Protein Kinase C (PKC) δ Regulates PKCα Activity in a Syndecan-4-dependent Manner*

The phosphorylation state of Ser183 in the cytoplasmic tail of syndecan-4 determines the binding affinity of the cytoplasmic tail to phosphatidylinositol 4,5-bisphosphate (PIP2), the capacity of the tail to multimerize, and its ability to activate protein kinase C (PKC) α. We sought to identify the kinase responsible for this phosphorylation and to determine its downstream effects on PKCα activity and on endothelial cell function. Among several PKC isoenzymes tested, only PKCα and -δ were able to specifically phosphorylate Ser183 in vitro. However, studies in cultured endothelial cells showed that the phosphorylation level of syndecan-4 was significantly reduced in endothelial cells expressing a dominant negative (DN) PKCδ but not a DN PKCα mutant. Syndecan-4/PIP2-dependent PKCα activity was significantly increased in PKCδ DN cells, while PKCδ overexpression was accompanied by decreased PKCα activity. PKCδ-overexpressing cells exhibited a significantly lower proliferation rate and an impaired tube formation in response to FGF2, which were mirrored by similar observations in PKCα DN endothelial cells. These findings suggest that PKCδ is the kinase responsible for syndecan-4 phosphorylation, which, in turn, attenuates the cellular response to FGF2 by reducing PKCα activity. The reduced PKCα activity then leads to impaired endothelial cell function. We conclude that PKCδ regulates PKCα activity in a syndecan-4-dependent manner.

The protein kinase C (PKC) 1 family of enzymes is one of the most extensively studied group of proteins involved in intracellular signal transduction. However, to date little information is available regarding specific regulation of function and activity of individual PKC isoforms (1)(2)(3)(4). Recently syndecan-4 has been shown to be able to activate PKC␣ in the presence of phosphatidylinositol 4,5-bisphosphate (PIP 2 ) and in the absence of Ca 2ϩ (5)(6)(7). Syndecan-4 is a member of the syndecan gene family, a group of heparan sulfate-carrying core proteins present in the plasma cell membrane (8). While sharing the ability of other syndecans to interact with heparin-binding proteins including fibroblast growth factors (FGFs), vascular endothelial growth factors, and numerous other partners, syndecan-4 has been specifically implicated in FGF2 signaling (9) and in regulation of cell cytoskeleton, focal adhesions, and migration (10,11).
FGF2-dependent activation of syndecan-4 signaling requires oligomerization of its cytoplasmic tails (12) that in turn depends on the phosphorylation state of Ser 183 , which regulates PIP 2 binding to the syndecan-4 tail (6). In previous studies we have demonstrated that phosphorylation of Ser 183 is carried out by a novel PKC (13). This chain of events, therefore, raises the possibility that one PKC isoform controls the activity of another isoform via the regulation of syndecan-4 phosphorylation. The present study was designed to explore that possibility. We found that PKC␦ is the PKC isoform responsible for syndecan-4 phosphorylation and that alterations in PKC␦ activity result in biologically meaningful alterations in PKC␣ function.
Construction of Rat Fat Pad Endothelial Cell (RFPEC)-derived Cell Lines-The dominant negative (DN) PKC␣ construct (15,16) was a generous gift from Dr. Dan Rosson (Lankenau Medical Research Center), and the dominant negative PKC⑀ construct was a gift from Dr. I. Bernard Weinstein (Colombia University). These expression plasmids were generated by replacing the conserved lysine in the ATP binding domain with arginine. PKC␦ cDNA containing a c-Myc tag sequence was subcloned into pRc/CMV vector (Invitrogen) between HindIII and XbaI sites. Dominant negative PKC␦ constructs were created by in vitro mutagenesis replacing the conserved lysine in the ATP binding domain in position 376 with tryptophan. These constructs were stably transfected into RFPECs using LipofectAMINE Plus and the protocol provided by the manufacturer (Invitrogen). Following neomycin (Geneticin, Invitrogen) selection, a number of clones were isolated and expanded. Expression of constructs was verified by Western blotting. At least two different clones were used for each construct-related experiment.
Syndecan-4 Phosphorylation Stoichiometry-Confluent RFPECs were incubated for 24 h in methionine-, sulfate-, and phosphate-free minimum Eagle's medium prepared from the MEM SELECT-AMINE kit (Invitrogen) with 1% bovine serum albumin (Invitrogen) and radiolabeled for 2 h with 2 mCi/ml [ 35 S]methionine (EasyTag Express, PerkinElmer Life Sciences) and 1 mCi/ml [ 32 P]orthophosphoric acid (PerkinElmer Life Sciences). Syndecan-4 was immunoprecipitated and gel-resolved, and the ratio between its 32 P and 35 S incorporation was measured by scintillation counting as described previously (13).
Proliferation Assay-2,000 cells were plated in 96-well tissue culture plates and incubated overnight in 10% FBS, M199 medium. After that cells were starved with 0.5% FBS for 24 h and then treated with 0, 5, or 25 ng/ml FGF2. For measurement of proliferation, 20 l of CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega) was added to the wells and incubated for 2 h. Absorbance at 490 nm was measured using a 96-well plate reader both before FGF2 application and 72 h later.
Matrigel Assay-Matrigel (Becton Dickinson) plates were prepared by adding 0.5 ml of thawed Matrigel to a 12-well tissue culture plate. The gel was allowed to solidify for 1 h at 37°C. 100,000 cells were plated in each well with 25 ng/ml FGF-2 in 0.5% FBS. Cells were imaged after a 24-h incubation at 37°C in a humidified chamber with 5% CO 2 . Analysis of Matrigel results was carried out as described previously (9).

RESULTS AND DISCUSSION
To define the PKC isoform responsible for phosphorylation of Ser 183 in the syndecan-4 cytoplasmic domain, we assayed the ability of all PKC isoforms expressed in RFPECs to phosphorylate Ser 183 in vitro. To determine the PKC isozyme preferentially phosphorylating Ser 183 over other potentially phosphorylatable residues in the syndecan-4 tail, we synthesized a peptide corresponding to the syndecan-4 cytoplasmic domain with Ser 183 replaced by Ala (SA peptide). While all PKC isoforms phosphorylated both the wild type and mutant peptides with similar efficiencies, only PKC␣ and PKC␦ preferentially phosphorylated the Ser 183 site (i.e. preferential phosphorylation of the wild type peptide compared with the SA peptide, Fig.  1A). Since our previously published PKC inhibitor studies (13) strongly argue against PKC␣ as a biologically relevant PKC isoform phosphorylating syndecan-4, we compared the extent of syndecan-4 cytoplasmic domain phosphorylation in growtharrested wild type RFPECs and in an RFPEC-derived cell line stably expressing dominant negative PKC␦, -⑀, or -␣ constructs (Fig. 1B). While PKC␦ dominant negative expression resulted in a 2.5-fold reduction in the stoichiometry of syndecan-4 cytoplasmic tail phosphorylation, the expression of another noncalcium-dependent PKC (PKC⑀) or calcium-dependent PKC␣ had no effect on syndecan-4 phosphorylation stoichiometry. At the same time, overexpression of PKC␦ increased the extent of Ser 183 site phosphorylation (Fig. 1B). Taken together with previously published observations (13), these results identify PKC␦ as the PKC isoenzyme responsible for syndecan-4 phosphorylation.
Since syndecan-4 phosphorylation affects its ability to activate PKC␣ in a PIP 2 -dependent manner, we reasoned that PKC␦ may inhibit PKC␣ activity by phosphorylating the cyto-plasmic tail of syndecan-4. To examine the role of PKC␦ in syndecan-4-dependent regulation of PKC␣ activity, we studied RFPEC-derived cell lines expressing PKC␦ and PKC␣ dominant negative constructs as well as a wild type PKC␦ construct. Since expression of dominant negative isoform-specific PKC constructs can potentially influence activities of PKCs other then the intended target (17), we assayed the activity of all PKC isoforms present in RFPECs expressing PKC␦ or PKC␣ dominant negative constructs. In both cases the dominant negative construct expression resulted in a significant inhibition of the intended target ( Fig. 2A) while not affecting activities of other present PKCs (Fig. 2B). At the same time, overexpression of PKC␦ resulted in a significant increase in PKC␦ activity in cells (Fig. 2C).
Since the extent of syndecan-4 phosphorylation can affect PKC␣ activity, we measured Ca 2ϩ -and PIP 2 -activated PKC␣ activity in wild type RFPECs and in clones expressing wild type or dominant negative PKC␦. While the expression of both constructs had no effect on Ca 2ϩ -activated activity, overexpression of PKC␦ significantly decreased syndecan-4/PIP 2 -activated PKC␣ activity, while expression of PKC␦ dominant neg- ative significantly increased syndecan-4/PIP 2 -activated PKC␣ activity (Fig. 3).
To study whether these changes in PIP 2 -activated PKC␣ activity translate into functionally relevant changes in cell behavior, we measured the ability of FGF2 to induce proliferation of wild type RFPECs or RFPECs expressing either PKC␣ and PKC␦ dominant negative constructs or unmodified PKC␦.
In accord with previously published results (18), expression of the PKC␣ dominant negative construct significantly inhibited cell growth. PKC␦-overexpressing cells also had a significantly lower proliferation rate compared with vector-transfected RF-PECs (Fig. 4A). In fact, the proliferation rate of PKC␦ overexpressors was close to that of cells expressing the PKC␣ dominant negative construct. At the same time, cells expressing the PKC␦ dominant negative construct demonstrated enhanced proliferation compared with vector-transfected cells. One interesting finding was a high rate of growth of cells expressing the PKC␦ dominant negative construct even in the absence of FGF2 presumably because FGF2-activated syndecan-4 phosphatase, in the absence of syndecan-4 phosphorylation, was no longer needed to activate PKC␣ (Fig. 4A). Similar results were obtained in an in vitro Matrigel angiogenesis assay with PKC␦ overexpressors and cells expressing the PKC␣ dominant negative construct demonstrating reduced vascular structure formation compared with vector-transfected RFPECs (Fig. 4B).
To further link the effect of PKC␦ on PKC␣ activity to changes in syndecan-4 Ser 183 site phosphorylation, we determined the stoichiometry of this site in RFPEC-derived cell lines expressing the PKC␦ construct. As expected, PKC␦ overexpression increased base-line syndecan-4 phosphorylation (0.28 Ϯ 0.05 versus 0.19 Ϯ 0.03 mol of phosphate/mol of protein, p Ͻ 0.05; PKC␦ overexpressor versus wild type syndecan-4). At the same time, FGF2 treatment had no appreciable effect on syndecan-4 phosphorylation in PKC␦-overexpressing cells (without FGF2, 0.28 Ϯ 0.05, and with FGF2, 0.27 Ϯ 0.06 mol of phosphate/mol of protein).
The results of this study show that PKC␦ is the kinase responsible for syndecan-4 cytoplasmic domain phosphorylation and that PKC␦ regulates PKC␣ activity via this mechanism. Several observations support these conclusions. The conclusion that PKC␦ is the PKC isoenzyme responsible for Ser 183 phosphorylation in the syndecan-4 cytoplasmic domain is supported by its ability to preferential phosphorylate the Ser 183 site in vitro.
While PKC␣ also preferentially phosphorylated this site in vitro, only PKC␦ had this activity in vivo as demonstrated by a decrease in the extent of Ser 183 phosphorylation in vivo in cells expressing a PKC␦ DN construct and an increase in cells overexpressing PKC␦. At the same time, expression of the PKC␣ dominant negative construct had no effect on syndecan-4 phosphorylation in vivo. It is not clear why the PKC␣ effect on syndecan-4 phosphorylation is different in in vitro versus in vivo settings. PKC␣ does not directly interact with syndecan-4 but rather binds to the syndecan-4-PIP 2 complex (6). It is quite possible that when such a complex is formed, as would be the case in vivo, the 183 Ser site is no longer accessible to PKC␣.
The modulation of the extent of syndecan-4 phosphorylation, achieved by expression of either PKC␦ or PKC␦ dominant neg-  Total cell lysates were immunoprecipitated using anti-syndecan-4 ectoplasmic antibody, and PKC␣ was detected with a specific antibody. ative constructs, affected its ability to activate PKC␣ in the PIP 2 -dependent manner. Interestingly the Ca 2ϩ -dependent PKC␣ activity in cells expressing the wild type PKC␦ or the PKC␦ dominant negative construct was not affected. These alterations in PKC␦ or its dominant negative construct expression (and corresponding changes in syndecan-4 Ser 183 phosphorylation) resulted in significant changes in cellular function as demonstrated by the proliferation and the in vitro Matrigel angiogenesis assays.
The changes in endothelial cell function induced by PKC␦ overexpression in these experiments, inhibition of endothelial cell growth and angiogenesis, are consistent with prior publications including growth inhibition in smooth muscle cells (19), fibroblasts (20), and capillary endothelial cells (21,22). Furthermore, the similarity of the functional effects between PKC␣ overexpression and the expression of a PKC␦ dominant negative construct, accompanied by increased PKC␣ activity, is in agreement with the previously reported positive effects of PKC␣ and inhibitory effects of PKC␦ on endothelial cell migration (21). It is also interesting to note that vascular endothelial growth factor-induced increase in endothelial cell migration and proliferation is accompanied by a decrease in PKC␦ activity (23).
The indirect modulation of the activity of one PKC isoform by another by means of regulation of the syndecan-4 phosphorylation state represents a novel mechanism of modulation of PKC activity. It is also interesting to note that changes in PIP 2 -but not Ca 2ϩ -dependent PKC␣ activation correlated with changes in cell function, suggesting that syndecan-4/PIP 2 -dependent regulation of PKC␣ activity reflects its physiological function.