Constitutive Release of α4 Type V Collagen N-terminal Domain by Schwann Cells and Binding to Cell Surface and Extracellular Matrix Heparan Sulfate Proteoglycans*

During peripheral nerve development, Schwann cells synthesize collagen type V molecules that contain α4(V) chains. This collagen subunit possesses an N-terminal domain (NTD) that contains a unique high affinity heparin binding site. The α4(V)-NTD is adhesive for Schwann cells and sensory neurons and is an excellent substrate for Schwann cell and axonal migration. Here we show that the α4(V)-NTD is released constitutively by Schwann cells both in culture and in vivo. In cultures of neonatal rat Schwann cells, α4(V)-NTD release is increased significantly by ascorbate treatment, which facilitates collagen post-translational modification and collagen trimer assembly. In peripheral nerve tissue, the α4(V)-NTD is localized to the region of the outer Schwann cell membrane and associated extracellular matrix. The released α4(V)-NTD binds to the cell surface and extracellular matrix heparan sulfate proteoglycans of Schwann cells. Pull-down assays and immunofluorescent staining showed that the major α4(V)-NTD-binding proteins are glypican-1 and perlecan. α4(V)-NTD binding occurs via a mechanism that requires the high affinity heparin binding site and that is blocked by soluble heparin, demonstrating that binding to proteoglycans is mediated by their heparan sulfate chains.

The development of the peripheral nervous system is critically dependent on the migration and proliferation of Schwann cells (1)(2)(3). These processes are needed to provide sufficient numbers of glial cells to ensheathe and myelinate peripheral axons. Schwann cell proliferation and migration are regulated by molecular signals derived from peripheral axons as well as the extracellular matrix (ECM) 1 that invests peripheral nerve tracts. A major source of the peripheral nerve ECM is Schwann cells (4).
The structure and composition of the ECM are modified during nerve development. These modifications contribute to the modulation of cellular functions that are regulated by ECM contact. During the late embryonic period when Schwann cell migration and proliferation are most active, the ECM consists mainly of a fibrillar matrix component (4). During the transition to axonal ensheathement and myelination, Schwann cell basal laminae appear. Previous studies (5,6) have shown that contact with the basal lamina triggers myelination of axons by Schwann cells.
A prominent component of the developing nerve ECM is ␣4 type V collagen (7). A high level of expression of the ␣4(V) polypeptide is restricted to only a few tissues, including developing peripheral nerve, suggesting that this collagen chain carries out a unique function (8). Native ␣4(V) polypeptides are secreted by Schwann cells as triple helical collagen heterotrimers that also contain the ubiquitously expressed ␣1(V) and ␣2(V) collagen chains (8). The ␣4(V) collagen polypeptide consists of a large central collagen domain that is flanked by a noncollagenous N-terminal domain (NTD) of 475 amino acids (including a signal peptide of 29 amino acids) and a C-terminal domain of 251 amino acids. The ␣4(V)-NTD contains a high affinity heparin binding site that is not present in other type V collagen polypeptides (9). The heparin binding site consists of multiple repeats of the consensus heparin binding sequence YYXY (where Y is lysine or arginine).
Studies (9 -11) of the function of ␣4(V) collagen molecules have revealed diverse effects on Schwann cell adhesion and axonal migration, which are mediated by distinct protein domains. The collagen domain inhibits axonal outgrowth and Schwann cell migration and blocks the adhesion and migration-promoting activities of other ECM proteins, such as collagen type IV (10). The ␣4(V)-NTD, in contrast, promotes axonal outgrowth and Schwann cell migration and is an excellent substrate for Schwann cell adhesion and spreading. These effects of the NTD are dependent on its heparin binding activity and are blocked by soluble heparin or heparan sulfate but not by function-blocking anti-integrin antibodies (9). These findings suggest that Schwann cell interaction with ␣4(V)-NTD is mediated by heparan sulfate proteoglycans. This conclusion is supported by the observation that syndecan-3, a Schwann cell transmembrane heparan sulfate proteoglycan, binds in vitro to ␣4(V)-NTD by a heparan sulfate-dependent mechanism (11).
This study extends these findings and presents data on the processing of the ␣4(V) collagen chain in Schwann cell cultures and in developing peripheral nerve tissue. The ␣4(V)-NTD is released constitutively by a proteolytic mechanism that is enhanced by collagen assembly. The released ␣4(V)-NTD accumulates in Schwann cell cultures and rat peripheral nerve tissue. ␣4(V)-NTD binds to the Schwann cell plasma membrane and ECM via binding to the heparan sulfate proteoglycans, glypican-1 and perlecan.

MATERIALS AND METHODS
Schwann Cell Cultures-Schwann cells were isolated from newborn rat sciatic nerves as described previously (12). The cells were cultured on poly-L-lysine-coated culture dishes in Dulbecco's modified Eagle's * This work was supported by National Institutes of Health Grant NS21925. 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.
‡ To whom correspondence should be addressed: Weis Center for Research, Geisinger Clinic, 100 N. Academy Ave., Danville, PA 17822-2601; Tel.: 570-271-6659; Fax: 570-271-5886. E-mail: djcarey@ geisinger.edu. 1 The abbreviations used are: ECM, extracellular matrix; NTD, Nterminal domain; PBS, phosphate-buffered saline; PI-PLC, phosphatidylinositol-specific phospholipase C; P, postnatal day (e.g. P2); ⌬HBS, mutated ␣4(V)-NTD that lacks the heparin binding site. medium with 10% fetal bovine serum and 2 M forskolin. To investigate the effects of ascorbate on ␣4(V) collagen, Schwann cells were grown to confluence and then L-ascorbic acid (Sigma) was added at a final concentration of 50 g/ml. Ascorbate was replenished after 24 h. After 48 h, the medium was harvested, centrifuged to remove unattached cells, and stored at Ϫ80°C. Cell extracts were prepared by washing the cells twice with PBS (150 mM NaCl, 50 mM sodium phosphate, pH 7.4) and then scraping them into 300 l of electrophoresis sample buffer with 2% SDS. Extracts were heated at 100°C for 10 min and stored at Ϫ80°C.
Preparation of Anti-collagen Type V Antibodies-The preparation of antibodies against the ␣4(V) collagen chain and ␣1(V) N-terminal domain was described previously (8,11). Anti-␣4(V)-NTD antibodies were also generated. The recombinant His-tagged ␣4(V)-NTD was expressed in BL23-pLys-S Escherichia coli cells (Novagen) that were transformed with pET-30a(ϩ) vector containing the NTD coding sequence. The protein was purified by chromatography on His-Bind resin (Novagen) followed by heparin-agarose chromatography (Sigma). The purified protein was used to immunize New Zealand White rabbits as described previously (11). Antibodies were affinity-purified on a column of ␣4(V)-NTD coupled to Affi-Gel-15 (Bio-Rad) by elution with 100 mM glycine, pH 2.5.
Expression of Myc-tagged ␣4(V) Collagen-A Myc epitope tag was inserted into rat ␣4(V) collagen cDNA (NCBI accession number AF272661) at nucleotide position 1163 by overlapping PCR. PCR was performed using XL DNA polymerase (Applied Biosystems) with 10 ng of ␣4(V) collagen cDNA as the template. Amplification conditions were 1 min at 94°C and 1 or 4 min at 60°C for 20 cycles. First round PCR products were gel-purified and used as templates in a subsequent PCR assay in which sense and antisense oligonucleotide primers flanked the initiation and stop codons. PCR conditions were as above except that the extension was at 60°C for 5 min. The resulting products were cloned into pcDNA 3.1-V5-His-TOPO (Invitrogen). Positive clones were sequenced using a Beckman CEQ 2000XL capillary sequencer. Schwann cells in 60-mm dishes (70% confluence) were transfected with 3 g of plasmid DNA, 8 l of PLUS reagent, and 12 l of Lipo-fectAMINE in 2.3 ml of Opti-MEM 1 (Invitrogen). After 3 h, the transfection medium was replaced with 3 ml of Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 2 M forskolin.
Immunoblot Analysis of Collagen Polypeptides-Aliquots of Schwann cell-conditioned medium were subjected to SDS gel electrophoresis on 6% polyacrylamide gels. Proteins were transferred to polyvinylidene difluoride membranes (Immobilon) at 70 V for 1.5 h. Membranes were blocked for 1 h in blocking buffer (5% nonfat dry milk, 100 mM NaCl, 50 mM Tris-HCl, pH 7.4) plus 1% Tween 20. Membranes were incubated overnight at 4°C in rabbit anti-Myc antibody (1:1000 dilution) (Upstate Biotechnology, Inc.), rabbit anti-␣4(V) collagen antibody (1:1500 dilution), rabbit anti-␣4(V)-NTD antibody, or rabbit anti-␣1(V)-NTD antibody. The membranes were washed three times for 10 min each in blocking buffer and incubated for 30 min with the appropriate horseradish peroxidase-conjugated secondary antibodies (Bio-Rad). The membranes were washed twice for 10 min each in blocking buffer, rinsed in water, and then washed twice for 10 min each in Tris-buffered saline plus 0.1% Tween 20. Blots were developed in SuperSignal West Pico (Pierce) and imaged on a Lumi-Imager (Roche Applied Science).
Immunocytochemistry-Immunostaining of Schwann cells was performed first by rinsing the cells with cold PBS and incubating them on ice with primary antibodies diluted in blocking buffer for 1 h. Next, the cells were rinsed with PBS and fixed for 30 min with 3% paraformaldehyde in PBS. The cells were then rinsed with PBS and incubated with Alexa-conjugated secondary antibodies (Molecular Probes) diluted in blocking buffer for 1.5 h at room temperature. Finally, the cells were rinsed with PBS, overlaid with coverslips using MOWIOL mounting solution and imaged using a Leica Microsystems scanning confocal microscope.
Sciatic nerves were isolated, embedded in Tissue-Tek freezing medium, and frozen at Ϫ20°C. Cryosections (7 m thick) were cut using a Reichert-Jung cryostat, placed on glass microscope slides, and fixed with 3% paraformaldehyde in PBS for 30 min. After fixing, sections were rinsed with PBS and incubated with primary antibodies diluted in blocking buffer for 1 h at room temperature. Following a rinse with PBS, sections were incubated with Alexa-conjugated secondary antibodies (Molecular Probes) diluted in blocking buffer for 1.5 h. After a final rinsing with PBS, the sections were overlaid with coverslips using MOWIOL mounting solution and imaged with a Leica Microsystems confocal microscope.
SC ␣4(V)-NTD Binding Assays-Schwann cells were plated in glass slide chambers (Nunc) coated with collagen type IV (2.5 g/cm 2 ) (BD Biosciences) and cultured for 3 days in Schwann cell growth medium. The cells were then placed on ice, rinsed with PBS, and incubated for 20 min in serum-free medium (50% Dulbecco's modified Eagle's medium, 50% Ham's F-12 medium, 100 g/ml apotransferrin, 100 ng/ml insulin) plus 1% bovine serum albumin with or without purified recombinant ␣4(V)-NTD (10 g/ml). Next, the cells were rinsed with PBS and prepared for immunocytochemistry as described above. To investigate the effect of removing glycosylphosphatidylinositol-anchored proteins, the cells were rinsed with 20 mM HEPES, 150 mM NaCl, and 2 mM CaCl 2 , pH 7.4, and incubated in HEPES buffer with or without PI-PLC (1 unit/ml Prozyme/Glyco) for 20 min at 37°C. ␣4(V)-NTD binding was assayed as described above.
Pull-down Assays of ␣4(V)-NTD Binding-Detergent extracts of cultured Schwann cells were clarified by centrifugation, mixed with purified recombinant ␣4(V)-NTD bound covalently to cyanogen bromideactivated Sepharose 4B (Sigma), and rocked gently at 4°C for 3 h. The beads were harvested by a brief centrifugation, and supernatant solutions containing unbound proteins were removed. The beads were rinsed three times with PBS, and bound proteins were eluted with 1.5 M sodium chloride and 10 mM Tris-HCl, pH 7.5. Eluted proteins were digested with heparitinase (4.5 milliunits/ml) (Seikagaku) for 18 h at 37°C and then separated by SDS gel electrophoresis. Immunoblotting was performed using antibodies to a heparan sulfate Neo epitope (3G10) (Seikagaku), perlecan (a gift from Dr. Peter Yurchenko), glypican-1 (13), or syndecan-3 (7).

Constitutive Release of ␣4(V)-NTD in Schwann Cell
Cultures-The domain structure of the ␣4(V) collagen polypeptide is shown in Fig. 1A. We have shown previously (8) that the addition of ascorbic acid to Schwann cell cultures stimulates the assembly of ␣4(V) polypeptides into triple helical collagen molecules that also contain ␣1(V) polypeptides.
As shown in Fig. 2, ascorbic acid also stimulated the release of the NTD from the ␣4(V) collagen polypeptide. This was demonstrated by immunoblot analysis of Schwann cell-conditioned medium from cultures grown without or with ascorbic acid. When the medium from cultures grown in the absence of ascorbic acid was analyzed, antibodies to ␣4(V) collagen or the ␣4(V)-NTD revealed a single immunoreactive band of ϳ200 kDa. Similar results were obtained when polypeptides from the medium of cultures transfected with ␣4(V) collagen cDNA containing a Myc epitope tag inserted into the NTD were stained with anti-Myc antibodies ( Fig. 2) (also see Fig. 1 for the location of the Myc epitope tag). Based on immunoreactivity with anti-␣4(V)-NTD and anti-Myc antibodies, the 200-kDa polypeptide contained the ␣4(V)-NTD but not the noncollagenous C-terminal domain (see Fig. 1). The latter conclusion is based on the observation that this polypeptide was not detected by anti-V5 antibodies when medium from cells transfected with ␣4(V) cDNA containing a V5 epitope tag at the C terminus was analyzed (data not shown).
Conditioned medium of Schwann cells incubated with ascorbic acid contained the 200-kDa immunoreactive band plus a more prominent band of ϳ95 kDa that was stained by anti-␣4(V) and anti-NTD antibodies (Fig. 2). Ascorbic acid produced a slight reduction in mobility of the 200-kDa polypeptide, consistent with increased post-translational modification. An apparently identical Myc-reactive polypeptide of 95 kDa was observed in medium from ascorbate-treated Schwann cells transfected with Myc-tagged ␣4(V) collagen cDNA. From these results, we conclude that the 95-kDa polypeptide is the intact released ␣4(V)-NTD. After 48 h of ascorbate treatment, the ratio of released ␣4(V)-NTD to full-length 200-kDa ␣4(V) polypeptide in Schwann cell-conditioned medium was ϳ10:1. The released ␣4(V)-NTD was not detected in Schwann cell lysates (data not shown), suggesting that the release occurred after secretion of the ␣4(V) collagen polypeptide. Together, these results demonstrate that the ␣4(V)-NTD is released constitutively from type V collagen molecules and that it then accumulates in conditioned medium of cultured Schwann cells. The ␣4(V)-NTD release is accelerated greatly by ascorbate treatment, which promotes the assembly of collagen type V trimers.
Type V collagen molecules secreted by Schwann cells are heterotrimers that contain both ␣4(V) and ␣1(V) collagen polypeptides. The ␣1(V) collagen polypeptide also contains a noncollagenous N-terminal domain that shows a high degree of sequence homology with the N-terminal half of the ␣4(V)-NTD. As shown in Fig. 2, conditioned medium of Schwann cells cultured without ascorbate contained a 200-kDa polypeptide that was stained with antibodies raised against the ␣1(V)-NTD. Analysis of the medium from cells cultured with ascorbic acid showed the 200-kDa band plus additional bands of ϳ180 kDa and a doublet of ϳ45 kDa. The lower molecular mass of the doublet polypeptides was approximately half the size of the released ␣4(V)-NTD. These results suggest that the ␣1(V)-NTD contains two protease cleavage sites, the presence of which results in the release of two nonoverlapping NTD fragments.
␣4(V)-NTD Release in Vivo-␣4(V) collagen is expressed by Schwann cells in newborn rat peripheral nerves and at significantly reduced levels in adult nerves (7). To investigate the fate of the ␣4(V) NTD in vivo, rat sciatic nerves were harvested  on postnatal day 2 (P2) and at 3 months of age. Nerve extracts were subjected to immunoblot analysis with anti-␣4(V) and anti-␣4(V)-NTD antibodies. As shown in Fig. 3, extracts of P2 nerves contained the 200-kDa form of ␣4(V) collagen and a much greater amount of the released ␣4(V)-NTD. In extracts of nerves from adult animals, the 200-kDa ␣4(V) collagen form was barely detectable, whereas the released ␣4(V)-NTD was detectable but present at lower levels than in extracts of P2 nerves. These results demonstrate that ␣4(V)-NTD is released and then accumulates in peripheral nerves in vivo. The highest levels of ␣4(V)-NTD are observed in developing nerves.
The localization of ␣4(V)-NTD in peripheral nerve tissue was investigated by confocal microscopy. Fig. 4 shows images of postnatal day 10 sciatic nerve sections stained with anti-␣4(V)- NTD, anti-laminin, and anti-neurofilament antibodies. Laminin is a major component of the basal lamina sheets that are apposed closely to the outer Schwann cell membrane (4). Antilaminin antibodies produced linear staining of the region outlining the outer Schwann cell membrane, corresponding to the location of the Schwann cell ECM. Anti-␣4(V)-NTD antibodies produced discontinuous and slightly more diffuse staining of the same region, suggesting that ␣4(V)-NTD is localized to the Schwann cell ECM. This was confirmed by dual staining of nerve sections with anti-␣4(V)-NTD and anti-laminin antibodies (Fig. 4, C-F).
␣4(V)-NTD Binding to Schwann Cells-Experiments were carried out to investigate the interaction between ␣4(V)-NTD and Schwann cells. The addition of soluble ␣4(V)-NTD to Schwann cells that were incubated in medium without ascorbic acid resulted in ␣4(V)-NTD binding to the Schwann cell surface, which was visualized readily by immunofluorescence microscopy (Fig. 5).
The role of the ␣4(V)-NTD high affinity heparin binding site in binding to Schwann cells was investigated. As shown in Fig.  5, the addition of soluble heparin completely blocked the binding of ␣4(V)-NTD to Schwann cells. In addition, ␣4(V)-NTD that was mutated to remove the heparin binding site (⌬HBS) (9) failed to bind to Schwann cells. These results suggest that ␣4(V)-NTD binds via its heparin binding site to heparan sulfate proteoglycans on the Schwann cell surface.
Ascorbic acid stimulates the assembly in Schwann cell cultures of fibrillar ECM that contains collagen types I, IV, and V and the heparan sulfate proteoglycan perlecan. In ascorbic acid-treated cultures, endogenous ␣4(V)-NTD was also associated with the fibrillar extracellular matrix (not shown). Exogenous ␣4(V)-NTD bound to the fibrillar ECM in ascorbatetreated Schwann cell cultures (see below). Together, these results suggest that the ␣4(V)-NTD binds to both the cell surface and ECM heparan sulfate proteoglycans of Schwann cells.
Identification of ␣4(V)-NTD-binding Proteoglycans in Schwann Cells-Schwann cell proteoglycans that bind to ␣4(V)-NTD were isolated by pull-down assays using ␣4(V)-NTD or ␣4(V)-NTD-⌬HBS immobilized on beads. Schwann cell lysates contained several heparan sulfate proteoglycan species with core proteins ranging in size from ϳ50 to Ͼ250 kDa, as revealed by heparitinase digestion and staining with an antibody that recognized the heparitinase digestion product on core proteins (Fig. 6). Heparan sulfate proteoglycans that bound to ␣4(V)-NTD-coated beads were identified by staining hepariti-nase-digested proteins with anti-heparan sulfate stub antibodies or specific anticore protein antibodies. ␣4(V)-NTD-bound proteoglycans included perlecan (core protein Ͼ250 kDa), a 200-kDa core protein, and glypican-1 (core protein ϳ64 kDa) (Fig. 6). Syndecan-3 was not detected in the ␣4(V)-NTD-bound samples. We believe the 200-kDa protein is a perlecan-derived fragment that does not react with the anti-perlecan antibody. No binding of heparan sulfate proteoglycans to NTD-⌬HBSconjugated beads was detected (Fig. 6), demonstrating that proteoglycan binding was mediated by the ␣4(V)-NTD high affinity heparin binding site. Similar results were obtained when extracts of P7 rat sciatic nerve were subjected to ␣4(V)-NTD pull-down assays (data not shown).
Glypican-1 is a lipid-anchored proteoglycan of the plasma membrane, and perlecan is a secreted ECM proteoglycan. To confirm that these proteoglycans are major binding sites for ␣4(v)-NTD, the localization of bound ␣4(V)-NTD and these proteoglycans was visualized by immunofluorescence microscopy. Fig. 7, A-C, shows micrographs of cultures that were incubated with soluble ␣4(V)-NTD and then stained with anti- ␣4(V)-NTD (green) and anti-glypican-1 (red) antibodies. Both antibodies produced a punctate staining pattern on the Schwann cell surface. Superimposition of ␣4(V)-NTD and glypican-1 images (Fig. 7C) shows a substantial overlap of the two proteins. To provide additional evidence that ␣4(V)-NTD bound to glypican-1, the effects of digestion with PI-PLC were investigated. As shown in Fig. 8, incubation with PI-PLC abolished cell surface glypican-1 staining. Incubation with PI-PLC also substantially reduced the binding of soluble ␣4(V)-NTD. These data also demonstrate that glypican-1 and surface-bound ␣4(V)-NTD are enriched in the numerous filopodia that extend from the Schwann cell surface. Fig. 7, D-F, shows micrographs of cultures that were incubated in medium with ascorbate to induce ECM assembly and then incubated with soluble ␣4(V)-NTD and stained with anti-␣4(V)-NTD (red) and anti-perlecan (green) antibodies. Both antibodies stained prominent fibrillar structures that correspond to the Schwann cell ECM. Superimposition of ␣4(V)-NTD and perlecan images (Fig. 7F) shows a substantial overlap of the two proteins. Together, these data support the conclusion that the major Schwann cell binding sites for ␣4(V)-NTD are glypican-1 and perlecan. DISCUSSION In this study we investigated the processing of the ␣4(V) collagen chain by Schwann cells. This collagen chain is expressed abundantly by Schwann cells during peripheral nerve development (14). An unusual feature of the structure of ␣4(V) collagen is the presence of a large noncollagenous N-terminal domain that contains a unique high affinity heparin/heparan sulfate binding site (8,9). Here we show that the ␣4(V)-NTD is released constitutively by Schwann cells both in culture and in vivo and that it accumulates in conditioned culture medium or the endoneurium ECM in vivo. The released ␣4(V)-NTD binds to the cell surface and to ECM heparan sulfate proteoglycans, especially glypican-1 and perlecan. This binding occurs via a mechanism that requires the high affinity heparin binding site and appears to be mediated by binding of the ␣4(V)-NTD to heparan sulfate chains. This process is summarized in Fig. 9.
The protease responsible for ␣4(V)-NTD release is not known. The proteolytic cleavage that results in ␣4(V)-NTD release appears to occur after secretion of the polypeptide because ␣4(V)-NTD is not detected in Schwann cell lysates. Release is increased dramatically by ascorbate treatment of the cells, which promotes assembly of collagen type V heterotrimers as a result of increased post-translational modification of amino acid residues in the collagen domain. After 48 h of ascorbate treatment, ϳ90% of the ␣4(V) collagen chains that accumulate in Schwann cell medium have undergone ␣4(V)-NTD release. Several protease inhibitors that have been reported to block proteolytic collagen processing events in other systems (15)  metalloproteinase inhibitor BB-94 (50 M) failed to produce demonstrable inhibition (data not shown).
Although the protease cleavage site was not identified directly, several observations suggest that the site is located within the second collagen GXX sequence interruption. Expression of ␣4(V) collagen with a Myc epitope tag near the Cterminal end of the noncollagenous NTD (see Fig. 1) resulted in the accumulation of Myc-tagged ␣4(V)-NTD in Schwann cell medium. Expression of ␣4(V) collagen with a Myc epitope tag located between the second and third GXX repeat of the uninterrupted collagen domain (amino acid 369) resulted in the accumulation of Myc-tagged ␣4(V) collagen domain (data not shown). When the Myc epitope tag was inserted within the second GXX interruption (amino acid 472), the epitope was destroyed. This region contains the sequence AQAQA, which has been suggested as the site for cleavage of the ␣1(XI) collagen chain (16). This putative cleavage site sequence is conserved in other members of this collagen gene family, including ␣1(V) and ␣4(V) collagen (Fig. 1). It is also of interest that ␣1(V) and ␣1(XI) collagens each contain an additional AQAQ motif near the middle of the noncollagenous N-terminal domain that is not present in the ␣4(V)-NTD (Fig. 1). This could explain the observation that the ␣4(V)-NTD is released as a single large fragment, whereas the ␣1(V) NTD is released as two smaller fragments of approximately equal size (Fig. 9).
We also investigated the fate of released ␣4(V)-NTD. In peripheral nerve tissue, anti-␣4(V)-NTD immunoreactivity was associated with the region corresponding to the Schwann cell outer membrane and the closely apposed Schwann cell ECM (Fig. 4). The immunostaining reflects primarily the localization of released ␣4(V)-NTD, because immunoblot analysis of nerve tissue revealed a strong preponderance of released ␣4(V)-NTD compared with ␣4(V) collagen chains containing full-length NTD (Fig. 3). It is noteworthy that a similar localization is observed for glypican-1 and perlecan in peripheral nerve tissue (12).
Experiments with cultured Schwann cells revealed that soluble ␣4(V)-NTD binds to the Schwann cell plasma membrane primarily through binding to glypican-1. Evidence for this is provided by the co-localization of bound ␣4(V)-NTD and cell surface glypican-1, the significant reduction in cell surface ␣4(V)-NTD binding by pretreatment with PI-PLC, and pulldown assays that demonstrated glypican-1 binding to ␣4(V)-NTD. ␣4(V)-NTD binds to heparan sulfate chains on glypican-1, based on the observations that the binding requires the high affinity heparin binding site in ␣4(V)-NTD and that the binding is blocked by low concentrations of exogenous soluble heparin. ␣4(V)-NTD also binds to the heparan sulfate proteoglycan perlecan in Schwann cell cultures. The association of ␣4(V)-NTD with perlecan in the ECM of Schwann cell cultures incubated in medium with ascorbic acid was observed.
In the experiments reported here, we failed to detect significant binding of ␣4(V)-NTD to the transmembrane heparan sulfate proteoglycan syndecan-3. We have shown previously (11) that ␣4(V)-NTD binds syndecan-3. It was this observation, in fact, that led to the initial identification and isolation of ␣4(V) collagen in our laboratory. The explanation for this apparent contradiction lies in the fact that the steady state concentration of cell surface-associated syndecan-3 is very low in Schwann cells. This is a result of constitutive matrix metalloproteinase-mediated shedding of the syndecan-3 extracellular domain from the plasma membrane (17). The ability to detect binding of ␣4(V)-NTD to glypican-1 and perlecan but not syndecan-3 likely reflects the relative steady state levels of these heparan sulfate proteoglycans on the Schwann cell surface and not the inherent specificity of the binding reaction.
Immobilized ␣4(V)-NTD is adhesive for Schwann cells and sensory neurons and is an excellent substrate for Schwann cell and axonal migration (9,10). In contrast, ␣4(V)-NTD is a poor substrate for fibroblast adhesion. The adhesive property of ␣4(V)-NTD for Schwann cells appears to be dependent on heparan sulfate-mediated binding, because adhesion is blocked by soluble heparin and abolished by deletion of the high affinity heparin binding site. The findings reported here suggest strongly that glypican-1 is the primary cell surface ␣4(V)-NTDbinding protein on Schwann cells.
The physiological function of ␣4(V)-NTD release is not known. A potential function of ␣4(V)-NTD release may be to terminate the adhesive activity of the molecule by untethering it from the ECM. It is also possible that the ␣4(V)-NTD possesses additional functional activity of an unknown nature and that heparan sulfate-mediated binding to the Schwann cell surface and nerve ECM provides a mechanism to immobilize the protein. Such immobilization could be used to target ␣4(V)-NTD to a specific subcellular localization. As revealed by the Schwann cell binding experiments, ␣4(V)-NTD bound to glypican-1 is enriched in the numerous filopodia that decorate the Schwann cell surface. Alternatively, ␣4(V)-NTD immobilization could provide a reservoir of protein to be released in response to a specific physiological stimulus. Such a mechanism has been proposed (18,19) for other heparan sulfatebinding proteins, including polypeptide growth factors and secreted proteases.