(Received for publication, September 18, 1996, and in revised form, October 7, 1996)
From the Sigfried and Janet Weis Center for Research, Geisinger Clinic, Danville, Pennsylvania 19822
The amino acid sequence of rat N-syndecan core
protein was deduced from the cloned cDNA sequence. The sequence
predicts a core protein of 442 amino acids with six structural domains:
an NH2-terminal signal peptide, a membrane distal
glycosaminoglycan attachment domain, a mucin homology domain, a
membrane proximal glycosaminoglycan attachment domain, a single
transmembrane domain, and a noncatalytic COOH-terminal cytoplasmic
domain. Transfection of human 293 cells resulted in the expression of
N-syndecan that was modified by heparan sulfate chain addition.
Heparitinase digestion of the expressed proteoglycan produced a core
protein that migrated on SDS-polyacrylamide gels at an apparent
molecular weight of 120,000, identical to N-syndecan synthesized by
neonatal rat brain or Schwann cells. Rat genomic DNA coding for
N-syndecan was isolated by hybridization screening. The rat N-syndecan
gene is comprised of five exons. Each exon corresponds to a specific
core protein structural domain, with the exception of the fifth exon,
which contains the coding information for both the transmembrane and cytoplasmic domains as well as the 3-untranslated region of the mRNA. The first intron is large, with a length of 22 kilobases. The
expression of N-syndecan was investigated in late embryonic, neonatal,
and adult rats by immunoblotting and Northern blotting analysis. Among
the tissues and developmental stages studied, high levels of N-syndecan
expression were restricted to the early postnatal nervous system.
N-syndecan was expressed in all regions of the nervous system,
including cortex, midbrain, spinal cord, and peripheral nerve.
Immunohistochemical staining revealed high levels of N-syndecan
expression in all brain regions and fiber tract areas.
The syndecans are a gene family of transmembrane cell surface proteoglycans (reviewed in Refs. 1 and 2). Syndecans play an important role in tissue morphogenesis and differentiation by virtue of their ability to bind a number of extracellular adhesive proteins and growth factors. Four different syndecan core proteins that are the products of different genes are synthesized by mammalian cells. Syndecan synthesis is highly regulated (3-7) and is dependent on both cell type and developmental state (8). Syndecan core proteins are characterized by highly conserved transmembrane and noncatalytic COOH-terminal cytoplasmic domains but structurally distinct extracellular domains. This modular structural design suggests that individual mammalian syndecans have evolved to carry out specific functions within the tissues where they are expressed, probably related to the binding of specific extracellular ligands, but that these may be linked to common intracellular activities.
We reported the cloning of a partial cDNA sequence from neonatal rat Schwann cells that coded for the syndecan core protein, which we called N-syndecan (9). This proteoglycan is also expressed in the central nervous system. A highly homologous cDNA was cloned from chick embryo limb buds (10, 11). The chick proteoglycan, which appears to be the homologue of rat N-syndecan, was named syndecan-3.
N-syndecan purified from neonatal rat brain has been used to study the
ligand binding properties of the proteoglycan. In contrast to what has
been reported for syndecan-1, brain N-syndecan binds poorly in
solid-phase assays to most extracellular matrix proteins, including
fibronectin, laminin, and collagens I, III, IV, and V. The proteoglycan
does bind with high affinity (KD = 5 × 1010 M) to basic fibroblast growth factor
(bFGF)1(12). N-syndecan has been proposed
to be an endogenous heparin-like cofactor for bFGF in the developing
brain. bFGF has been postulated to have a number of effects during
development of the central nervous system, e.g. regulation
of oligodendrocyte terminal differentiation (13, 14). N-syndecan also
binds to heparin binding growth-associated molecule (HB-GAM)(15), an
extracellular matrix-associated adhesive protein that is expressed in
the developing brain (16), and may function as a receptor for HB-GAM in
neonatal brain tissue. Recently, we reported the purification of a
novel, peripheral nerve-specific, extracellular matrix protein, p200,
that binds N-syndecan (17). p200 is expressed in peripheral nerves
during early postnatal development. Purified p200 promotes adhesion and spreading of Schwann cells.
A determination of the biological function of N-syndecan would be facilitated by knowledge of the primary structure of the core protein and its expression in recombinant form, as well as information on the patterns of expression of N-syndecan in vivo. In this study, we report the complete amino acid sequence and expression of rat N-syndecan, the cloning and characterization of the rat N-syndecan gene, and an analysis of the tissue- and development-specific expression of N-syndecan.
The cloning and sequencing of a partial
cDNA coding for rat N-syndecan was reported previously (9). This
cDNA contained the coding sequence for the transmembrane and
cytoplasmic domains of the core protein and a portion of the ectodomain
but lacked a translational start site and signal peptide. Additional 5
cDNA sequence was obtained by sequence-specific reverse
transcriptase catalyzed primer extension using total RNA isolated from
2-day-old rat brain as template. RNA was isolated by using UltraSpec
RNazol (Biotecx Laboratories, Houston, TX) as described previously (9). Antisense primers based on the known N-syndecan cDNA sequence were
used as primers for cDNA synthesis by avian myeloblastosis virus
reverse transcriptase (Promega Corp.). After tailing the 3
end of the
resulting cDNA products with terminal deoxynucleotidyl transferase
and dCTP, the specific products were amplified by two rounds of PCR
using oligodeoxyguanosine as the sense primer and nested
sequence-specific antisense primers. The PCR products were gel purified
and ligated into plasmid pCRII (Invitrogen). Plasmids were isolated
after transformation into bacteria. DNA sequences were analyzed using
the dye terminator cycle sequencing method (Perkin Elmer) and an
Applied Biosystems 373A DNA sequencer.
A
rat Schwann cell N-syndecan cDNA probe was used to screen an EMBL3
rat liver genomic library (Clonetech). A phage clone was isolated
(JJH1) that by DNA sequence analysis was found to contain the coding
information for exons 3, 4, and 5 and part of intron II. Rescreening
the genomic library with JJH1 as a probe resulted in the isolation of
an overlapping genomic clone (NG11) that extended the sequence upstream
by approximately 10 kb but did not contain exon 1 coding information.
Another round of screening with NG11 as a probe resulted in the
isolation of an overlapping genomic clone (GNG26) that extended the
sequence upstream by approximately 8 kb but did not contain exon 1. Repeated screening of the EMBL3 library with GNG26 did not yield clones
that hybridized to cDNA sequence corresponding to exon 1. The
remaining sequence was obtained by screening a Lambda FIX II rat kidney
genomic library (Stratagene) with a 9-kb SalI restriction
fragment from the 5 end of clone GNG26. This yielded clone GNG31 that
overlapped clone GNG26 at its 3
end and extended upstream to contain
exon 1. The N-syndecan genomic clones were mapped by restriction enzyme
digestion and Southern hybridization using defined segments of the
N-syndecan cDNA as probes. Restriction fragments of interest that
included all exon sequences and intron/exon splice junctions were
subcloned into pGEM7Z (Promega) for DNA sequence analysis.
Attempts to isolate cDNAs
containing the entire protein coding sequence or to generate such
cDNAs by reverse transcriptase-linked PCR amplification with rat
brain or Schwann cell RNA as template or by overlapping PCR with the
corresponding partial cDNAs as templates were all unsuccessful.
This may be due to the extremely high G-C content of the the 5 region
of the N-syndecan cDNA (see "Results"). An expression construct
that contained cDNA coding information for the mature N-syndecan
core protein was constructed by ligating a cDNA containing the
coding sequence for the mature N-syndecan core protein to the rat
syndecan-1 signal peptide (6). This cDNA was subcloned into an
expression plasmid that uses the constitutively active cytomegalovirus
promoter and was used to transfect human 293 cells as described
previously (18). Expression of N-syndecan was detected by immunoblot
analysis, using affinity-purified polyclonal anti-rat N-syndecan
antibodies, as described previously (9).
The panel of tissues indicated in "Results" were excised from day 18 rat embryos, postnatal day 2, and adult (more than 3 months of age) rats. Brain cortex and spinal cord tissues were also obtained from rats ranging in age from embryonic day 15 to adult. The tissues were homogenized in ice-cold phosphate-buffered saline (PBS, 0.05 M sodium phosphate, 0.15 M sodium chloride, pH 7.5) plus 0.5 mM phenylmethylsulfonyl fluoride with a Polytron (2.5 ml/g of tissue). The homogenates were centrifuged at 17,000 × g for 45 min. The PBS-soluble material was removed, and the resulting pellets were extracted with 1% Triton X-100 in PBS and centrifuged as above. The Triton X-100 extracted material was removed, and the pellets were dissolved in electrophoretic sample buffer with 2% SDS. Aliquots of the extracts were subjected to SDS gel electrophoresis on polyacrylamide gels and electrophoretically transferred (4-6 h, 70 V) to Immobilon P membranes (Millipore, Bedford, MA). The blots were stained with affinity-purified anti-N-syndecan antibodies. Bound antibodies were detected by enhanced chemiluminescence (Amersham Life Science, Inc.). In some experiments, N-syndecan was treated with heparitinase (Seikagaku America, Rockville, MD) as described previously (9, 12). Preparation and characterization of affinity-purified antibodies directed against bacterially expressed N-syndecan have been described (9). Anti-peptide antibodies directed against a synthetic peptide corresponding to the COOH-terminal seven amino acids (KQEEFYA) of the N-syndecan core protein cytoplasmic domain were also prepared. The peptide was covalently coupled to keyhole limpet hemocyanin (Pierce) and injected into rabbits along with synthetic adjuvant (RIBI Immunochemicals, Hamilton, MT). The antibodies were affinity-purified on a column containing immobilized bacterially expressed core protein as described previously (9).
Northern Blot AnalysisTotal RNA was isolated from rat
tissues using UltraSpec RNazol, fractionated on 1.5%
agarose-formaldehyde gels (20 µg/lane), transferred to nylon
membranes (Schleicher and Schuell), and immobilized by UV
cross-linking. The membranes were hybridized to a 1.6-kb 32P-labeled N-syndecan cDNA probe from the
3-untranslated region. After hybridization, the membranes were washed
with 0.1 × SSC, 0.1% SDS at 65 °C. Hybridization signals were
visualized by autoradiography using DuPont Reflection film with
intensifying screens at
70 °C. Hybridization signals were
quantitated by scanning the autoradiograms with a Molecular Dynamics
laser densitometer and normalized to the quantity of 28 S rRNA loaded
on the gel, determined by scanning photographs of ethidium
bromide-stained gels.
Paraformaldehyde-fixed and paraffin-embedded brain and
spinal cord tissue obtained from postnatal rats were immunostained as
described previously (19). Following dewaxing, the sections were
immersed in 0.3% H2O2 to block endogenous
peroxidase activity and then treated with 1 mg/ml hyaluronidase
(Sigma) for 30 min at room temperature. The tissues
were blocked by incubation for 1 h with 10% goat serum and then
incubated overnight with affinity-purified anti-N-syndecan antibodies.
After rinsing, the bound antibodies were detected by the
peroxidase-antiperoxidase method using 3,3-diaminobenzidine (Sigma) as substrate.
Previously, we reported the cloning of a partial rat
N-syndecan cDNA that was truncated at the 5 end. Additional
cDNA sequence was obtained by carrying out primer extension and PCR
amplification as described under "Materials and Methods." The
resulting product extended the cDNA sequence by approximately 380 bp in the 5
direction and provided the remaining protein coding
sequence. The cDNA and deduced core protein sequences are shown in
Fig. 1.
The deduced amino acid sequence predicts a polypeptide of 442 amino
acids. The proposed initiation methionine is preceded by a purine (G)
at position 3 and is, thus, in a "strong" initiation context
(20). Several important structural features can be predicted by perusal
of the deduced amino acid sequence. The linear sequence can be divided
into six structural domains (Fig. 2A). The
sequence immediately downstream of the initiation codon contains an
uninterrupted stretch of hydrophobic amino acids. This is consistent
with the membrane topology of syndecans (type I transmembrane proteins) and identifies this region as a putative signal peptide. Applying proposed rules for predicting signal peptide cleavage sites (21), the
cleavage is predicted to occur on the COOH-terminal side of either
Ala-39 or Gly-44. The latter site would place the cleavage precisely at
the end of exon 1 (see below). The nucleotide sequence of the
5
-untranslated region and putative signal peptide is extremely rich in
G and C residues (79% of the total nucleotides). There are strings of
uninterrupted G or C sequences of 22, 19, and 18 bases, plus one in
which 19 of 20 bases are G or C. This region is expected to form rather
stable secondary structures, which may account for the difficulty in
isolating this part of the cDNA (9).
The potential glycosaminoglycan acceptor sites, identified by the consensus sequence Ser-Gly, occur in two clusters. One cluster is in a region of approximately 50 amino acids immediately following the signal peptide that contains five consensus glycosaminoglycan attachment sequences, including three in tandem. The second cluster of three potential glycosaminoglycan attachment sites is in a domain of approximately 90 amino acids located adjacent to the transmembrane domain. The membrane distal and membrane proximal glycosaminoglycan attachment domains are separated by a domain of approximately 200 amino acids that is rich in proline and threonine residues and shows significant sequence homology to mucin-like proteins. These domains, which comprise the core protein ectodomain, are followed by a stretch of 24 hydrophobic residues that constitute the transmembrane domain and a COOH-terminal domain of 34 amino acids that form the noncatalytic cytoplasmic domain. The structural features of the latter two domains, including their high degree of amino acid sequence homology with other members of the syndecan family of core proteins, have been described previously (9, 18, 22).
Structure of the Rat N-syndecan GeneRat N-syndecan genomic
sequences were isolated by hybridization screening of lambda phage rat
genomic libraries as described under "Materials and Methods." Four
overlapping genomic clones that contained the coding information for
N-syndecan were isolated. A composite map of the rat N-syndecan gene
derived by analysis of these clones is shown in Fig. 2B. The
gene consists of five exons and has a total length of approximately 30 kb. Each of the five exons corresponds to identifiable domains of the
core protein and mRNA (Fig. 2C). Exon 1 encodes the
5-untranslated mRNA sequence plus the protein signal peptide; exon
2 encodes the membrane distal glycosaminoglycan attachment domain; exon
3 encodes the proline- and threonine-rich spacer domain with homology
to mucin-like sequences; exon 4 encodes the membrane proximal
glycosaminoglycan attachment domain; and exon 5 encodes the
transmembrane and cytoplasmic domains of the polypeptide plus the
3
-untranslated region of the mRNA. The first intron is rather
large, with an estimated length of 22 kb. The other introns range in
length from 0.6 to 2.5 kb. The sequences at the exon splice sites are
typical of what has been found in other mammalian genes and are shown
in Fig. 3. With respect to overall organization and exon
and domain structure, the rat N-syndecan gene shows striking similarity
to the mouse syndecan-1 gene (23, 24).
Expression of N-syndecan cDNA
The N-syndecan expression
vector described under "Materials and Methods" was used to
transfect human 293 cells, which lack endogenous immunoreactive
N-syndecan. Immunoblot analysis of transfected cultures with
affinity-purified anti-N-syndecan antibodies revealed the synthesis of
a high molecular weight immunoreactive smear, which is characteristic
of proteoglycans, that comigrated with native N-syndecan synthesized by
rat Schwann cells or rat brain (Fig. 4). After digestion
with heparitinase, the immunoreactive protein produced by the
transfected cells was shifted to an apparent molecular weight of
approximately 120,000. The heparitinase-digested product of the
transfected cells comigrated with heparitinase-digested N-syndecan
extracted from neonatal rat brain or Schwann cells (Fig. 4). These
results demonstrate that the cloned cDNA contains the information
necessary to encode mature N-syndecan core protein that is modified by
addition of heparan sulfate chains in a manner similar to the native
proteoglycan.
N-syndecan Expression in Vivo
Regulated expression of cell
surface proteoglycans provides a mechanism for restricting the
biological activity of these molecules (8). Knowledge of N-syndecan
expression in vivo would provide important information
related to its function. Immunoblot analysis with affinity-purified
anti-N-syndecan antibodies was used to measure the steady-state levels
of N-syndecan in a variety of tissues obtained from late embryonic,
early postnatal, and adult rats. As shown in Fig.
5A, N-syndecan was detected in extracts of
peripheral and central nervous system tissue of early postnatal rats.
N-syndecan was found in all major subdivisions of the central nervous
system, including cortex, midbrain, and spinal cord. N-syndecan was not
detected by immunoblot analysis in the other neonatal rat tissues
examined, including eye, tongue, lung, liver, kidney, stomach, and
skeletal muscle (Fig. 5A), as well as spleen, heart, and
ovary (data not shown). Immunoblot analysis failed to detect N-syndecan
in extracts of adult rat tissues (data not shown), including brain and
spinal cord (see below). N-syndecan was not detected in tissue extracts
of late embryonic rats (data not shown), with the exception of the
nervous system (see below).
The distribution of N-syndecan mRNA in rat tissues was determined by Northern blot analysis. The mRNA was detected in peripheral nerve, brain, and spinal cord of early postnatal rats but not in eye, tongue, lung, liver, or kidney (Fig. 5B). Thus, the distribution of N-syndecan mRNA correlated with the pattern of expression of the proteoglycan.
The developmental time course of N-syndecan expression was measured in
rat brain and spinal cord tissue. As shown in Fig. 6A, N-syndecan was present at low levels in
late embryonic (E18) rat brain and spinal cord extracts. N-syndecan
levels increased during the early postnatal period and reached a
maximum on postnatal day 7. Thereafter, the N-syndecan levels declined
to very low levels in adult tissues. The temporal pattern of N-syndecan
expression was reflected in the levels of N-syndecan mRNA in brain
and spinal cord tissue, as measured by Northern blot analysis (Fig.
6B). The highest levels of expression were observed in the
early postnatal period, reaching a maximum at 7 days postnatal, with
very low levels of expression in late embryonic and adult tissue. An
almost identical temporal pattern of expression in the brain of HB-GAM, an N-syndecan ligand, has been reported previously (16).
Differential Extraction of N-syndecan from Rat Brain and Evidence for Shedding in Vivo
The deduced protein sequence of N-syndecan
predicts a transmembrane polypeptide. Studies with cultured cells have
shown that soluble forms of syndecans can be generated by proteolytic
cleavage of the ectodomains at a site very close to the membrane
spanning domain, a process referred to as membrane shedding (8, 25). This generates products that contain all of the glycosaminoglycan chains but not the transmembrane or cytoplasmic domains and are only
slightly smaller than the full-length proteoglycans. To assess the
membrane association of N-syndecan in vivo, neonatal rat
brain tissue was fractionated into soluble and particulate fractions by
homogenization in PBS without detergent, followed by centrifugation. The particulate fraction was extracted with buffer containing 1%
Triton X-100, and the Triton-insoluble material was solubilized in
buffer containing 2% SDS. Immunoblot analysis of the resulting fractions was carried out, using antibodies directed against the ectodomain or antibodies directed against a peptide antigen in the
COOH-terminal cytoplasmic domain. As shown in Fig. 7,
the anti-ectodomain antibodies detected a significant amount of the total N-syndecan in the PBS-soluble fraction. The N-syndecan in the PBS
extract could not be pelleted, even after centrifugation at
100,000 × g for 2 h (data not shown). The soluble
N-syndecan might result from membrane shedding of the proteoglycan.
This was supported by the observation that this form of N-syndecan was
not stained by antibodies directed against the cytoplasmic domain (Fig.
7). The Triton X-100 insoluble form of N-syndecan also failed to react
with the cytoplasmic domain antibodies, suggesting that it may be a
shed form of the proteoglycan that is associated with an insoluble
brain matrix. The Triton X-100 soluble fraction contained an additional
immunoreactive band that migrated at a position similar to that of
heparitinase treated N-syndecan (Mr 120,000).
This might be newly synthesized N-syndecan molecules that lack heparan
sulfate chains. This explanation was supported by the finding
that this form of N-syndecan was stained by the anti-cytoplasmic domain
antibodies (Fig. 7).
Immunocytochemical Staining of N-syndecan in Rat Brain and Spinal Cord
Anti-N-syndecan antibodies were used to examine the
distribution of the proteoglycan in neonatal rat brain and spinal cord by immunohistochemistry. Consistent with the immunoblot and Northern blot data, N-syndecan immunoreactivity was low in embryonic and adult
tissues (data not shown). Strong immunostaining was observed beginning
around postnatal day 1 and increased during the first 2 weeks after
birth. The staining was distributed throughout most regions of the
brain including the cortex (Fig. 8, A and
F), hippocampus (Fig. 8C), thalamus, basal
ganglia (data not shown), and fiber tract areas such as the hippocampal
commissure (Fig. 8D), fornix, and corpus callosum (data not
shown). In the cortex, the most intense staining was in the molecular
layer and white matter (Fig. 8, A, E, and F).
When immunostaining was visualized by indirect immunofluorescence, the
staining was confined largely to the surfaces of small, round cells
throughout the brain (data not shown). Results very similar to these
were obtained with spinal cord (Fig. 8I).
The sequence of the rat N-syndecan core protein deduced from the
cDNA sequence reveals several features of the core protein structure that are similar to other syndecans, as well as some distinct
features. These similarities and differences most likely reflect common
functional roles, such as extracellular ligand binding, that are
carried out within the context of cell type-specific differences in
function. It is presumably the latter that dictates the need for
multiple forms of syndecans in mammalian cells. Syndecans show the
greatest amino acid sequence homology in the transmembrane and
cytoplasmic domains. The details and possible functional consequences of the high degree of structural conservation in these domains have
been discussed previously (9, 18, 22, 26). In contrast, there is a
relatively low level of sequence homology among the ectodomains. On the
basis of sequence homology, the syndecans can be divided into
syndecan-1/N-syndecan and syndecan-2/syndecan-4 subfamilies. Although
the amino acid sequences of the ectodomains of N-syndecan and
syndecan-1 are not strikingly similar, the two core proteins appear to
reflect a common structural organization, with membrane proximal and
membrane distal glycosaminoglycan attachment domains separated by a
proline-rich spacer domain. This similarity in overall structural
organization is reflected in the exon structure of the corresponding
genes (23, 24). In both genes, each protein domain is encoded by a
separate exon, with the exception of exon 5, which encodes both the
transmembrane and cytoplasmic domains as well as the relatively long
3-untranslated sequences. Also conserved is the unusual length of the
first intron. These striking similarities in gene organization and
protein structure provide strong support for the suggestion that the
syndecans arose by gene duplication during mammalian evolution.
The deduced amino acid sequence of N-syndecan predicts a polypeptide with a relative molecular mass of approximately 50 kDa. This is considerably lower than the observed mass of native heparitinase treated N-syndecan of 120 kDa. This discrepancy could result from a combination of anomalous migration of the polypeptide on SDS-polyacrylamide gels, which is a hallmark of syndecan core proteins, as well as the presence of an unidentified posttranslational modification. Interpretation of the migration on SDS gels is also complicated by the fact that N-syndecan core protein forms SDS-resistant dimers and tetramers (18). Deglycosylation of native N-syndecan with heparitinase, nitrous acid, or trifluoromethanesulfonic acid all yield a product with a relative molecular mass of approximately 120 kDa. Digestion with chondroitinase has no effect on N-syndecan electrophoretic mobility (9, 12). The band that is observed following digestion of N-syndecan with these reagents is diffuse, however, suggesting the possible presence of additional modifications. The presence of the mucin homology domain in the N-syndecan core protein suggests the possibility of mucin-type oligosaccharides, but no additional data to support this conclusion have been obtained. Digestion of N-syndecan with neuraminidase has no effect on the core protein migration (data not shown). The cDNA sequence reported here contains sufficient coding information to direct the synthesis of a heparan sulfate proteoglycan that is indistinguishable in overall size and sensitivity to heparitinase from native N-syndecan extracted from neonatal rat brain tissue or Schwann cells.
Immunoblot analysis of proteoglycan steady-state levels and Northern blot analysis of mRNA levels in rat tissues revealed that the highest levels of N-syndecan expression were restricted to cells of the central and peripheral nervous system of neonatal rats. These results do not rule out lower levels of synthesis by other tissues. We have found, for example, that rat heart and arterial tissue contain N-syndecan mRNA. In the heart, N-syndecan mRNA increases dramatically at birth and, in contrast to what was observed in the nervous system, persists in adult animals.2 In spite of this, however, we were unable to detect N-syndecan in heart tissue.
The cell types in the neonatal brain that are responsible for high levels of N-syndecan expression are not known. Temporally, the period of highest expression correlates well with the period of oligodendrocyte differentiation, which in rats begins at the time of birth and persists for the first three weeks of postnatal life (13). Interestingly, these cells express FGF receptors. Administration of exogenous bFGF in culture results in cell proliferation and inhibition of terminal differentiation (13). These effects are prevented when the cells are grown in the presence of sodium chlorate, and the chlorate inhibition can be reversed by exogenous heparin (14). Brain N-syndecan binds bFGF with high affinity (12) and is, thus, a good candidate for the endogenous heparan sulfate molecules that function as co-receptors for bFGF activation.
Additional functions for N-syndecan in the developing nervous system
are suggested by other findings. Brain N-syndecan binds with high
affinity (KD = 5 × 1010
M) to HB-GAM (15), an 18-kDa secreted protein expressed in early postnatal brain and other tissues. Affinity chromatography experiments have shown that N-syndecan is the major protein in brain
with HB-GAM binding activity. Although the exact function of HB-GAM is
not known, it has been shown to have neurite outgrowth activity
in vitro (27). The high affinity and selectivity of N-syndecan binding to HB-GAM and the striking correspondence of their
temporal patterns of expression in brain strongly suggest that these
two proteins together carry out an important
adhesion-dependent activity during early postnatal
development of the brain.
Another interesting question is the physiological function of membrane shedding of N-syndecan. Evidence was presented for a large amount of N-syndecan in brain tissue that was not attached to membranes, consistent with the shedding of syndecans that has been observed with cultured cells. Membrane shedding would provide a mechanism for terminating functional activity that is dependent on attachment to the plasma membrane. This could include cell-cell adhesion activity or binding events that were dependent on cytoskeletal attachment, or generated intracellular signals through the cytoplasmic domain. There is evidence that syndecan cytoplasmic domains can interact with actin filaments (28, 29), and syndecan expression has been shown to alter the morphology and cytoskeletal organization of epithelial cells (30) and Schwann cells (31).
N-syndecan that is released by shedding could also become incorporated into the brain extracellular matrix. Although biochemically and ultrastructurally distinct from collagen-rich fibrous extracellular matrices present in peripheral tissue, there is increasing evidence for the existence of an extracellular matrix in the brain. In adult brain tissue, a prominent component of this matrix appears to be large chondroitin sulfate proteoglycans (32). Association with an extracellular matrix is suggested by the identification of a PBS- and Triton X-100-insoluble pool of N-syndecan in brain tissue. Matrix-associated N-syndecan could provide sites for cell adhesion or pathways for cell migration for cells that express N-syndecan binding proteins on their surface. Alternatively, it could provide a reservoir of heparin-binding growth factors.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U73184[GenBank].