630

BMP-1/Tolloid-like Proteinases Cleave Osteoglycin

FIG. 4. BMP-1 cleaves pro-OGN to produce two detectable fragments. Electrophoretic patterns on an SDS-PAGE gel are compared for recombinant pro-OGN incubated in the absence (⫺) or presence (⫹) of BMP-1. MW denotes molecular mass markers, with approximate sizes marked in kDa. Pro-OGN, Mat, and Propep denote proOGN, mature OGN, or the cleaved OGN prodomain, respectively. Bands were visualized by staining with Coomassie Blue.

FIG. 5. Cleavage assays of recombinant pro-OGN incubated separately with the four mammalian BMP-1-related proteinases. Western blots are shown of recombinant pro-OGN incubated alone (No enzyme), or in the presence of BMP-1 (BMP1), mTLD (TLD), mTLL-1 (TLL1), or mTLL-2 (TLL2). S.M. denotes starting material. Immunoblots employed either antibodies directed against the NH2terminal His tag (␣ His), or antibodies directed against sequences within mature OGN (␣ OGN). ␣ His antibodies detected only 39-kDa pro-OGN, whereas ␣ OGN antibodies detected both 39-kDa pro-OGN and the 27-kDa OGN cleavage product.

17-kDa product represented a cleaved N-prodomain of a proform of OGN, and that the 27-kDa product represented mature OGN. To determine whether the 27-kDa protein indeed represented mature OGN and, if so, to determine the site of cleavage by BMP-1-like proteinases, 2 ␮g of OGN was incubated with 50 ng of BMP-1, and the 27-kDa fragment was isolated and subjected to automated Edman degradation for determination of NH2-terminal amino acid sequences. Ten cycles of Edman degradation produced the sequence DEVIPSLTK, which corresponds to the NH2 terminus of mature keratan sulfate-modified OGN isolated from bovine cornea (33) (Fig. 3). The OGN cleavage site in both species resembles the cleavage sites of various previously identified substrates of BMP-1-like proteinases, in having an Asp at the P1⬘ position. However, although the bovine sequence, like the sequences of various previously identified substrates of BMP-1-like proteinases, has a residue with an aromatic side chain within 5 residues NH2-terminal to the cleavage site, this Phe is not conserved in the murine sequence (Fig. 3). A Product or Products of the Bmp1 and Tll1 Genes Are Responsible for all Detectable Pro-OGN Processing by MEFs— Previously, comparison of cleavage patterns of proteins in cultures of wild-type MEFs, to cleavage patterns of proteins in

FIG. 6. Absence of proteolytic processing of pro-OGN in cultures of MEF from embryos doubly homozygous null for the Bmp1 and Tll1 genes. Western blot analysis was employed, using antibodies specific for sequences within the mature region of OGN, to monitor pro-OGN processing within cultures of wild-type (WT) MEF or MEF doubly homozygous null (Bmp1/Tll1 null) for the Bmp1 and Tll1 genes. Cultures were either treated (⫹) or not treated (⫺) with 2 ng/ml TGF-␤. ␣ OGN antibodies detected the 39-kDa pro-OGN and the 27kDa OGN cleavage products in TGF-␤-treated WT MEF cultures, but only detected 39-kDa pro-OGN in TGF-␤-treated Bmp1/Tll1 null MEF cultures.

cultures of MEFs derived from embryos homozygous null for both the Bmp1 gene, which encodes alternatively spliced mRNAs for Bmp1 and mTLD (3), and the Tll1 gene, which encodes mTLL-1, has led to identification and verification of a number of in vivo substrates of the mammalian BMP-1-like proteinases (9, 10, 13, 23, 31). In the present study, OGNrelated bands were only faintly detected on immunoblots of proteins from the media of wild-type and Bmp1/Tll1 doubly null MEFs grown under normal growth conditions (Fig. 6). However, TGF-␤, which induces increased levels of expression of both ECM components (40) and BMP-1-like proteinases (41), induced levels of OGN expression to the extent that OGN was easily visualized in immunoblots of MEF media proteins (Fig. 6). Under such conditions, OGN in wild-type MEF media is detected primarily as the cleaved fully mature 27-kDa form, with lesser amounts of the full-length 39-kDa pro-OGN form. In contrast, OGN in the media of Bmp1/Tll1 doubly homozygous null MEFs is only detectable as the full-length 39-kDa pro-OGN species, with no evidence of processed OGN (Fig. 6). Pro- and Mature OGN Affect Collagen Fibrillogenesis Differently—Characterization of mice with null alleles for various SLRP genes has suggested that some SLRPs play roles in the regulation of collagen fibrillogenesis in vivo (32, 36, 42– 46). To begin testing for possible effects of the different forms of OGN on collagen fibrillogenesis, increasing amounts of pro-OGN were added to an in vitro type I collagen fibrillogenesis assay. As can be seen (Fig. 7A), 1 ␮g ml⫺1 pro-OGN was found to retard the rate of fibrillogenesis, but not to reduce the final turbidity of the reaction. However, 5–30 ␮g ml⫺1 pro-OGN was found to retard the rate of fibrillogenesis and to dramatically decrease the final turbidity of the reaction (Fig. 7A). These results indicated that pro-OGN is able to inhibit fibrillogenesis of type I collagen in a concentration-dependent manner. However, a concern was that the additional His6 tag at the pro-OGN NH2 terminus might affect the structure/function of this protein. To control for this possibility the His tag was removed via thrombin cleavage at an adjacent site, and the effects of proOGN with or without His tag were compared in an in vitro fibrillogenesis assay. Results demonstrated that pro-OGN has the same effects on retarding the rate and reducing the final turbidity in the type I collagen fibrillogenesis assay, regardless of the presence or absence of the His tag (Fig. 7B). We next tested whether processing of pro-OGN to mature OGN by BMP-1-like proteinases might influence the effects of this SLRP on fibrillogenesis. Toward this end, pro-OGN and mature OGN were separately added to the in vitro fibrillogenesis assay and their effects on type I collagen fibrillogenesis were compared. Both pro-OGN and mature OGN were found to

BMP-1/Tolloid-like Proteinases Cleave Osteoglycin

41631

FIG. 8. Comparison of the effects of pro- and mature OGN on type I collagen fibrillogenesis. Pro-OGN and mature OGN (open and closed circles, respectively) differ in their effects on type I collagen fibrillogenesis. Squares, type I collagen in the absence of added SLRPs. Collagen was at a concentration of 0.5 mg ml⫺1 and SLRP concentrations were 10 ␮g ml⫺1, with reactions performed at 37 °C. Absorbance was measured at 313 nm.

FIG. 7. Effects of pro-OGN on type I collagen fibrillogenesis. A, pro-OGN retards fibrillogenesis and reduces final turbidity of the fibrillogenesis reaction in a concentration-dependent manner. The turbiditytime curves shown are for experiments run with collagen alone, at a concentration of 0.5 mg ml⫺1, or in the presence of His-tagged pro-OGN at concentrations of 1 (open circles), 5 (open triangles), 10 (closed triangles), or 30 (closed circles) ␮g ml⫺1 at 37 °C. B, pro-OGN with and without a His tag (open and closed circles, respectively) similarly inhibit the kinetics of fibrillogenesis and decrease final turbidity of the reaction. Turbidity-time curves were run with collagen alone (squares) at a concentration of 0.5 mg ml⫺1, or in the presence of pro-OGN at a concentration of 5 ␮g ml⫺1, at 37 °C. Absorbance was measured at 313 nm.

retard the rate of fibrillogenesis and to reduce final turbidity of the reaction (Fig. 8). However, mature OGN was markedly more effective in reducing final turbidity than was pro-OGN. An additional observation was that fibrillogenesis appeared to have a shorter lag time in the presence of mature OGN, compared with fibrillogenesis of type I collagen in the presence of pro-OGN, or in the absence of additional proteins. To compare the effects of pro-OGN and mature OGN on fibrillogenesis with the effects of previously characterized SLRPs, mature biglycan was added to the same type of fibrillogenesis assay, and was found to potentiate the rate of fibril formation (Fig. 9). In contrast, mature decorin was found to markedly delay the rate of fibrillogenesis and to reduce final turbidity levels, as previously reported (47). The shapes of the various curves were reproducible for the various SLRPs over

FIG. 9. Comparison of the roles of different SLRPs on collagen fibrillogenesis. Mature BGN (open circles) and mature DCN (open triangles) were used as controls, and for comparison to the role of mature OGN (closed circles) on collagen fibrillogenesis. Squares, fibrillogenesis of type I collagen in the absence of added SLRPs. Collagen was at a concentration of 0.5 mg ml⫺1 and SLRP concentrations were 10 ␮g ml⫺1, with reactions performed at 37 °C and absorbance measured at 313 nm.

the course of five experiments, and in each case there was a pronounced increase in the ability of OGN to inhibit fibrillogenesis compared with the ability of pro-OGN to exert an effect in the same assay. The data thus indicate that processing of the pro-peptide region has a marked effect on the ability of OGN to regulate the rate of collagen fibrillogenesis. DISCUSSION

Multiple OGN forms of different sizes have previously been extracted from different tissues. Specifically, a 12-kDa glycoprotein corresponding to the 105 most COOH-terminal amino acids of OGN, and previously designated osteoinductive factor, has been isolated from bovine bone (33, 48, 49); and a 25-kDa keratan sulfate proteoglycan, designated mimecan and corre-

41632

BMP-1/Tolloid-like Proteinases Cleave Osteoglycin

sponding to the 223 most COOH-terminal amino acids of OGN, has been isolated from bovine cornea (33, 50, 51). In addition, a 44 – 46-kDa glycoprotein has been detected by Western blot analysis with an anti-OGN monoclonal antibody in bovine tissues such as aorta, cartilage, sclera, and skin (33, 50, 51). However, although the possibilities of alternatively spliced mRNAs and/or proteolytic processing have been suggested as possible mechanisms by which the differently sized forms of OGN might arise in vivo, neither the mechanism(s) by which the various sized species of OGN are formed, nor the possible biological significance underlying the formation of various forms have been ascertained. In fact, the biology, including the biosynthesis and functions, of the class III SLRPs has remained largely uncharacterized. Here we have begun such characterization of OGN via the successful production, purification, and analysis of recombinant murine OGN. Most notably, we have succeeded in further expanding the range of known activities for the mammalian BMP-1/Tolloid-like proteinases, by demonstrating that such metalloproteinases are capable of processing OGN at a physiologically relevant site, and at no other site, and that such proteinases are responsible for this activity in cells. Previously, we have shown the mammalian BMP-1-like proteinases to play key roles in formation of the ECM via biosynthetic processing of precursors to produce the mature functional forms of collagens I–III, V, VII, and XI, lysyl oxidase, and laminin 5 (5–10, 12–14, 19 –21). We have also previously shown that dentin matrix protein-1 is processed by BMP-1-like proteinases in a manner that suggests that such proteinases play important roles in the mineralization of ECM in hard tissues (23). In addition, we have previously demonstrated that a 21-amino acid NH2-terminal prodomain of the class I SLRP biglycan is proteolytically removed by the BMP-1-like proteinases, although the biological significance of this event remains to be defined (31). The current study shows that biglycan is not unique among SLRPs in being synthesized as a precursor form with an NH2-terminal prodomain that is cleaved by the BMP1-like proteinases. Rather, we demonstrate that the class III SLRP OGN has a 75-amino acid prodomain that is cleaved by the same enzymes. Because the processing by BMP-1-like proteinases of previously identified substrates involved in formation of the ECM appears to involve the processing of precursors to mature functional forms of the molecules, we suggest that it is likely that processing of biglycan and OGN by these enzymes may play similar roles. Previous studies employing antibodies have shown that removal of the NH2-terminal prodomain of biglycan induces conformational changes that affect the availability of epitopes in both the NH2- (52) and COOH-terminal (53) portions of the mature protein. Here we show that conversion of OGN from the precursor to mature form, through cleavage by BMP-1, has profound effects on OGN biological activity, and it may be that such a change in activity reflects a conformational change in the protein. The ramifications of prodomain removal have not been ascertained for the previously reported activities of biglycan in binding TGF-␤ and possibly interacting with various collagen types. Nevertheless, proteolytic removal of the NH2terminal prodomains from biglycan and OGN by BMP-1-like proteinases appears to have significant effects on the properties of both proteins. Inactivation of individual SLRP genes has resulted in knockout mice, which present with relatively mild phenotypes that involve improper organization of collagen fibers, thus demonstrating roles for SLRPs in collagen fibrillogenesis (32, 36, 42– 46). In particular, mice homozygous null for the gene that encodes OGN showed both an increased collagen fibril diame-

ter and an increase in the variability of collagen fibril diameters compared with wild type (36). The latter result, plus preliminary data showing that recombinant OGN binds to collagen types I, II, and III,2 which are abundant in tissues expressing OGN, prompted us to test for a possible role for OGN in collagen fibrillogenesis. Interestingly, although both pro-OGN and mature OGN were shown to hinder the rate of fibrillogenesis, mature OGN had a markedly greater effect in inhibiting the final levels of turbidity in the assay. The effects of both pro- and mature OGN on fibrillogenesis were most pronounced in the growth and plateau phases of turbidity, in which the lateral growth of fibrils plays a major role (54), and were less pronounced in affecting the duration of the lag phase at the beginning of the assay, during which there is no detectable change in turbidity. Because the lag phase is thought to involve primarily linear elongation of thin fibrils (54), the results thus suggest that both pro- and mature OGN act primarily to limit lateral accretion of collagen monomers to growing fibrils. Indeed, similar inhibition in fibrillogenesis assays by decorin, as previously reported (47) and as recreated here, has been shown to result in fibrils markedly smaller in diameter than those formed by collagen in the absence of decorin (55). The biochemical results presented here are thus consistent with the previous finding of thicker collagen fibrils in tissues of mice homozygous null for the gene that encodes OGN (36), and demonstrate direct roles for pro- and mature OGN in regulating fibrillogenesis. It should be noted that proteins produced in a baculovirus system have high mannose Asn-linked saccharides, rather than the complex Asn-linked glycosylation found in proteins produced by mammalian cells (56). Thus, because recombinant OGN forms used in the current study were produced in a baculovirus system, they are differently glycosylated than OGN forms from mammalian tissues. High mannose Asnglycosylation may increase affinity to collagen (57). However, because both pro-OGN and mature OGN in the present study are identically glycosylated, differences in their effects on fibrillogenesis are manifestly because of BMP-1 cleavage, and are glycosylation-independent. The finding that mature OGN is a more potent regulator of fibrillogenesis than is pro-OGN marks another role that the BMP-1-like proteinases appear to play in regulating formation of the ECM in general and collagen fibrillogenesis in particular. The fact that pro-OGN is itself somewhat functional in regulating fibrillogenesis indicates that its conversion to the mature form by BMP-1-like proteinases does not represent an “on switch” that permits fibrillogenesis to proceed, but rather that this conversion may represent fine-tuning of fibrillogenesis regulation affecting fibril diameter. However, it seems quite possible that the proteolytic processing of pro-OGN and its effects on OGN secondary structure affect OGN interactions with other proteins, some of which may also affect fibrillogenesis in vivo. Thus, the processing of pro-OGN to mature OGN may have more profound effects on collagen fibrillogenesis in vivo than in vitro. Additional effects that the processing of OGN by BMP-1-like proteinases may have on protein-protein interactions and functional properties of OGN remain to be determined. Acknowledgments—We are grateful to Dr. Mary Estes, Dr. Carl Zeng, and Sue Crawford (Department of Molecular Virology and Microbiology, Baylor Collage of Medicine) for technical support in the production of recombinant baculoviruses. We thank Dr. David McQuillan and Dr. Rick Owens (LifeCell Corporation) for thoughtful comments on the collagen fibrillogenesis assay.

2

N.-S. Seo and M. Ho¨ o¨ k, unpublished data.

BMP-1/Tolloid-like Proteinases Cleave Osteoglycin REFERENCES 1. Bond, J. S., and Beynon, R. J. (1995) Protein Sci. 4, 1247–1261 2. Greenspan, D. S. (2005) Top. Curr. Chem. 229, in press 3. Takahara, K., Lyons, G. E., and Greenspan, D. S. (1994) J. Biol. Chem. 269, 32572–32578 4. Takahara, K., Brevard, R., Hoffman, G. G., Suzuki, N., and Greenspan, D. S. (1996) Genomics 34, 157–165 5. Scott, I. C., Blitz, I. L., Pappano, W. N., Imamura, Y., Clark, T.G., Steiglitz, B. M., Thomas, C. L., Maas, S. A., Takahara, K., Cho, K. W. Y., and Greenspan, D. S. (1999) Dev. Biol. 213, 283–300 6. Kessler, E., Takahara, K., Biniaminov, L., Brusel, M., and Greenspan, D. S. (1996) Science 271, 360 –362 7. Li, S., Sieron, A. L., Fertala, A., Hojima, Y., Arnold, W. V., and Prockop, D. J. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 5127–5130 8. Suzuki, N., Labosky, P. A., Furuta, Y., Hargett, L., Dunn, R., Fogo, A. B., Takahara, K., Peters, D. M., Greenspan, D. S., and Hogan, B. L. (1996) Development 122, 3587–3595 9. Pappano, W. N., Steiglitz, B. M., Scott, I. C., Keene, D. R., and Greenspan, D. S. (2003) Mol. Cell. Biol. 23, 4428 – 4438 10. Uzel, M. I., Scott, I. C., Babakhanlou-Chase, H., Palamakumbura, A. H., Pappano, W. N., Hong, H.-H., Greenspan, D. S., and Trackman P. C. (2001) J. Biol. Chem. 276, 22537–22543 11. Kagan, H., and Trackman, P. C. (1991) Am. J. Respir. Cell Mol. Biol. 5, 206 –210 12. Imamura, Y., Steiglitz, B. M., and Greenspan, D. S. (1998) J. Biol. Chem. 42, 27511–27517 13. Unso¨ ld, C., Pappanoo, W. N., Imamura, Y., Steiglitz, B. M., and Greenspan, D. S. (2002) J. Biol. Chem. 277, 5596 –5602 14. Medeck, R. J., Sosa, S., Morris, N., and Oxford, J. T. (2003) Biochem. J. 376, 361–368 15. Birk, D. E., Fitch, J. M., Babiarz, J. P., and Linsenmayer, T. G. (1988) J. Cell Biol. 106, 999 –1008 16. Birk, D. E., Fitch, J. M., Babiarz, J. P., Doane, K. J., and Linsenmayer, T. F. (1990) J. Cell Sci. 95, 649 – 657 17. Toriello, H. V., Glover, T. W., Takahara, K., Byers, P. H., Miller, D. E., Higgins, J. V., and Geenspan, D. S. (1996) Nat. Genet. 13, 361–365 18. Wenstrup, R. J., Langland, G. T., Willing, M. C., D-Souza, V. N., and Cole, W. G. (1996) Hum. Mol. Genet. 5, 1733–1736 19. Amano, S., Scott, I. C., Takahara, K., Koch, M., Champliaud, M.-F., Gerecke, D. R., Keene, D. R., Hudson, D. L., Nishiyama, T., Lee, S., Greenspan, D. S., and Burgeson, R. E. (2000) J. Biol. Chem. 275, 22728 –22735 20. Veitch, D. P., Nokelainen, P., McGowan, K. A., Nguyen, T.-T., Nguyen, N. E., Stephenson, R., Pappano, W. N., Keene, D. R., Spong, S. M., Greenspan, D. S., Findell, P. R., and Marinkovich, M. P. (2003) J. Biol. Chem. 278, 15661–15668 21. Rattenholl, A., Pappano, W. N., Koch, M., Keene, D. R., Kadler, K. E., Sasaki, T., Timpl, R., Burgeson, R. E., Greenspan, D. S., and Bruckner-Tuderman, L. (2002) J. Biol. Chem. 277, 26372–26378 22. Burgeson, R. E., and Christiano, A. M. (1997) Curr. Opin. Cell Biol. 9, 651– 658 23. Steiglitz, B. M., Ayala, M., Narayanan, K., George, A., and Greenspan, D. S. (2004) J. Biol. Chem. 279, 980 –986 24. Butler, W. T. (1998) Eur. J. Oral Sci. 106, Suppl. 1, 204 –210 25. Wolfman, N., McPherron, A. C., Pappano, W. N., Davies, M. V., Song, K., Tomkinson, K. N., Wright, J. F., Zhao, L., Sebald, S. M., Greenspan, D. S., and Lee, S.-J. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, 15842–15846 26. Iozzo, R. V. (1998) Annu. Rev. Biochem. 67, 609 – 652 27. Hocking, A. M., Shinomura, T., and McQuillan, D. J. (1998) Matrix Biol. 17, 1–19 28. Iozzo, R. V. (1999) J. Biol. Chem. 274, 18843–18846

41633

29. Henry, S. P., Takanosu, M., Boyd, T. C., Mayne, P. M., Eberspaecher, H., Zhou, W., de Crombrugghe, B., Ho¨ o¨ k, M., and Mayne, R. (2001) J. Biol. Chem. 276, 12212–12221 30. Lorenzo, P., Aspberg, A., Onnerfjord, P., Bayliss, M. T., Neame, P. J., and Heinegard, D. (2001) J. Biol. Chem. 276, 12201–12211 31. Scott, I. C., Imamura, Y., Pappano, W. N., Troedel, J. M., Recklies, A. D., Roughley, P. J., and Greenspan, D. S. (2000) J. Biol. Chem. 275, 30504 –30511 32. Xu, T., Bianco, P., Fisher, L. W., Longenecker, G., Smith, E., Goldstein, S., Bonadio, J., Boskey, A., Heegaard, A.-M., Sommer, B., Satomura, K., Dominguez, P., Chengyan, Z., Kulkarni, A. B., Robey, G., and Young, M. F. (1998) Nat. Genet. 20, 78 – 82 33. Funderburgh, J. L., Corpuz, L. M., Roth, M. R., Funderburgh, M. L., Tasheva, E. S., Conrad, G. W. (1997) J. Biol. Chem. 272, 28089 –28095 34. Johnson, H. K., Rosenberg, L., Choi, H. U., Garza, S., Ho¨ o¨ k, M., and Neame, P. J. (1997) J. Biol. Chem. 272, 18709 –18717 35. Johnson, J., Shinomura, T., Eberspaecher, H., Pinero, G., Decrombrugghe, B., and Ho¨ o¨ k, M. (1999) Dev. Dyn. 216, 499 –510 36. Tasheva, E. S., Koester, A., Paulsen, A. Q., Garrett, A. S., Boyle, D. L., Davidson, H. J., Song, M., Fox, N., and Conrad, G. W. (2002) Mol. Vis. 8, 407– 415 37. Laemmli, U. K. (1970) Nature 227, 680 – 685 38. Birk, D. E., and Silver, F. H. (1984) Arch. Biochem. Biophys. 235, 178 –185 39. Pace, C. N., Vajdos, F., Fee, L., Grimsley, G., and Gray, T. (1995) Protein Sci. 4, 2411–2423 40. Massague´ , J. (1990) Annu. Rev. Cell Biol. 6, 597– 641 41. Lee, S., Solow-Cordero, D. E., Kessler, E., Takahara, K., and Greenspan, D. S. (1997) J. Biol. Chem. 272, 19059 –19066 42. Danielson, K. G., Baribault, H., Holmes, D. F., Graham, H., Kadler, K. E., and Iozzo, R. V. (1997) J. Cell Biol. 136, 729 –743 43. Svensson, L., Aszodi, A., Reinholt, F. P., Fassler, R., Heinegard, D., and Oldberg, A. (1999) J. Biol. Chem. 274, 9636 –9647 44. Chakravarti, S., Magnuson, T., Lass, J. H., Jepsen, K. J., LaMantia, C., and Carroll, H. (1998) J. Cell Biol. 141, 1277–1286 45. Liu, C. Y., Birk, D. E., Hassell, J. R., Kane, B., and Kao, W. W. (2003) J. Biol. Chem. 278, 21672–21677 46. Ameye, L., and Young, M. F. (2002) Glycobiology 12, 107R–116R 47. Vogel, K. G., and Trotter, J. A. (1987) Collagen Relat. Res. 7, 105–114 48. Bentz, H., Nathan, R. M., Rosen, D. M., Armstrong, R. M., Thompson, A. Y., Segarini, P. R., Mathews, M. C., Dasch, J. R., Piez, K. A., and Seyedin, S. M. (1989) J. Biol. Chem. 264, 20805–20810 49. Madisen, L., Neubauer, M., Plowman, G., Rosen, D., Segarini, P., Dasch, J., Thompson, A., Ziman, J., Bentz, H., and Purchio, A. F. (1990) DNA Cell Biol. 9, 303–309 50. Funderburgh, J. L., and Conrad, G. W. (1990) J. Biol. Chem. 265, 8297– 8303 51. Funderburgh, J. L., Funderburgh, M. L., Mann, M. M., and Conrad, G. W. (1991) J. Biol. Chem. 266, 14226 –14231 52. Yeo, T.-K., Torok, M. A., Kraus, H. L., Evans, S. A., Zhou, Y., and Marcum, J. A. (1995) J. Vasc. Res. 32, 175–182 53. Roughley, P. J., White, R. J., and Mort, J. S. (1996) Biochem. J. 318, 779 –784 54. Gelman, R. A., Williams, B. R., and Piez, K. A. (1979) J. Biol. Chem. 254, 180 –186 55. Vogel, K. G., Paulsson, M., and Heinegård, D. (1984) Biochem. J. 223, 587–597 56. Jarvis, D. L., and Summers, M. D. (1992) in Recombinant DNA Vaccines: Rationale and Strategies (Isaacson, R. E., ed) pp. 265–291, Dekker, New York 57. Kaufmann, B., Mu¨ ller, S., Hanisch, F.-G., Hartmann, U., Paulsson, M., Maurer, P., and Zaucke, F. (2004) Glycobiology 14, 609 – 619