Bisindolylmaleimide I Suppresses Fibroblast Growth Factor-mediated Activation of Erk MAP Kinase in Chondrocytes by Preventing Shp2 Association with the Frs2 and Gab1 Adaptor Proteins*

Fibroblast growth factors (FGFs) inhibit chondrocyte proliferation via the Erk MAP kinase pathway. Here, we explored the role of protein kinase C in FGF signaling in chondrocytes. Erk activity in FGF2-treated RCS (rat chondrosarcoma) chondrocytes or human primary chondrocytes was abolished by the protein kinase C inhibitor bisindolylmaleimide I (Bis I). Bis I inhibited FGF2-induced activation of MEK, Raf-1, and Ras members of Erk signaling module but not the FGF2-induced tyrosine phosphorylation of Frs2 or the kinase activity of FGFR3, demonstrating that it targets the Erk cascade immediately upstream of Ras. Indeed, Bis I abolished the FGF2-mediated association of Shp2 tyrosine phosphatase with Frs2 and Gab1 adaptor proteins necessary for proper Ras activation. We also determined which PKC isoform is involved in FGF2-mediated activation of Erk. When both conventional and novel PKCs expressed by RCS chondrocytes (PKCα, -γ, -δ, and -ϵ) were down-regulated by phorbol ester, cells remained responsive to FGF2 with Erk activation, and this activation was sensitive to Bis I. Moreover, treatment with PKCλ/ζ pseudosubstrate lead to significant reduction of FGF2-mediated activation of Erk, suggesting involvement of an atypical PKC.

Activating mutations in fibroblast growth factor receptor 3 (FGFR3) cause several human dwarfisms characterized by diminished long bone growth (1). In cartilage, FGFR3 alters chondrocyte proliferation and differentiation by up-regulation of cell cycle inhibitors and stimulation of cartilage matrix degradation (2)(3)(4)(5). The anti-proliferative action of FGF 2 signaling in cartilage contrasts with the usual mitogenic response of cells to FGF stimulus (6), but the molecular basis of this paradox remains unclear. Recently, Erk MAP kinase was found as a candidate for FGFR3-mediated inhibition of chondrocyte proliferation and differentiation (7)(8)(9)(10).
FGF signaling in chondrocytes leads to long term Ras/Erk activation, which appears to account for the growth inhibitory outcome of FGF treatment (8,9). To date, little is known about chondrocyte properties of FGF signaling permitting prolonged Erk activity, although slow down-regulation of mutated FGFR3 appears to be involved (30,31). The requirement of PKC for sustained Ras/Erk signaling prompted us, in this study, to investigate the role of PKC in FGF signaling in chondrocytes.
Signal Transduction Studies-Cells were serum-starved for 12 h before treatment with 10 ng/ml FGF2 (R&D Systems, Minneapolis, MN) for 30 min in the presence of heparin (1 g/ml; Invitrogen). When Bis I, Gö6983, Gö6976, GSK3␤ inhibitor I (4-benzyl-2-methyl-1,2,4-thiadiazolidine-3,5-dione), H89, Raf1 inhibitor I (5-iodo-3-((3,5-dibromo-4-hydroxyphenyl) methylene)-2-indolinone), Ro-31-8220, SU5402 (Calbiochem), U0126, or rapamycin (Cell Signaling) were used, cells were treated for 30 min prior to FGF2 treatment. When PMA (200 nM; Sigma) was used, cells were treated for 5 min or 12 h, respectively. The phosphorylation status of Raf-1, Erk, and MEK was detected by WB using the antibodies described above. Erk activity was determined using a kinase assay kit (Cell Signaling). Briefly, Erk was immunoprecipitated from 200 g of total protein and incubated with recombinant Elk-1 in the presence of ATP. Phosphorylation of Elk-1 at Ser-383 was determined by WB. For the Raf-1 kinase assay, Raf-1 immunocomplexes were washed with Raf-1 kinase buffer (25 mM Tris, pH 7.5, 5 mM ␤-glycerolphosphate, 2 mM dithiothreitol, 0.1 mM Na 3 VO 4 , 10 mM MgCl 2 ), and the kinase reaction was performed for 30 min at 30°C in the presence of 20 M ATP and 500 ng of recombinant MEK1 (Santa Cruz Biotechnology) in 40 l of kinase buffer. MEK1 phosphorylation was determined by WB. For the FGFR3 kinase assay using recombinant FGFR3 intracellular domain, the Frs2 immunocomplexes were washed with FGFR3 kinase buffer (60 mM HEPES-NaOH, pH 7.5, 3 mM MgCl 2 , 3 mM MnCl 2 , 3 M Na 3 VO 4 , 1.2 mM dithiothreitol), and the kinase reaction was performed for 30 min at 30°C in the presence of 2.5 g of polyethylene glycol, 10 M ATP, and 300 ng of recombinant FGFR3 intracellular domain (Glu-322-Thr-806; Cell Signaling) in 50 l of kinase buffer. For the FGFR3 kinase assay using full-length FGFR3, Chinese hamster ovary cells were transfected with vectors carrying C-terminally FLAG-tagged human wild-type or K650E-FGFR3. Ninety-six hours later, FGFR3 was immunoprecipitated with FLAG antibody. Immunocomplexes were washed with FGFR3 kinase buffer, and the kinase reaction was performed for 30 min at 30°C in the presence of 10 M ATP and 150 ng of recombinant Frs2 (Abnova, Taipei City, Taiwan) in 40 l of kinase buffer. Ras and Rap1 activation was determined using a Ras or the Rap-1 activation assay (Upstate Biotechnology). Briefly, active GTPase was purified from cell lysates using the agarose-bound glutathione S-transferase fusion protein containing the Rasbinding domain of Raf-1 or Rap1-binding domain of Ral GDS and detected by WB. Myristoylated atypical PKC pseudosubstrate (N-myristoyl-SIYRRGARRWR KL) was obtained from Biomol (Plymouth, PA). Cells were treated with pseudosubstrate for 30 min, treated with 10 ng/ml FGF2 for 30 min, and analyzed for active Erk by WB.
Vectors, Cell Transfection, and Cell Sorting-The pRK7-FGFR3 vector was made by cloning the full-length human wildtype or K650E-FGFR3 cDNA into the HindIII site of pRK7. The FLAG-tagged FGFR3 constructs were prepared by PCR amplification of pRK7-FGFR3 segment between the MluI site and the 3Ј end using primers that added the FLAG tag and a BamHI site immediately 3Ј of the stop codon (5Ј-CTGGAGTCCAACGC-GTCCATGAGCTC-3Ј and 5Ј-GTTGGGGATCCAGTGGCC-CTTCACTTATCGTCGTCATCCTTGTAATCCATCGTCC-GCGAGCCCCCACTGC-3Ј). The PCR product was recloned into the pRK7 vector at the MluI and BamHI sites. Cells were transfected with FuGENE 6 (Roche Diagnostics, Penzberg, Germany) according to the manufacturer's protocol. For cell sorting, cells were co-transfected with a GFP-expressing vector (pCCEY) and pRK7-FGFR3 vector in a 1:3 ratio. Twenty-four hours later, the GFP-positive cells were isolated using a FACStarϩ cell sorter (BD Biosciences) and plated.

Bis I Inhibits FGF2-mediated Activation of Erk MAP Kinase in Chondrocytes-
To determine whether PKC is involved in FGF signaling in chondrocytes, we asked whether the FGF2mediated activation of Erk MAP kinase was sensitive to the chemical inhibition of PKC. We treated RCS chondrocytes with FGF2 alone or together with three different PKC inhibitors (Gö6983, Bis I, Gö6976) and determined their effect on the FGF2-mediated Erk activity. Fig. 1A shows that Bis I almost entirely abolished Erk activation at concentrations higher than 5 M in contrast to Gö6983 or Gö6976 that only partially modulated the Erk activation. Since sustained Erk activation underlies the growth inhibitory effect of FGF signaling in chondrocytes (8,9), we determined the effect of Bis I on long term FGF2-mediated Erk activation. Fig. 1B demonstrates that Bis I inhibits both short term and long term FGF2-mediated Erk activation in RCS cells. Similar data were obtained with human primary chondrocytes (Fig. 1C). The effect of Bis I on Erk activ-ity is not direct since no inhibition of Erk activity was detected by kinase assay in the presence of up to 50 M Bis I (Fig. 1D).
Bis I Inhibits FGF2-mediated Activation of the Erk Pathway Upstream of Ras- Fig. 1 demonstrates that Bis I inhibits activation of Erk by FGF signaling upstream of Erk. We asked at which level Bis I targets the Erk module. FGF2 treatment led to activatory phosphorylation of both MEK and Raf-1, and this phosphorylation was abolished by Bis I (Fig. 2, A and B). In the following experiment, Raf-1 was immunoprecipitated from cells treated with FGF2 alone or in combination with Bis I and probed for its kinase activity using recombinant MEK1 as a substrate. Raf-1 kinase activity was inhibited by Bis I, thus confirming that Bis I targets the Erk module either at the level or upstream of Raf-1 (Fig. 2C). Next, we tested whether Bis I targets Raf-1 directly. Active Raf-1 was immunoprecipitated from FGF2-treated cells and subjected to a kinase assay with Bis I added into the kinase reaction. No significant Raf-1 inhibition was detected in the presence of up to 50 M Bis I (Fig. 2D), suggesting that Bis I inhibits the Erk module upstream of Raf-1.
It is well established that FGF receptors (FGFR) activate the Erk module via Frs2-mediated recruitment of Grb2-Sos complexes that in turn activate Ras (33). Recently, an alternative mode was described in nerve growth factor-treated PC12 cells whereby Frs2 recruits Crk-C3G complexes and activates Erk through Rap1 GTPase and B-Raf kinase (34). We therefore determined whether FGF signaling in RCS cells utilizes Ras or Rap1 for Erk activation. Although FGF2 triggered significant and prolonged Ras activity, a weak, if any, activation of Rap1 was detected (Fig. 3A). FGF2-mediated activation of Ras was abolished by Bis I (Fig. 3B).
Bis I Does Not Inhibit FGFR3- Fig. 3B demonstrates that Bis I inhibits FGF2-mediated activation of the Erk pathway upstream of Ras. We asked whether Bis I inhibits FGFR3 kinase itself. First, we determined the FGF2-and Bis I-induced changes in the tyrosine phosphorylation status of Frs2, which reflects FGFR3 activity. Fig. 4A shows that FGF2 treatment led to an increase of Frs2 tyrosine phosphorylation that was not FIGURE 1. Bis I inhibits FGF2-mediated activation of Erk in cells but not in a kinase assay. A, RCS cells were serum-starved for 12 h, treated with PKC inhibitors for 30 min prior to FGF2 treatment (30 min), and analyzed for Erk activatory phosphorylation by WB. The WB signal was quantified by densitometry and graphed. The levels of total Erk2 serve as a loading control. Circled P indicates phosphorylation. B and C, RCS cells (B) or (C) human primary chondrocytes were serum-starved for 12 or 24 h, pretreated with 10 M Bis I for 30 min prior to FGF2 treatment for the indicated times, and analyzed for Erk activatory phosphorylation by WB. The levels of total Erk2 or Erk1/2 serve as a loading controls. D, RCS cells were serum-starved for 12 h and treated with FGF2 for 30 min. Active Erk was immunoprecipitated and analyzed for kinase activity with Elk as a substrate, and Bis I was added into the kinase reaction. Elk phosphorylation was monitored by WB. DMSO, Me 2 SO. FIGURE 2. Bis I inhibits FGF2-mediated activation of Raf-1 in cells but not in a kinase assay. A and B, RCS cells were serum-starved for 12 h, treated with Bis I for 30 min prior to FGF2 treatment (30 min), and analyzed for MEK1 or Raf-1 activatory phosphorylation by WB. The levels of total MEK1 and Raf-1 serve as a loading control. Circled P indicates phosphorylation. C, RCS cells were treated as indicated, and Raf-1 was immunoprecipitated (IP) and analyzed for its kinase activity using MEK1 as a substrate. Raf-1-mediated phosphorylation of MEK1 was determined by WB. The levels of total Raf-1 and MEK1 serve as a control for immunoprecipitation and substrate quantity, respectively. D, RCS cells were serum-starved for 12 h and treated with FGF2 for 30 min. Raf-1 was immunoprecipitated and subjected to a kinase assay with MEK1 as a substrate and Bis I, or a Raf-specific inhibitor was added into the kinase reaction. MEK1 phosphorylation was monitored by WB. The levels of total Raf-1 and MEK1 serve as controls for immunoprecipitation and substrate quantity, respectively. DMSO, Me 2 SO. inhibited by Bis I. However, a significant electrophoretic mobility shift of Frs2 was induced by FGF2 that was eliminated by Bis I (Fig. 4B). This shift appears to be caused by Erk-mediated phosphorylation since it was nearly eliminated by the MEK inhibitor U0126.
The tyrosine phosphorylation status of Frs2 (Fig. 4A) represents an indirect method of monitoring the FGFR3 activity. Since we could only detect weak Frs2 tyrosine phosphorylation and Bis I is structurally similar to staurosporine (both staurosporine and its derivatives such as PKC412 inhibit FGFRs (35)), we considered the evidence in Fig. 4A insufficient to rule out the possibility that Bis I directly inhibits FGFR3. Although we routinely detected endogenous FGFR3 in RCS cells by WB, we failed to immunoprecipitate enough FGFR3 to monitor the levels of FGFR3 activity directly by WB with antibody recognizing Tyr-653/Tyr-654-phosphorylated FGFR or 4G10 phospho-tyrosine antibody (not shown). When transfected into the RCS cells, no transgenic FGFR3 was detected 72 h later. When FGFR3 was co-transfected with GFP-expressing vector and GFP-positive cells were isolated by cell sorting 24 h later, they showed poor attachment and no growth in contrast to cells transfected with GFP vector alone, which grew normally (not shown), thus demonstrating the inhibitory effect of FGFR3 overexpression on RCS growth.
We therefore probed the Bis I inhibitory activity toward FGFR3 by a kinase assay utilizing the recombinant, kinase active intracellular domain of FGFR3 and Frs2, immunoprecipitated from RCS cells, as a substrate. Fig. 4C shows that in this system, Bis I did not significantly inhibit the Frs2 tyrosine phosphorylation induced by recombinant FGFR3. This contrasted with the FGFR-specific inhibitor SU5402 that completely inhibited Frs2 tyrosine phosphorylation. Next, we tested the ability of Bis I to inhibit the activity of full-length FGFR3. C-terminally FLAG-tagged wild type or constitutively active FGFR3 mutant (K650E) was transfected into Chinese hamster ovary cells, immunoprecipitated, and subjected to a kinase assay with recombinant Frs2 as a substrate, and Bis I was added into the kinase reaction. Fig. 4D shows that Bis I did not inhibit FGFR3 activity at concentrations up to 20 M, although partial inhibition was apparent at 50 M.
Bis I Prevents FGF2-induced Shp2 Association with Frs2 and Gab1 Adaptor Proteins- Fig. 4 shows that FGFR3 is not directly inhibited by Bis I concentrations sufficient to abolish FGF2mediated activation of Erk pathway in cells, suggesting that Bis I inhibits the Erk pathway immediately upstream of Ras (Figs. 1,  3, and 4). To activate Ras, FGFRs phosphorylate several adaptor proteins such as Frs2, Gab1, Shc, and Shb, which recruit Ras guanine nucleotide exchange factor Sos, complexed with Grb2 or Shp2-Grb2, to the site of FGFR activation (33, 36 -38). We therefore asked whether Bis I affects the assembly of adaptor signaling complexes. FGF2 treatment led to stable formation of Frs2-Shp2-Grb2 and Gab1-Shp2-Grb2 complexes, whereas Shc recruited mostly Grb2 (Fig. 5A). No significant amounts of Grb2 or Shp2 associated with the Shb upon the FGF2 treatment (not shown). Next, we asked whether FGF2-triggered Shp2 and Grb2 binding to Frs2, Gab1, and Shc is sensitive to Bis I. Fig. 5B shows that Bis I impairs binding of both Shp2 and Grb2, although Shp2 binding appears to be affected to a greater extent.
Next, we determined which PKC isoform is involved in FGF2-mediated Erk activation. At the protein level, RCS cells expressed PKC␣, -␥, -␦, -⑀, -, and - (Fig. 6A). We first asked whether phorbol ester-responsive PKCs are involved in Erk activation by FGF2. Brief PMA treatment led to potent Erk activation that was completely inhibited by 5 M of Gö6983 or Bis I, demonstrating the ability of PMA-responsive PKCs to activate Erk (Fig. 6B) (14). Chronic PMA treatment led to FIGURE 4. Bis I does not inhibit FGFR3 kinase activity. A, RCS cells were treated as indicated, and Frs2 was immunoprecipitated (IP) and analyzed for its tyrosine phosphorylation status by WB (4G10 antibody). The levels of active and total Erk1/2 in lysates used for Frs2 immunoprecipitation are shown. The 4G10 signal was quantified by densitometry and graphed. Note that although the level of FGF2-induced tyrosine phosphorylation of Frs2 appears unaffected by Bis I, its electrophoretic mobility increases. Circled P indicates phosphorylation. B, RCS cells were serum-starved for 12 h, treated with Bis I, the FGFR inhibitor SU5402, or the MEK inhibitor U0126 for 30 min prior to FGF2 treatment (30 min), and probed for Frs2 and activated Erk1/2 by WB. The levels of total Erk2 serve as a loading control. Note the marked electrophoretic mobility shift of Frs2 that is inhibited by Bis I, SU5402, and also by U0126, suggesting that it is due to Erk-mediated phosphorylation. I.O.D., integrated optical density. C, the effect of Bis I on recombinant FGFR3 kinase domain using Frs2 immunoprecipitated from RCS cells as a substrate and Bis I or SU5402 added to the kinase reaction. Frs2 phosphorylation was determined by WB. The total levels of Frs2 serve as a control for substrate quantity. DMSO, Me 2 SO. D, the effect of Bis I on immunoprecipitated FLAG-tagged wild-type (wt) or K650E-mutated FGFR3 using recombinant Frs2 as a substrate and Bis I or SU5402 added to the kinase reaction. The levels of FGFR3 autophosphorylation and Frs2 phosphorylation were determined by WB. The levels of total FGFR3 and Frs2 serve as loading controls. Ϫ ATP indicates no ATP added into the kinase reaction.
down-regulation of PKC␣, -␥, -␦, and -⑀ (42), and their absence was functionally confirmed by brief PMA treatment, which failed to activate Erk (Fig. 6, A and B). When cells with downregulated PKC␣, -␥, -␦, and -⑀ were treated with FGF2, they responded with the usual levels of Erk activation that remained sensitive to Bis I, suggesting that a PMA-unresponsive aPKC (PKC and/or PKC) is involved in FGF2-mediated Erk activation in RCS cells (Fig. 6, A and C). To challenge this hypothesis, we determined the effect of cell-permeable, aPKC-specific pseudosubstrate peptide (SIYRRGARRWRKL) on FGF2-mediated Erk activation. The pseudosubstrate suppresses PKC catalytic activity by blocking the substrate-binding site and thus can be used as an isoform-specific PKC inhibitor (43). Treatment with PKC/ pseudosubstrate inhibited FGF2-mediated Erk activation in RCS cells (Fig. 6D).

DISCUSSION
Although it is well established that FGF signaling inhibits chondrocyte growth, the molecular mechanism of this effect is not clearly defined. Considering the usual mitogenic outcome of FGF signaling, the effect of the FGF signal in chondrocytes appears to be unique. It was recently shown that the Ras/Erk pathway is a candidate for FGF-mediated growth arrest, and in contrast to most other cell types, the FGF stimulus elicits a remarkably prolonged Erk activation in chondrocytes (8,9). This correlates with the known cellular responses to Erk, where transient activation appears to be crucial for mitogenic signaling of growth factors, but sustained activity frequently leads to growth arrest (44). Therefore, a part of the unique chondrocyte response to FGF may lie in their ability to maintain sustained Erk activity following the FGF stimulus. The molecular mechanism of this feature is unknown. The role of PKC in sustained or in constitutive activation of Erk in growth factor and oncogenic Ras signaling (26 -29) prompted us to explore the role of PKC in FGF signaling in chondrocytes.
We found that the PKC inhibitor Bis I abolishes FGF2-mediated activation of Erk in chondrocytes. Both short term and long term Erk activity were equally inhibited by Bis I with somewhat higher effectiveness than either FGFRspecific or MEK-specific inhibitors, suggesting a critical role for Bis I-inhibited molecule in both the initiation and the maintenance of Erk activation following FGF2 treatment (Figs. 1 and  4). Bis I inhibited FGF2-mediated MEK activity, demonstrating that it targets the Erk pathway upstream of MEK. It is well established that FGFRs activate the Erk pathway through phos- phorylation of the adaptor protein Frs2, which subsequently recruits multiple Grb2-Sos complexes that in turn activate Erk through Ras and Raf-1 (33). Recently, an alternative mechanism was described in nerve growth factor-treated PC12 cells, where Frs2 recruits Crk-C3G complexes that activate Rap1 GTPase, which employs B-Raf as the MEK kinase. Importantly, the Crk-C3G/Rap1/B-Raf but not the Grb2-Sos/Ras/Raf-1 pathway accounted for prolonged Erk activation, leading to growth arrest and neuronal differentiation (34). Since the RCS and PC12 models appear similar in several aspects of their response to both FGF and nerve growth factor (8,(45)(46)(47), we determined whether FGF signaling in RCS cells utilizes Ras or Rap1 for Erk activation. Although FGF treatment leads to prolonged Ras activation, it activates Rap1 very weakly, if at all (Fig. 3). Since Raf-1 but not B-Raf is activated by FGF2-treatment in RCS cells (5), it appears that Raf-1 is the principal MEK kinase employed by FGF signaling, and its activation is mediated by Ras (Figs. 2  and 3). In RCS cells, Ras activity was inhibited by Bis I, showing that it targets the Erk pathway upstream of Ras (Figs. 2 and 3).
We further ruled out direct inhibition of FGFR3 kinase activity by Bis I (Fig. 4), thus leaving the signal relay from FGFR3 to Ras as an area of Bis I action.
Unlike other receptor tyrosine kinases, activated FGFRs do not directly recruit Grb2-Sos but rather use adaptors that provide docking sites for Grb2-Sos or Shp2-Grb2-Sos complexes that in turn activate Ras. Although Frs2 is the major adaptor employed by FGFRs to activate Erk, they can utilize several other adaptors including Gab1, Shc, and Shb (33, 36 -38). We demonstrate that at least three different adaptors, Frs2, Gab1, and Shc, are employed by FGFR3 to recruit Grb2 and Shp2 in RCS cells. Although the importance of a particular adaptor for activation of the Erk module in RCS cells is unclear, it appears that both Gab1 and Shc recruit more Grb2 when compared with Frs2 (Fig. 5A). In cells treated with Bis I prior to FGF2, both Shp2 and/or Grb2 binding to Frs2, Gab1, and Shc was impaired, albeit to different degrees. Although Shp2 binding was completely inhibited by Bis I, the inhibition of Grb2 was only partial (Fig. 5B). Since both Frs2 and Gab1 recruit Grb2 either directly (Grb2-Sos complex) or via Shp2 (Shp2-Grb2-Sos complex), it is likely that binding of only Shp2complexed Grb2 is inhibited by Bis I. Frs2, Gab1, and particularly Shc recruited significant amounts of Grb2 in the presence of Bis I (Fig. 5B), thus leading to the question of whether the inhibition of Shp2 translocation accounts for the Bis I inhibition of the Erk pathway. Shp2 is known to be a critical mediator of Erk activation in FGF signaling that operates by disabling the action of two potent negative regulators of the Ras/Erk pathway, Sprouty and RasGAP (45, 48 -50). Since failure of Shp2 translocation to Frs2 leads to potent inhibition of Erk activation in FGF signaling (45), we conclude that the effect of Bis I on Shp2 translocation is a candidate for its inhibitory effect on the Erk activation in RCS cells (Fig. 7).
How does Bis I execute its effect? In vitro, Bis I showed no direct inhibition of three out of four kinases directly involved in the Erk pathway (Erk, Raf-1, and FGFR3; Figs. 1, 2, and 4). Furthermore, we excluded a role for the known PKC-unrelated Bis I targets cAMP-dependent protein kinase, GSK-3, p70S6K, and MAPKAP-1␤ kinases (39 -41) in Erk activation in RCS cells (not shown). Although our data do not rule out the possibility of , murine tissue biopsy is given to control for antibody. The levels of actin serve as a loading control. Note the complete down-regulation of PMA-responsive PKC variants in contrast to PKC and -, which were not affected by PMA. B, RCS cells were treated as indicated, and activated Erk1/2 was determined by WB. The WB signal was quantified by densitometry and graphed. Note that PKC-mediated Erk activation by short PMA treatment was not observed in cells chronically treated with PMA, functionally confirming the down-regulation of PMA-responsive PKCs. The levels of total Erk2 serve as a loading control. Circled P indicates phosphorylation. I.O.D., integrated optical density. C, RCS cells were treated as indicated, and activated Erk1/2 was determined by WB. The levels of total Erk2 serve as a loading control. Note that cells with down-regulated PMA-responsive PKCs still respond to FGF2 with Erk activation that is sensitive to Bis I. D, RCS cells were serum-starved for 12 h, treated with an atypical PKC pseudosubstrate (aPKC phosphatidylserine; specific to PKC/ and PKC) for 30 min prior to FGF2 treatment, and analyzed for activated Erk by WB. The levels of total Erk2 serve as a loading control. The WB signal was quantified by densitometry and graphed. E, RCS cells were serum-starved for 12 h and then treated as indicated, and PKC was immunoprecipitated (IP). The immunocomplexes were probed for Shp2 and PKC by WB.
a yet unknown target of Bis I being involved in FGF2-mediated Erk activation, they suggest the involvement of PKC.
We therefore attempted to identify the nature of the PKC isoform(s) involved in FGF2-mediated Erk activation. PMAinduced down-regulation of conventional PKC and novel PKC variants (␣, ␥, ␦, and ⑀) expressed in RCS cells did not affect the FGF2-mediated activation of Erk or sensitivity of this activation to Bis I (Fig. 6, A-C), thus excluding PMA-responsive PKCs from regulation of the Erk pathway. With the exception of PKC, which we were unable to detect with a commercially available antibody, the PKC isoforms that remain present in RCS cells after chronic PMA administration are aPKC and - (Fig. 6A). The FGF-mediated activation of Erk was completely inhibited by higher Bis I concentrations that are necessary to inhibit PKC and - (Fig. 1A) (51), and the treatment with PKC/-specific pseudosubstrate abolished FGF-mediated Erk activation (Fig. 6D), both demonstrating the role of an aPKC. It is unlikely that PKC is involved in FGF-mediated activation of Erk since Gö6983, which is a potent PKC inhibitor (52), had only minor effect on FGF-mediated Erk activation (Fig. 1A). In addition, PKC but not PKC co-immunoprecipitated with Shp2 in RCS cells (Fig. 6E), suggesting that PKC is the candidate PKC involved in FGF-mediated activation of Erk in RCS cells.
How does the PKC regulate Shp2? In an inactive state, the N-SH2 domain of Shp2 occupies the catalytic cleft of the proteintyrosine phosphatase domain, resulting in mutual allosteric inhibition. In contrast, the C-SH2 domain is available to interact with tyrosyl-phosphorylated ligands, such as Frs2 or Gab1. During ligand binding, both SH2 domains are ultimately engaged in phosphorylated-tyrosyl binding, thus relieving the inhibition of the protein-tyrosine phosphatase domain. Alternatively, Shp2 can be tyrosine phosphorylated at Tyr-542/Tyr-580, creating docking sites for its own SH2 domains that lead to protein-tyrosine phosphatase activation (53). Since some Shp2 ligands, such as Gab1, are also its substrates (50), the first mode of Shp2 activation ensures proximity to its substrate. The alternative mode, in turn, results in Shp2 activation without simultaneous localization to its substrate. It has been demonstrated that PKCs phosphorylate Shp2 in the vicinity of Tyr-542/Tyr-580 (Ser-576/Ser-591 (54)), potentially disturbing the binding of SH2 domains and interfering with the alternative mode of Shp2 activation. Since FGF signaling activates Shp2 both ways (55), PKC action may shift the equilibrium toward the substrate (i.e. Frs2 or Gab1)-bound Shp2. Experiments are now ongoing to test this hypothesis.
It is well established that Frs2 represents not only the critical mediator of Erk activation in FGF signaling but also a site of negative feedback (56,57). Erk phosphorylates Frs2 on multiple threonine residues, leading to diminished recruitment of Grb2 and down-regulation of Ras/Erk activity (57). In RCS cells, FGF2 induced rapid, Erk-mediated phosphorylation of the entire Frs2 moiety. Surprisingly, however, this Frs2 was able to recruit significant amounts of both Shp2 and Grb2 for the entire length of FGF2 treatment (Fig. 5). Whether the Erk-mediated feedback does not affect the capacity of Frs2 to recruit Grb2/Shp2 in RCS cells or whether we are observing residual Grb2/Shp2 recruitment after the negative feedback has taken place is a matter for future investigation. It is, however, likely that insufficient Frs2/Erk-negative feedback contributes to prolonged Erk signaling in FGF2-treated RCS chondrocytes.
Taken together, we show that Bis I inhibits the FGF-mediated Erk activation in chondrocytes distal to FGFR, through disturbing the assembly of its proximal signaling complexes (Fig. 5). In RCS cells, the Erk pathway was likely activated by both FGFR2 and FGFR3 (32). Since all FGFRs appear to use a similar mechanism to activate Erk, Bis I action might not be limited to chondrocytes. In support of this hypothesis, we found that Bis I inhibits the FGF2-mediated activation also in 3T3 fibroblasts, HeLa, and Chinese hamster ovary cells, which are not only of a different origin but also express FGFR1 in addition to FGFR2 (not shown).