(Received for publication, February 20, 1997, and in revised form, March 18, 1997)
From the Program in Cellular and Molecular Biology and Department of Pathology and Laboratory Medicine, University of Wisconsin-Madison, Madison, Wisconsin, 53706-1532
The syndecan family of cell surface proteoglycans
regulates cell adhesion and growth factor signaling by binding
components of the extracellular matrix and growth factors. To date, all
known ligand interactions are via the covalently attached
glycosaminoglycan chains. To assay for potential extracellular
interactions via the core proteins directly, the recombinant
extracellular domain of syndecan-4 (S4ED), one of the four syndecan
family members, was tested as a substratum for the attachment of
mammalian cells. Human foreskin fibroblasts bind to mouse S4ED, and
both mouse and chicken S4ED can block this binding, with 50%
inhibition observed between 0.1 and 1 × 107
M. The extracellular domain of another syndecan family
member, syndecan-1, fails to compete for cell binding to mouse S4ED.
Amino acids 56-109 of the 120-amino acid mouse S4ED compete fully,
suggesting that the cell binding domain is within this region. The
ability of syndecan-4 to interact with molecules at the cell surface
via its core protein as well as its glycosaminoglycan chains may
uniquely regulate the formation of cell surface signaling complexes
following engagement of this proteoglycan with its extracellular
ligands.
Syndecan-4, also known as ryudocan or amphiglycan, is one of four members of the syndecan family of cell surface proteoglycans. These type I transmembrane proteins are unified by their homologous transmembrane and cytoplasmic domains and the covalent addition of glycosaminoglycan chains, predominantly heparan sulfate, to their extracellular domains. Their heparan sulfate chains bind components of the extracellular matrix, such as fibronectin and collagen, as well as heparin-binding growth factors such as the FGFs1 (1-4). In tissues, these proteoglycans show distinct expression patterns, although their expression can be overlapping and more than one syndecan family member may be found on a single cultured cell type (3, 5-7). The syndecans are present at the plasma membrane and are also in the medium of cultured cells as a result of shedding the extracellular domain from the cell surface (7, 8).
Despite the emphasis on the binding of ligands to their glycosaminoglycan chains, the syndecan core proteins are also likely to have important roles in the regulation of cell adhesion and cell morphology. Syndecan-1 expressed in Schwann cells co-aligns with actin filaments in response to antibody ligation, a process dependent on a specific tyrosine residue within the syndecan-1 cytoplasmic domain (9). Syndecan-1 expressed in a B cell line enables the cells to bind and spread on heparan sulfate ligands and on core protein-specific antibodies (10). Antibody-mediated spreading does not require the cytoplasmic domain of syndecan-1 or its glycosaminoglycan chains, suggesting that the signals that convey syndecan-1-specific cell spreading result from interactions between the transmembrane and/or extracellular protein domains of syndecan-1 and other cell surface or transmembrane molecules (10).
Engagement of cell surface proteoglycans, such as the syndecans, with
other receptors may be a common growth factor and adhesion signaling
mechanism. In FGF signaling, the high affinity signaling complex
consists of a receptor tyrosine kinase(s), the FGF ligand, and heparan
sulfate proteoglycans such as the syndecans (11-13). Similar
multi-molecular complexes may assemble during cell adhesion. NG2, a
cell surface chondroitin sulfate proteoglycan, cooperates with 4
1
integrin in the development of focal adhesions (14). Simultaneous
engagement of NG2 and
4
1 signals melanoma cells to spread and
form focal adhesions in a tyrosine kinase-dependent process. When either receptor is individually engaged, cells bind but
fail to change shape. Similarly, the interaction between cell surface
heparan sulfate proteoglycans and fibronectin stimulates focal adhesion
formation but only in cooperation with integrins, such as the
5
1
integrin (15, 16). In this assay, the engagement of proteoglycan can be
replaced by activating protein kinase C, suggesting that adhesive
interactions with cell surface proteoglycans can lead to intracellular
signaling events (17). Syndecan-4, unlike the other syndecan family
members, localizes to focal adhesions, implicating this proteoglycan as
part of their signaling mechanism (18, 19). To better understand how
syndecan core proteins might regulate the formation of multi-molecular
cell adhesion and growth factor signaling complexes, studies were
undertaken to identify interactions between the extracellular protein
domain of syndecan-4, the most widely expressed of the syndecan family members, and other receptors at the cell surface.
Neonatal human foreskin fibroblasts were cultivated in DMEM (Life Technologies, Inc.), supplemented with 10% calf serum (Hyclone), and used prior to 60 days in culture.
Plasmid ConstructsDr. Merton Bernfield (Harvard Medical
School, Boston, MA) kindly provided the cDNAs encoding mS4ED and mouse
syndecan-1 (7, 20). Chicken syndecan-4 cDNA was kindly provided by Dr.
Paul F. Goetinck (Massachusetts General Hospital, Charlestown, MA) (21). His-mS4ED was constructed by cloning mS4ED into the
BamHI (5) and EcoRI (3
) sites of the pET-30a
vector (Novagen), Full-length and truncated mS4ED, cS4ED (nucleotides
114-482) and mS1ED (nucleotides 291-989) cDNAs were PCR amplified
using primers containing restriction enzyme sites, BamHI
(5
) or EcoRI (3
), for cloning into pGEX-2T (Pharmacia
Biotech Inc.). The EcoRI site is immediately preceded by the
in-frame stop codon TAA. PCR amplified cDNAs were confirmed by
sequencing.
S4ED protein alignment, calculation of amino acid similarity, and homology searches were performed with Genetics Computer Group, Inc. software (PileUp, Gap, FastA) and the National Center for Biotechnology Information network service (BLAST), which access GenBank and EMBL sequence data bases.
Production and Quantification of Recombinant S4EDGST fusion proteins were isolated using glutathione-Sepharose beads according to the manufacturer's instructions for "bulk purification of fusion proteins" (Pharmacia) with the following exceptions. After induction, bacterial pellets were resuspended in phosphate-buffered saline (pH 7.4) containing 25% sucrose, 1 mM EDTA, 5 mM dithiothreitol, 10 µg/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride. Lysozyme (1 mg/ml) and 1% TX-100 were added, followed by sonication and DNase I (1500 NIH units) treatment prior to centrifugation. His-mS4ED protein was prepared similarly, except that purification was achieved by nickel chromatography (Qiagen) followed by imidazole elution. GST fusion proteins bound to glutathione beads were digested with 5 NIH units thrombin/mg fusion protein for 90 min at 4 °C, and thrombin was removed with benzamidine-Sepharose. Glutathione-eluted fusion proteins were quantified by absorbance at 280 nm. For GST and GST constructs containing mS4ED sequences, protein concentrations were calculated using the formula 1 A280 = 0.5 mg/ml. For GST-cS4ED and GST-mS1ED, 1 A280 = 0.4 mg/ml. His-mS4ED and thrombin-released peptides were quantified relative to a standard curve of thrombin-released mS4ED by Coomassie Blue or SYPRO®-Red (Molecular Probes) staining of proteins in polyacrylamide gels.
Substrata for Cell Binding AssaysHigh binding 96-well
plates (Costar, 2585) were used for all cell binding assays.
Recombinant proteins were diluted in HEPES/DMEM to 0.2 µM, unless otherwise noted, and incubated overnight at 4 °C. Plasma fibronectin was diluted in phosphate-buffered saline (pH 7.4) to approximately 45 nM (10 µg/ml) and plated for
2 h at 37 °C. For the experiment shown in Fig. 2B,
the plating concentration of GST-mS4ED was 45 nM (1.75 µg/ml). Wells were blocked by incubating with HEPES/DMEM containing
1% heat-denatured BSA prior to the addition of cells.
Cell Binding Assays
Cells were washed twice with HEPES/DMEM and treated with 25 µg/ml cycloheximide in HEPES/DMEM for 2 h prior to trypsinization. After washing with 20 mM Tris (pH 7.6) containing 1 mM EDTA, 150 mM NaCl, and 25 µg/ml cycloheximide, cells were trypsinized for 2 min at 37 °C by incubating with 0.25% trypsin dissolved in the same buffer. The trypsin was inactivated by washing cells with 0.5 mg/ml soybean trypsin inhibitor dissolved in HEPES/DMEM containing 25 µg/ml cycloheximide. For all experiments, 50,000 cells were added per well except for Fig. 2B, where 25,000 cells were added. Competing proteins were added to each well after blocking but prior to the addition of cells. Cell binding assays were performed at 37 °C in HEPES/DMEM containing 0.2% heat-denatured BSA and 25 µg/ml cycloheximide in a total volume of 100 µl/well. Nonadherent cells were removed by washing the wells three times with HEPES/DMEM. Fixed, adherent cells were quantified by colorimetric detection with 1% bromphenol blue (22).
Recombinant mS4ED was
engineered and expressed in bacteria, providing glycosylation-free core
protein. Two independent expression vectors, each of which encodes
mS4ED as a C-terminal fusion protein with distinct amino sequences,
were used (Fig. 1A). One encodes a GST fusion
protein containing the entire 120-amino acid mS4ED attached to the GST
C terminus. This construct also contains a thrombin cleavage site,
allowing isolation and testing of the mS4ED domain (Fig.
1A). As another means of ensuring that the activity of the
fusion protein is due to mS4ED alone, a second fusion construct was
used, namely, His-mS4ED (Fig. 1A) His-mS4ED is comprised of
six histidines and an immunological tag (S·TagTM) at the N terminus
of mS4ED. Polyacrylamide gel electrophoresis of the affinity purified
protein products reveals one major protein band following purification
of GST, GST-mS4ED, and thrombin-released mS4ED (Fig. 1B).
Immunoblotting with GST-mS4ED serum, which recognizes both mS4ED and
GST, confirmed the identity of the protein bands and demonstrated that
little, if any, GST protein contaminated the thrombin-released mS4ED
preparation (data not shown). The His-mS4ED preparation contains three
major protein bands. Only the slowest migrating band of the three
reacts with GST-mS4ED serum, thus identifying this protein as intact
His-mS4ED (data not shown). The two smaller proteins may represent
co-purifying proteins or degradation products. All proteins containing
mS4ED sequence migrate 5-7 kDa higher than their calculated molecular mass. This is not unexpected, as deglycosylated, full-length syndecan-4 has been shown to migrate with an apparent molecular mass 10-15 kDa
greater than its actual molecular mass (3, 23). Importantly, none of
the minor contaminating bands in any one mS4ED preparation is visibly
shared by all three mS4ED preparations (GST-mS4ED, His-mS4ED, and
thrombin-released mS4ED).
Human Foreskin Fibroblasts Bind to mS4ED Substrata
To detect interactions between recombinant mS4ED and other receptors at the cell surface, a solid phase cell binding assay was developed. Substrata were prepared by incubating microtiter wells with the various recombinant forms of mS4ED, including intact fusion proteins or thrombin-released peptide. After blocking, the individual substrata were assessed for their ability to support cell binding over time (Fig. 2A). Human foreskin fibroblasts, pretreated with cycloheximide prior to suspension with trypsin, bind avidly to all substrata containing mS4ED but fail to bind to a substratum composed of GST alone. Binding is maximal by 60 min at 37 °C, and greater than 75% of the input cells bind to GST-mS4ED. Binding occurs despite the presence of cycloheximide, suggesting that binding is via a trypsin-resistant cell surface receptor.
To compare the binding of mS4ED to that of a known adhesive ligand, microtiter wells were coated with fibronectin. Cell binding was monitored over time and quantified (Fig. 2B). At equimolar plating concentrations, the extent of cell binding to GST-mS4ED and fibronectin was comparable, although the kinetics of binding to GST-mS4ED were slightly slower.
Cell Binding to mS4ED Is SpecificTo assess the specificity
and affinity of cell binding, soluble proteins were evaluated for their
ability to compete in the binding assay. Binding to GST-mS4ED is
reduced in a dose-dependent manner by GST-mS4ED (Fig.
3A). The IC50 in this experiment is 0.13 µM, suggesting that the interaction between mS4ED
and its putative cell surface receptor(s) is of moderate affinity,
similar to that of some integrins for their ligands (24).
Thrombin-released mS4ED peptide competes as effectively or better than
the GST-mS4ED fusion protein (Fig. 3B). Six independent
experiments yielded an average IC50 of 0.04 µM for thrombin-released mS4ED, with a range of 0.01-0.1
µM (Fig. 4). The extracellular domain of
another syndecan family member, syndecan-1, was also evaluated for its ability to compete for binding of cells to a GST-mS4ED substratum. Neither full-length GST-mS1ED nor thrombin-released mS1ED compete for
binding when added at 13 and 3.5 µM, respectively (Fig.
3, A and B). Interestingly, chicken syndecan-4
extracellular domain (cS4ED) competes as well as mS4ED (cloned from
mouse) for binding of cells to GST-mS4ED (Fig. 3A),
suggesting that the cell binding domain of mS4ED is conserved across
species.
Although cells from both human and mouse species bind to GST-mS4ED, not all cell types bind, suggesting that the receptor(s) for mS4ED may be expressed in a cell type-specific manner. Cells that bind avidly to GST-mS4ED include human foreskin fibroblasts, mouse aortic endothelial cells, and mouse NIH 3T3 and Swiss 3T3 fibroblast cell lines. Cells that fail to bind include mouse NMuMG epithelial cells and the human Raji and ARH-77 B cell lines (data not shown).
The S4ED56-109 Fragment Contains the Cell Binding DomainA domain containing the cell binding activity was identified by testing truncated mS4ED peptides as competitors in the cell binding assay. Complementary DNAs encoding mS4ED peptides were created by PCR amplification and were expressed as GST fusion proteins. IC50 values were calculated for both intact GST fusion proteins and their thrombin-released peptides (Fig. 4). Thrombin-released mS4ED and mS4ED56-109 compete equally well for cell binding to GST-mS4ED, demonstrating that amino acids 56-109 can fully replace the cell binding activity of the full-length mS4ED. Further, thrombin-released mS4ED1-80 and mS4ED81-120 compete 100-fold less effectively than mS4ED56-109. These two peptides together contain all 120 amino acids that constitute mS4ED but bisect amino acids 56-109, demonstrating that the cell binding domain is not reiterated within the mS4ED sequence and that the binding activity does not reside solely in either half of the mS4ED56-109 domain.
Some thrombin-released peptides compete better than the GST fusion proteins from which they were derived, suggesting that proximity to GST can sterically hinder the cell binding domain. For example, GST-mS4ED and thrombin-released mS4ED compete with similar effectiveness, whereas GST-mS4ED56-109 competes 1000 times less effectively than thrombin-released mS4ED56-109. Importantly, thrombin-released mS4ED and thrombin-released mS4ED56-109 compete identically.
Alignment of Vertebrate Syndecan-4 Extracellular DomainsAn
alignment of the cloned syndecan-4 extracellular domains reveals the
positioning and degree of conservation of amino acids 56-109 (Fig.
5). Although chicken and mouse are only 34% identical within this domain, pairwise alignment of the sequences indicate similarity values ranging from 62 to 70%. The S4ED56-109
domain begins 15 amino acids C-terminal to the last possible glycosaminoglycan acceptor site in the extracellular domain of syndecan-4. This distinguishes the cell binding domain spatially from
the glycosaminoglycan acceptor sites. Syndecan-4 is expressed as a
heparan sulfate proteoglycan and to a lesser extent as a hybrid
proteoglycan bearing both heparan sulfate and chondroitin sulfate
chains (3, 18, 21, 25). Because heparin and chondroitin sulfate A, B,
and C are unable to compete for binding of cells to GST-mS4ED (up to
100 µg/ml, data not shown), it seems unlikely that the
glycosaminoglycan decoration of endogenous syndecan-4 would compete
directly for cell binding domain interactions. Because no other
proteins currently in the data base share significant homology to this
region of syndecan-4, the interaction between amino acids 56-109 and
the cell surface may be unique to syndecan-4.
The cell binding activity of the syndecan-4 core protein constitutes a new class of interactions for the syndecans, namely, the association of the extracellular core protein domain of a syndecan and other cell surface molecules. This domain may interact with molecules on adjacent cells to mediate cell-cell adhesion or potentially function as a shed or matrix-embedded adhesive ligand. Because syndecan-4 is the major heparan sulfate proteoglycan to localize to focal adhesions, it is intriguing to speculate that the interactions mediated by the cell binding domain of syndecan-4 may regulate its localization to or the formation of such adhesion structures (18, 19). Given the propensity of the syndecans to homodimerize (26, 27) and potentially multimerize in response to ligand engagement (28), it seems equally likely that the syndecan-4 cell binding domain might regulate the formation of other multi-molecular cell surface signaling complexes.