Regulation of Syndecan-4 Phosphorylation in Vivo *

Recent studies suggest that some of the heparan sulfate-carrying proteoglycans may directly participate in signaling via their cytoplasmic tail. The present investigation addresses the potential involvement of syndecan-4, a widely expressed transmembrane proteoglycan, in this process. We found that the cytoplasmic tail of syndecan-4 is phosphorylated on a single serine residue (Ser183) in growth-arrested NIH 3T3 fibroblasts, with a stoichiometry of 0.3 mol Pi/mol syndecan-4. Treatment of the cells with a protein kinase C (PKC)-activating phorbol ester lead to a 2.5-fold increase in Ser183 phosphorylation. This increase was inhibited by a generic PKC inhibitor but not by an inhibitor specific to the calcium-dependent conventional PKCs, suggesting that the cytoplasmic tail of syndecan-4 is phosphorylated by a calcium-independent novel PKC isozyme. Application of 10–30 ng/ml basic fibroblast growth factor (bFGF) produced a 2–3-fold reduction in the phosphorylation of syndecan-4. Because treatment with the phosphatase inhibitor calyculin prevented the bFGF-induced decrease in syndecan-4 phosphorylation, the effect of bFGF appears to be mediated by a protein serine/threonine phosphatase type 1 or 2A. We conclude that the cytoplasmic tail of syndecan-4 is subject to in vivo phosphorylation on Ser183, which is regulated by the activities of a novel PKC isozyme and a bFGF-dependent serine/threonine phosphatase.

Recent studies suggest that some of the heparan sulfate-carrying proteoglycans may directly participate in signaling via their cytoplasmic tail. The present investigation addresses the potential involvement of syndecan-4, a widely expressed transmembrane proteoglycan, in this process. We found that the cytoplasmic tail of syndecan-4 is phosphorylated on a single serine residue (Ser 183 ) in growth-arrested NIH 3T3 fibroblasts, with a stoichiometry of 0.3 mol P i /mol syndecan-4. Treatment of the cells with a protein kinase C (PKC)-activating phorbol ester lead to a 2.5-fold increase in Ser 183 phosphorylation. This increase was inhibited by a generic PKC inhibitor but not by an inhibitor specific to the calciumdependent conventional PKCs, suggesting that the cytoplasmic tail of syndecan-4 is phosphorylated by a calcium-independent novel PKC isozyme. Application of 10 -30 ng/ml basic fibroblast growth factor (bFGF) produced a 2-3-fold reduction in the phosphorylation of syndecan-4. Because treatment with the phosphatase inhibitor calyculin prevented the bFGF-induced decrease in syndecan-4 phosphorylation, the effect of bFGF appears to be mediated by a protein serine/threonine phosphatase type 1 or 2A. We conclude that the cytoplasmic tail of syndecan-4 is subject to in vivo phosphorylation on Ser 183 , which is regulated by the activities of a novel PKC isozyme and a bFGF-dependent serine/threonine phosphatase.
Although growth factor signaling generally occurs through specific high affinity receptors, several growth factors interact with additional membrane-anchored co-receptors. In particular, bFGF 1 requires binding to a specific sequence of sulfated polysaccharides in the extracellular heparan sulfate glycosaminoglycan (GAG) chain (1) to bind to its high affinity receptor (2) and to exert its effect on target cells (3,4). The current picture of the role of heparan sulfate in the binding mechanism of bFGF consists of dimerization of the growth factor (2,5,6), as well as direct heparan sulfate binding to the high affinity receptor (7,8). Together, these events lead to receptor multi-merization (9) and to tyrosine transphosphorylation of adjacent FGF receptor cytoplasmic tails, followed by phosphorylation of other downstream substrates (10).
Heparan sulfates are primarily associated with two classes of cell surface proteoglycans: syndecans and glypicans. The syndecans are a widely distributed, four-member family of transmembrane proteins capable of carrying both heparan and chondroitin sulfate chains (11)(12)(13). Although there are significant differences between the sequences of their ectoplasmic domains, the four syndecans share a highly conserved cytoplasmic tail containing four invariant tyrosines and one invariant serine (14). This degree of conservation may reflect functional similarities between the cytoplasmic tails of all the syndecans. However, unlike the well established involvement of the ectoplasmic domain in growth factor binding through the GAG chains, there is still no consensus regarding the function of the cytoplasmic tail. Several reports (15,16) point to transient association of the cytoplasmic tail of syndecan-1 to the actin cytoskeleton, which seems to be highly dependent on the presence of one of the four conserved tyrosines (17).
The 28-amino acid-long cytoplasmic tail of syndecan-4 is the least homologous to the other three syndecans, containing a unique nine-residue sequence (RMKKKDEGSYDLG-KKPIYKKAPTNEFYA). Syndecan-4 is incorporated into focal adhesions of fibroblasts in a PKC-dependent manner (18), and its cytoplasmic tail appears to bind and activate PKC␣ (19). These capacities are probably special to the cytoplasmic tail of syndecan-4 and not shared by the other syndecans, because they are mediated through oligomerization of its unique nineresidue sequence (20).
The presence of the five conserved phosphorylatable residues in the cytoplasmic tails of all the syndecans suggests that the cytoplasmic tail could be a kinase substrate in vivo. However, although in vitro phosphorylation by calcium-dependent PKC of serine residues in partial or complete synthetic cytoplasmic tails was reported for syndecan-2 and syndecan-3, it could not be produced for syndecan-1 or syndecan-4 (21,22). Serine phosphorylation in situ was detected in syndecan-2 of carcinoma cells cultured in the presence of serum (23). This phosphorylation was attributed to a serine residue in the cytoplasmic tail of syndecan-2, contained within a sequence that conforms to a phosphorylation motif of cAMP and cGMP-dependent kinases. In situ phosphorylation of the cytoplasmic tail of syndecan-1 was produced in mammary gland cells by treatment with orthovanadate or pervanadate, both of which inhibit tyrosine phosphatase (24). Accordingly, this treatment resulted predominantly in tyrosine phosphorylation, although a lesser degree of serine phosphorylation was also detected. One of the four tyrosines in the cytoplasmic tail of syndecan-1 is contained within a tyrosine kinase phosphorylation motif (25) conserved in all the syndecans and may at least partially account for the orthovanadate and pervanadate-produced phosphorylation.
Given recent reports of the involvement of syndecan-4 in PKC binding and activation (19,20), we set out to investigate whether the syndecan-4 molecule itself is subject to phosphorylation and whether this phosphorylation is affected by bFGF binding to the cell surface.
Isolation of Syndecan-4 Core Proteins-NIH 3T3 cells (American Type Culture Collection, Bethesda, MD) were grown to confluence in 100-mm plates in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (Life Technologies Inc.) at 37°C in a 5% CO 2 humidified atmosphere. The cells were harvested by scraping in 1 ml of lysis buffer (150 mM NaCl, 20 mM NaF, 20 mM Na 4 P 2 O 7 , 5 mM EDTA, 5 mM EGTA, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 1% Triton X-100, 50 mM HEPES, pH 7.4). The lysate was cleared by centrifugation at 9000 ϫ g for 30 min and then subjected to DEAEchromatography as described (26). The eluates were dialyzed twice against 10 mM NH 4 HCO 3 , 1 mM ␤-mercaptoethanol and concentrated by evaporation under vacuum. The concentrated samples were resuspended in 50 l of digestion buffer (50 mM NaCl, 4 mM CaCl 2 , 20 mM Tris, pH 7.4), and GAG chains were cleaved off the proteoglycan core proteins by 4 h of incubation in a mixture of 0.06 unit of chondroitinase ABC and 1 unit each of heparinases I, II, and III (Sigma) at 37°C.
Radiolabeling of Cultured Cells-Confluent NIH 3T3 cells were washed twice in phosphate-free DMEM and incubated for 24 h at 37°C in a 5% CO 2 humidified atmosphere in phosphate-free DMEM supplemented with 0.5% fetal bovine serum. The cells were washed twice with methionine, phosphate, and serum-free DMEM and incubated for 6 h in the same medium, supplemented with 400 Ci/ml [ 35 S]methionine (New England Nuclear, Boston, MA). At the onset of the last 2 h of incubation, 500 Ci/ml [ 32 P]orthophosphoric acid (New England Nuclear) was added to the medium.
Immunoprecipitation of Cytoplasmic and Ectoplasmic Syndecan-4 Domains-Cells were washed with PBS (137 mM NaCl, 10 mM Na 2 HPO 4 , 3.6 mM KCl, 1.8 mM KH 2 PO 4 , pH 7.4), dissociated by 0.05% trypsin, 0.5 mM EDTA (Life Technologies Inc.) in PBS for 10 min at 22°C, and sedimented by 2000 ϫ g centrifugation at 4°C for 5 min. The syndecan-4 ectoplasmic domain was immunoprecipitated from 0.5 ml of medium collected before cell trypsinization or from 0.5 ml of supernatant of the latter centrifugation. The cytoplasmic tail was immunoprecipitated from the pellet after a 30-min extraction at 4°C in 0.5 ml of lysis buffer supplemented with 100 M leupeptin, 2 M pepstatin, and 10 nM okadaic acid (Sigma). Total protein concentrations in each fraction were measured by spectrophotometry at 595 nm (DU 640, Beckman, Fullerton, CA) of an aliquot developed for 10 min in Protein Assay Dye Reagent (Bio-Rad). Bovine serum albumin (Life Technologies Inc.) was used as standard. The medium, trypsinization supernatant, and extracted pellet fractions were precleared by adding 30 l of 1:1 (v/v) slurry of protein G plus/protein A-agarose beads (Calbiochem), and 10 l of nonimmune rabbit serum (Life Technologies Inc.). After a 2-h incubation at 4°C in rotating tubes, the beads were sedimented by 5 min, 5000 ϫ g centrifugation at 4°C. The cleared samples were supplemented with 40 l of 1:1 (v/v) slurry of the above beads and 10 l of rabbit polyclonal antiserum (syndecan-4 ectoplasmic domain-specific antiserum was added to the medium and trypsinization supernatant samples, and cytoplasmic tail-specific antiserum was added to the extracted pellet fraction; both antisera were generous gifts of Dr. N. W. Shworak, MIT (26)) and incubated in rotating tubes for 18 h at 4°C. The agarose beads were collected by centrifugation as above, washed three times in heparinase digestion buffer, and resuspended in 40 l of digestion buffer, and the GAG chains of the bead-attached ectoplasmic domains from the medium and from the trypsinization-supernatant were cleaved as above. The ectoplasmic and cytoplasmic tails were dissociated from the beads by a 10-min incubation in SDS buffer at 95°C, and the beads were sedimented by a 5 min, 13,000 ϫ g centrifugation at 4°C.
Electrophoresis, Transfer, Autoradiography, and Immunoblotting-Immunoprecipitated full-length syndecan-4 core proteins were resuspended in Laemmli sample buffer (2% SDS, 10% glycerol, 0.5% ␤-mercaptoethanol, 0.004% bromphenol blue, 50 mM Tris-HCl, pH 6.8) resolved by SDS-PAGE on a 10% slab gel, and transferred to a polyvinylidene fluoride (PVDF) membrane (Immobilon-P, Millipore, Bedford, MA) for 12 h at 25 mA in 150 mM glycine, 20 mM Tris-HCl, and 20% methanol. The ectoplasmic and cytoplasmic syndecan-4 domains were resolved on a 15% slab gel and transferred for 90 min at 20 mA in 150 mM glycine, 20 mM Tris-HCl, and 30% methanol to a low porosity PVDF membrane (Immobilon-P SQ , Millipore). Radiolabeled bands detected by exposure to film (XAR, Kodak, Rochester, NY) were excised, and their radioactivity was measured in both the 32 P and 35 S spectra by scintillation counting (LS 6000IC, Beckman, Fullerton, CA). In some cases, the same membranes were used for immunoblotting prior to band excision. After blocking in PBS containing 5% nonfat milk powder for 1 h at 22°C, the membrane was incubated in the same solution supplemented with 1:3000 (v/v) dilution of either ectoplasmic or cytoplasmic tail-specific antiserum for 2 h, washed with PBS, and incubated for 1 h in 5% milk powder-PBS containing 1:2000 diluted goat anti-rabbit IgG conjugated to peroxidase (Vector Laboratories, Burlingame, CA). The secondary antibody was detected, after an additional PBS wash, by chemiluminescence (Western Blot Chemiluminescence Reagent Plus, New England Nuclear). Molecular weights were estimated by comparison with the electrophoretic mobility of standards (Kaleidoscope Prestained Standards, Bio-Rad). Densitometry of digitized images of immunoprobed membranes (ScanJet 4c, Hewlett Packard) was performed using ImageQuant software (Molecular Dynamics, Sunnyvale, CA).
Thin Layer Chromatography-Bands excised from PVDF membranes were hydrolyzed for 75 min in 6 N HCl at 110°C. Solvent was evaporated under vacuum, and the sediment was washed thrice with H 2 O. The sediment was resuspended in 5 l of H 2 O after the third evaporation, applied to a thin layer cellulose acetate plate (Sigma-Aldrich), and underwent electrophoresis at 1000 V in 5% acetic acid, 0.5% pyridine, pH 3.0. The radiolabeled phosphoamino acids were detected by phospholuminescence (PhosphorImager, Molecular Dynamics). Phosphorylated, unlabeled Ser, Thr, and Tyr (Sigma) were used as standards and were detected by spraying with ninhydrin.

Phosphorylation of Syndecan-4 Cytoplasmic Tail in Vivo-To
determine the presence and extent of phosphorylation of the syndecan-4 cytoplasmic tail, full-length heparan and chondroitin sulfate-carrying core proteins were isolated from serum-starved, 32 P-labeled NIH 3T3 cells. Autoradiography of NIH 3T3 GAG-lysed core proteins is shown in Fig. 1A (lane 1). To identify the syndecan-4 band, the autoradiogrpahed membrane was probed with an antiserum specific to the cytoplasmic tail of the syndecan-4 core protein (26). The immunoblotting highlighted a single band that ran at an approximate molecular mass of 36 kDa (Fig. 1A, lane 2). A similar syndecan-4 electrophoretic mobility lower than its predicted molecular mass of 20 kDa (14) was observed before with the same antiserum (26). As illustrated in Fig. 1A, the antiserum-detected band superimposed precisely on the second band from the bottom in the autoradiograph.
Phosphoamino acid analysis of the syndecan-4 band produced a single, intensely labeled spot that corresponded to the electrophoretic mobility of phosphorylated serine (Fig. 1B). The syndecan-4 core protein sequence contains multiple serines (16 in the human syndecan-4 (27) and 15 in the rat (14)), all but one of which are located in the ectoplasmic domain. To determine which domain contains the phosphorylated serine, we exploited the susceptibility of the ectoplasmic domain of the syndecans to trypsinization (28). Thus the core protein of syndecan-4 was cleaved at the cell surface concurrently with the trypsin dissociation of the 32 P-labeled, adherent cells from the culture plates.
By analogy with syndecan-1 (28), the trypsinization site is most likely between Arg 147 and Thr 148 preceding the transmembrane domain. Following trypsinization and detergent extraction, the cleaved syndecan-4 fragment was isolated by immunoprecipitation with the cytoplasmic tail-specific antiserum, which recognizes a 14-residue cytoplasmic sequence (26). As with the full-length core protein, the 32 P-labeled immunoprecipitate was separated by SDS-PAGE and transferred to a membrane. The band routinely detected in the autoradiographs of these membranes migrated at an approximate molecular mass of 5 kDa, slightly less than the predicted 7-kDa size of the fragment encompassing the trypsinized transmembrane and cytoplasmic tails ( Fig. 2A). This lower apparent molecular mass may have resulted from partial degradation during the isolation process or may reflect a higher electrophoretic mobility than the molecular mass standard used for estimating the band size. To verify the identity of this band, we reprobed the same membrane with the antiserum that recognizes the cytoplasmic tail. The immunoblotted band overlapped the 32 Plabeled one ( Fig. 2A), confirming that the latter is comprised of the Ser 183 -phosphorylated cytoplasmic tail of syndecan-4.
To rule out phosphorylation of additional serines in the syndecan-4 core protein outside the cytoplasmic tail, we examined the phosphorylation in three different fractions: (a) the medium, which could contain shed ectoplasmic syndecan-4; (b) the supernatant of the sedimented trypsinized cells, containing the cleaved ectoplasmic domain; and (c) the detergent-soluble fraction extracted from the pellet of the sedimentation, containing the transmembrane and cytoplasmic domains. The ectoplasmic domain of syndecan-4 was immunoprecipitated from the first two fractions with an antiserum specific to this domain (26), and the third fraction was immunoprecipitated with the cytoplasmic tail-specific antiserum. Autoradiography of the SDS-PAGE-separated fractions revealed a single band in the cytoplasmic fraction lane. No radioactive bands were detected in the lanes of the other two fractions (Fig. 2B). This clearly localizes the phosphorylation to the single serine residue in the cytoplasmic tail of the core protein of syndecan-4.
To determine the stoichiometry of the basal phosphorylation of syndecan-4, as well as the effect of bFGF and of other compounds on this phosphorylation, the cells were doubly labeled with [ 35 S]methionine and [ 32 P]orthophosphoric acid. The syndecan-4 carboxyl-terminal proteolytic fragment produced by trypsinization between Arg 147 and Thr 148 (referring to the rat sequence numbering) contains a single methionine (Met 176 ). Because we have shown above that the phosphorylation of syndecan-4 occurs at a single Ser 183 , the molar ratio of 32 P/ 35 S, as calculated from their specific activities, should be equivalent to the ratio of mol P i /mol syndecan-4, assuming the two radio probes have similar incorporation efficiencies. Because this quantitation method is ratiometric, the result is independent of the absolute amounts of protein processed. Using this approach, the stoichiometry of the basal phosphorylation of syndecan-4 in cells starved for 24 h in 0.5% serum, followed by 6 h of serum-free starvation, was 0.31 Ϯ 0.12 (mean Ϯ S.D., n ϭ 5) of mol P i /mol syndecan-4.

Regulation of Syndecan-4 Phosphorylation by bFGF and PKC-
The participation of the syndecan ectoplasmic domain in bFGF binding (2, 3) raises the question whether this binding is accompanied by intracellular modifications of syndecan-4, such as phosphorylation of its cytoplasmic tail. Treatment with 10 ng/ml of bFGF during the last 5 h of the serum-free starvation decreased the phosphorylation stoichiometry of syndecan-4 to 0.16 Ϯ 0.02 (n ϭ 5), approximately half its basal level (Fig. 3A). Larger bFGF dosages of 20 and 30 ng/ml further decreased the phosphorylation stoichiometry of syndecan-4 to 0.12 Ϯ 0.06 (n ϭ 3), but this decrease was not statistically different from the effect of 10 ng/ml bFGF (Fig. 3B). To test for the possible involvement of a phosphatase in the bFGF-induced decrease of syndecan-4 phosphorylation, we applied the phosphatase 1/2A inhibitor calyculin (5 nM) to bFGF (10 ng/ml)-treated cells. Calyculin countered the effect of bFGF, maintaining the syndecan-4 phosphorylation at its basal level (Fig. 3B). Moreover, when the same calyculin dose was applied to cells in the absence of bFGF, syndecan-4 phosphorylation was increased more than 2.5-fold relative to the basal level. If, contrary to our assumption, the incorporation efficiency of 35 S is higher than that of 32 P, the bFGF-induced decrease in syndecan-4 phosphorylation could solely result from bFGF up-regulation of syndecan-4 synthesis. To address this possibility, the syndecan-4 expression levels in control and in bFGF-treated cells processed identically to those in the phosphorylation assays were compared by immunoblotting cell lysates containing equal amounts of total protein. The syndecan-4 bands, which similar to immunoprecipitated samples ( Fig. 2A) ran at an approximate molecular mass of 5 kDa, were detected with the antiserum specific to the ectoplasmic domain, and the amount of protein in each band was quantified by densitometry. In cells treated by 10 and by 30 ng/ml bFGF, the level of syndecan-4 expression was 85% (Fig. 3B, inset) and 93% of the control cells, respectively.
The possible involvement of PKC in syndecan-4 phosphorylation is suggested by several reports on functional relationships between the kinase and the proteoglycan (18 -20). To up-regulate PKC, cells were treated with the PKC-activating phorbol ester PMA (0.5 M) during the last 5 h of the serumfree starvation. This treatment increased only the Ser 183 phosphorylation of syndecan-4, without having a detectable effect on the phosphorylation of threonine or tyrosine residues in the cytoplasmic tail (Fig. 4A). The stoichiometry of the phosphorylation of syndecan-4 in the PMA-treated cells was 0.81 Ϯ 0.33 (n ϭ 3), close to 3-fold higher than the basal level. This result indicates that syndecan-4 is either a direct or an indirect PKC substrate.
To further examine the role of PKC in syndecan-4 phosphorylation, we applied the PKC-specific inhibitor chelerythrine to PMA-stimulated cells. The phosphorylation of syndecan-4 started to decline at chelerythrine concentrations above 1.5 M and was reduced to an undetectable level at 6 M chelerythrine (Fig. 4, B and C). The latter concentration is less than 10% of the IC 50 of chelerythrine for the inhibition of protein tyrosine kinases (29). Although supporting the role of PKC in the phosphorylation of syndecan-4, these results do not identify the specific isozyme involved, because both PMA and chelerythrine affect all the four known calcium-dependent cPKCs, as well as the five calcium-independent nPKCs (30). To further narrow down the group of possible PKC isozymes, we applied the indolocarbazole Gö 6976, which selectively inhibits calcium-dependent PKC isozymes (31) to PMA (0.5 M)-treated cells. The phosphorylation of syndecan-4 was not reduced, however, by Gö 6976 concentrations up to 100 nM, more than 10-fold its IC 50 for cPKC (data not shown). It is likely, therefore, that the syndecan-4 cytoplasmic tail is phosphorylated by one of the nPKC isozymes. DISCUSSION We have shown that the cytoplasmic tail of syndecan-4 is phosphorylated in cultured fibroblasts and that the extent of its phosphorylation is determined by activities of a nPKC enzyme and a bFGF-activated phosphatase. The phosphorylation site was localized to Ser 183 , immediately upstream of a nine-amino acid segment involved in binding to and activation of PKC␣ (19). Phosphorylation of a cytoplasmic serine residue was previously detected in the cytoplasmic tail of syndecan-2 (23) and to a lesser extent in syndecan-1 (24). In our studies we observed a relatively high degree of syndecan-4 phosphorylation in growth-arrested cells, which could be further increased by treatment with PMA or decreased by bFGF. Because Ser 183 is part of an invariant seven-residue sequence (KKDEGSY), these findings may be relevant to all four members of the syndecan family.
The PMA-induced increase in the phosphorylation of syndecan-4 and its decrease by chelerythrine strongly suggest the involvement of PKC in this phosphorylation. Syndecan-4 could not be phosphorylated in vitro, however, by cPKC isozymes (21,22). In agreement with this observation, we were unable to suppress the PMA-induced phosphorylation of syndecan-4 by a cPKC-specific inhibitor, pointing to the participation of a nPKC isozyme in the phosphorylation. Although the amino acid sequence around Ser 183 in the cytoplasmic tail of syndecan-4 does not ideally fit a particular PKC isozyme-specific substrate sequence motif (32), it does contain the glycine residue that immediately precedes the phosphorylatable serine/threonine in the motifs of all the PKC isozymes. The nPKC isozymes ␦, ⑀, and were observed to be membrane-associated in NIH 3T3 fibroblasts (33), the same cell type used in the present study. The substrate sequence motif of PKC␦ (AKRKRKGSFFYGG, (32)) has the highest similarity among all PKC isozymes to the amino acid sequence around Ser 183 in syndecan-4.
A phosphatase inhibitor reversed the bFGF-induced reduction in syndecan-4 phosphorylation observed in our study. This suggests that bFGF binding up-regulates a phosphatase and/or down-regulates a kinase involved in controlling the level of Ser 183 phosphorylation. This effect could be mediated either through the bFGF high affinity tyrosine kinase receptor or through the syndecan-4 molecule. Tyrosine phosphorylation was reported to inhibit PKC␦ (34), although an opposite stimulatory effect of this phosphorylation has also been observed (35). Alternatively, dimerization of bFGF molecules bound to heparan sulfate chains of adjacent syndecan-4 core proteins could cross-link these molecules. This would facilitate transdephosphorylation of their cytoplasmic tails by a putative tailassociated phosphatase, similar to the association of tyrosine phosphatase to the cytoplasmic tails of growth factor receptors (10). Although the identity of the phosphatase cannot be determined from our data, its susceptibility to calyculin indicates that it is likely to be a serine/threonine phosphatase type 1 or 2A.
Although these findings demonstrate multi-factorial regulation of syndecan-4 cytoplasmic tail phosphorylation, the functional impact of this event is not known. Recent findings have suggested that syndecan-4 may play an important role in regulating the distribution and activity of PKC␣. These functions are mediated via the oligomerization of a unique nine-amino acid domain (Leu 186 -Lys 194 ) starting three residues downstream of the phosphorylated serine (19). The state of Ser 183 phosphorylation may conceivably affect syndecan-4 oligomerization. Indeed, phosphorylation of a cysteine residue appended to the amino terminus of a Leu 186 -Lys 194 synthetic peptide reduced the tendency of this peptide to oligomerize in vitro (20). The location of the phosphorylated cysteine is only two residues downstream of Ser 183 . Thus, the bFGF-induced dephosphorylation of the syndecan-4 cytoplasmic tail on Ser 183 , which we have detected in vivo, may be required for oligomerization of the core protein, as well as for other possible processes. These may include co-assembly of syndecan and high affinity bFGF receptors into signaling complexes (9,12), association with the actin cytoskeleton (17), and recruitment into focal adhesions (18). These and other potential consequences of syndecan-4 phosphorylation are currently being investigated.