RAFTK/Pyk2 activation is mediated by trans-acting autophosphorylation in a Src-independent manner.

The related adhesion focal tyrosine kinase (RAFTK), also known as Pyk2, undergoes autophosphorylation upon its stimulation. This leads to cascades of intracellular signaling that result in the regulation of various cellular activities. However, the molecular mechanism of RAFTK autophosphorylation is not yet known. Using various RAFTK constructs fused with two different tags, we found that the autophosphorylation of RAFTK was mediated by a trans-acting mechanism, not a cis-acting mechanism. In addition, overexpression of kinase-mutated RAFTK inhibited wild type RAFTK autophosphorylation in a dose-dependent manner by a trans-acting interaction. Trans-acting autophosphorylation was also observed between endogenous and exogenous RAFTK upon potassium depolarization of neuroendocrine PC12 cells. Using immunoprecipitation and affinity chromatography, we detected RAFTK self-association that was not affected by deletion of a single region or domain of RAFTK. Furthermore, RAFTK autophosphorylation occurred only at site Tyr402 in a Src kinase activity-independent manner. However, Src significantly enhanced RAFTK-mediated paxillin phosphorylation, suggesting a key role for Src in RAFTK activation and phosphorylation of downstream substrates. Our results indicate that the activation of RAFTK occurs in several steps. First, upon stimulus, RAFTK trans-autophosphorylates Tyr402. Second, phosphorylated Tyr402 recruits and activates Src kinase that in turn phosphorylates RAFTK and enhances its kinase activity. Lastly, the enhanced RAFTK activity induces the activation of downstream signaling molecules. Taken together, these studies provide insights into the molecular mechanism of RAFTK autophosphorylation and the specific role of Src in the regulation of RAFTK activation.

The related adhesion focal tyrosine kinase (RAFTK), also known as Pyk2, undergoes autophosphorylation upon its stimulation. This leads to cascades of intracellular signaling that result in the regulation of various cellular activities. However, the molecular mechanism of RAFTK autophosphorylation is not yet known. Using various RAFTK constructs fused with two different tags, we found that the autophosphorylation of RAFTK was mediated by a trans-acting mechanism, not a cis-acting mechanism. In addition, overexpression of kinase-mutated RAFTK inhibited wild type RAFTK autophosphorylation in a dose-dependent manner by a trans-acting interaction. Trans-acting autophosphorylation was also observed between endogenous and exogenous RAFTK upon potassium depolarization of neuroendocrine PC12 cells. Using immunoprecipitation and affinity chromatography, we detected RAFTK self-association that was not affected by deletion of a single region or domain of RAFTK. Furthermore, RAFTK autophosphorylation occurred only at site Tyr 402 in a Src kinase activity-independent manner. However, Src significantly enhanced RAFTK-mediated paxillin phosphorylation, suggesting a key role for Src in RAFTK activation and phosphorylation of downstream substrates. Our results indicate that the activation of RAFTK occurs in several steps. First, upon stimulus, RAFTK trans-autophosphorylates Tyr 402 . Second, phosphorylated Tyr 402 recruits and activates Src kinase that in turn phosphorylates RAFTK and enhances its kinase activity. Lastly, the enhanced RAFTK activity induces the activation of downstream signaling molecules. Taken together, these studies provide insights into the molecular mechanism of RAFTK autophosphorylation and the specific role of Src in the regulation of RAFTK activation.
The related adhesion focal tyrosine kinase (RAFTK) 1 (1), also known as Pyk2 (2), CAK-␤ (3), and CADTK (4), is a nonreceptor tyrosine kinase related to focal adhesion kinase (FAK). RAFTK transduces key extracellular signals through tyrosine phosphorylation, leading to various cellular responses, and is also implicated in the regulation of numerous cellular activities (5,6). FAK and RAFTK exhibit ϳ48% amino acid identity (65% similarity) and have a similar domain structure: a unique N terminus, a centrally located protein tyrosine kinase domain, and two proline-rich regions at the C terminus (6,7). Neither kinase contains SH2 and SH3 domains. Analyses of FAK activation events have shown that phosphorylation of FAK occurs at six sites in vivo: two sites (Tyr 397 and Tyr 407 ) within the FAK N-terminal region; two sites (Tyr 576 and Tyr 577 ) within the kinase domain activation loop; and two sites (Tyr 861 and Tyr 925 ) within the C-terminal region. Four FAK tyrosine phosphorylation sites (Tyr 397 , Tyr 576 , Tyr 577 , and Tyr 925 ) are conserved at analogous positions in RAFTK (Tyr 402 , Tyr 579 , Tyr 580 , and Tyr 881 ). Among these four tyrosine phosphorylation sites, Tyr 402 is known to be autophosphorylated (2). It is well established that RAFTK autophosphorylation plays a key role in various cellular signaling processes, such as mitogen-activated protein kinase activation mediated by G protein-coupled receptors (8), the activation of T cells through T cell antigen receptor cross-linking (9), nephrocystin signaling in a subset of renal epithelial cells (10), integrin-mediated osteoclast motility (11), association of RAFTK with nerve terminals (12), cardiac remodeling mediated by the reorganization of focal adhesion contacts (13), and adhesion-induced osteoclast spreading and bone resorption (14). Upon stimulation, RAFTK is known to autophosphorylate Tyr 402 , which recruits Src through binding of the Src-SH2 domain to the RAFTK-Tyr 402 site, leading to the activation of Src (6,7). The activated Src, in turn, phosphorylates RAFTK at Tyr 579 and Tyr 580 , which enhances the activity of RAFTK (6,15,16). Although the autophosphorylation of RAFTK is reported to play an important role in various systems, the detailed mechanism of this event, such as the nature and sites of its autophosphorylation, are unknown.
In this study, we focused on the following questions: 1) Is the autophosphorylation of RAFTK a cis-or trans-acting event? 2) What are the essential components needed for RAFTK autophosphorylation? 3) Is Tyr 402 the only autophosphorylation site? 4) Is RAFTK self-associated similar to the heterodimerization of other receptor-type tyrosine kinases? Better understanding of the initial steps of RAFTK activation will provide insights into RAFTK-mediated cellular functions, such as cell adhesion and migration, stress response, and induction of neuronal long term potentiation.
Chemicals and Antibodies-Protease inhibitors and all other reagents were purchased from Sigma. Protein G-Sepharose and recombinant Protein G-agarose were purchased from Pierce and Invitrogen, respectively. Normal rabbit serum and normal mouse serum were obtained from Sigma. Mouse anti-phospho-specific Tyr(P) 402 -RAFTK/ Pyk2 and anti-phosphotyrosine antibodies (4G10) were kind gifts from Upstate Biotechnology, Inc. Mouse anti-FLAG antibody (M2) was purchased from Sigma. Mouse anti-paxillin and anti-RAFTK/Pyk2 antibodies were purchased from Transduction Laboratory. Mouse anti-Myc tag antibody and goat anti-RAFTK/Pyk2 antibodies (N terminus; N- 19) were purchased from Santa Cruz Biotechnology. Mouse anti-GFP antibody was purchased from Clontech or Medical and Biological Laboratory (MBL). Horseradish peroxidase-conjugated sheep anti-mouse Ig antibodies and donkey anti-rabbit Ig antibodies were obtained from Amersham Biosciences. Secondary antibodies of horseradish peroxidase-conjugated rabbit anti-goat Ig were obtained from Santa Cruz Biotechnology. Polyclonal antibodies against RAFTK were generated by immunizing New Zealand White rabbits with a bacterially expressed fusion protein consisting of glutathione S-transferase and human RAFTK cDNA subcloned into the pGEX-2T expression vector as described (17).
Transient Transfection of PC12 Cells, SYF, and SYF ϩ Src Cells-The RAFTK cDNA in the pcDNA3-neo vector was constructed as described in our previous studies (1,17). A kinase-negative mutant of RAFTK (KM) was constructed by replacing Lys 475 with an Ala residue using a site-directed mutagenesis kit (Clontech, Palo Alto, CA). A point mutant of RAFTK (Y402F) was also constructed by replacing Tyr 402 with a Phe residue using site-directed mutagenesis. The GFP-tagged wild type (WT) or kinase mutant (KM) RAFTK was subsequently prepared by subcloning the RAFTK construct as an EcoRI/EcoRI fragment in-frame at the 3Ј end of a GFP tag in a pEGFP-C3 vector, according to the manufacturer's protocol (Clontech). Various RAFTK deletion mutants were generated using human wild type RAFTK as a template for PCR, followed by subsequent cloning as a HindIII/ClaI fragment inframe at the 5Ј end of a FLAG tag in pcDNA3. The RAFTK deletion mutant del C-FLAG encoding amino acids 1-679 was amplified by PCR using the forward primer 5Ј-GGAATTCCTAAGCTTACCATGTCTGGG-GTGTCCGAGCCCC and the reverse primer 5Ј-GCTCTAGACATCGA-TGAGGCTGCACACCAGCTCGGTGAAGC. The RAFTK deletion mutant del N-FLAG encoding amino acids 415-1009 was amplified with the forward primer 5Ј-TTGCATAAGCTTACCATGGGTCCACAGTATG-GCATTGCCC and the reverse primer 5Ј-GCTCTAGACATCGATGGC-CTCTGCAGGTGGGTGGGCCAG. The RAFTK deletion mutant del NK-FLAG encoding amino acids 680 -1009 was amplified with the forward primer 5Ј-TTGCATAAGCTTACCATGTCTGGGGTGTCCGAG-CCCC and the reverse primer 5Ј-GCTCTAGACATCGATGAGTGACG-TTTATCAGATGGAGAAG. Lastly, the RAFTK deletion mutant del NC-FLAG encoding amino acids 415-679 was amplified with the forward primer 5Ј-TTGCATAAGCTTACCATGGGTCCACAGTATGGCATT-GCCC and the reverse primer 5Ј-GCTCTAGACATCGATGAGGCTGC-ACACCAGCTCGGTGAAGC. PC12 cells, SYF, and SYF ϩ Src cells were plated on poly-D-lysine-coated plates a day before transfection and then transiently transfected with various cDNA constructs using Lipo-fectAMINE 2000 reagent according to the manufacturer's instructions (Invitrogen). The cell lysates were prepared with modified RIPA buffer and immunoprecipitated with specific antibodies as indicated. The immunoprecipitates were washed and analyzed for tyrosine phosphorylation as described below.
Transient Transfection of HEK 293 Cells-Transient transfection of cDNA constructs into HEK 293 cells was performed using the calcium phosphate precipitation method. Empty vector cDNA was used as a transfection control. After 24 or 48 h of transfection, the cells were examined for expression of the tagged GFP proteins under an immunofluorescent microscope. The cell lysates were prepared with mod-ified RIPA buffer and immunoprecipitated with specific antibodies, followed by immunoblotting as described below.
Preparation of Cell Lysates, Immunoprecipitations, and Immunoblotting-The cells were lysed in modified RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 0.25% sodium deoxycholate, and 1 mM EDTA) containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 g/ml aprotinin, 1 g/ml leupeptin, and 1 g/ml pepstatin) and phosphatase inhibitors (1 mM Na 3 VO 4 and 1 mM NaF). Immunoprecipitations were performed with the indicated antibodies, followed by protein G incubation. The immunoprecipitates were separated by either 8% SDS-PAGE or 7% Tris-acetate-NuPAGE (Novex, San Diego, CA) under reducing conditions, electrophoretically transferred to Immobilon polyvinylidene difluoride (Millipore, Bedford, MA), and processed for immunoblotting using the enhanced chemiluminescence technique (Amersham Biosciences). Levels of tyrosine phosphorylation, Tyr 402 -specific phosphorylation, and the amounts of RAFTK protein were quantified using a software program for densitometry analysis (Un-Scan-It; Silk Scientific Corp.). Each band was normalized to the total RAFTK level in the respective sample.
In Vitro Pull-down Assay-The fast protein liquid chromatography affinity chromatography column was prepared with mouse anti-FLAG® M2 antibody-attached agarose beads according to the manufacturer's instructions (Sigma). HEK 293 cells were co-transfected with 2 g of FLAG-tagged WT RAFTK and 2 g of GFP-tagged WT RAFTK, followed by preparation of whole cell lysates with modified RIPA lysis buffer. 4 ml of the cell lysates (1 mg/ml) were loaded on anti-FLAG affinity column beads. After extensive washing with Tris-buffered saline buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.4), each 1-ml fraction of elutes was collected after loading 5 ml of FLAG® peptide (100 mg/ml in Tris-buffered saline buffer; Sigma) for competitive elution according to the manufacturer's instructions (Sigma). The collected fraction samples were resolved on 8% SDS-PAGE followed by immunoblotting.

Autophosphorylation of RAFTK Requires Both Intact RAFTK
Kinase Activity and the Tyrosine Residue at 402-RAFTK, a FAK subfamily member, is an intracellular tyrosine kinase that mediates tyrosine phosphorylation in various signaling pathways (6). Upon extracellular stimulation, such as by cell adhesion, intracellular Ca 2ϩ increase, or growth factors, RAFTK undergoes autophosphorylation followed by subsequent molecular interaction and the activation of downstream signaling cascades. To analyze the molecular mechanism of RAFTK autophosphorylation, various deletion and point mutations of RAFTK were constructed (Fig. 1A). HEK 293 cells have no endogenous expression of RAFTK. These cells were transfected with the RAFTK constructs and analyzed for the tyrosine phosphorylation status of RAFTK by immunoblotting with anti-phosphotyrosine (4G10) and autophosphorylation at Tyr 402 by immunoblotting with phospho-specific anti-Tyr(P) 402 -RAFTK antibody. As shown in Fig. 1B, although wild type RAFTK (WT-FLAG and WT-GFP) was highly phosphorylated, a kinase mutant of RAFTK (KM-GFP) and RAFTK with a point mutation (tyrosine to phenylalanine) of the autophosphorylation site at 402 (Y402F-FLAG) were not phosphorylated. The pan RAFTK tyrosine phosphorylation was correlated with the RAFTK Tyr 402 phosphotyrosine. This demonstrates that intact kinase activity and the tyrosine residue 402 were essential for RAFTK autophosphorylation. These data were further supported by the expression of an N terminus deletion mutant (del N-FLAG), which showed no phosphorylation because of the lack of the Tyr 402 site, although it contained an intact kinase domain. Interestingly, autophosphorylation was detected in a C terminus deletion mutant (del C-FLAG) that contained the kinase domain and the autophosphorylation site at Tyr 402 , suggesting that the C terminus region was not essential for RAFTK autophosphorylation at residue 402.
RAFTK Autophosphorylation Was Mediated by a Trans-acting Mechanism Not by a Cis-acting Mechanism-Autophosphorylation of a kinase can be accomplished by two different mechanisms. One is intramolecular cis-autophosphorylation and the other is intermolecular trans-autophosphorylation. We hypothesized that if autophosphorylation can be mediated by a transacting mechanism, the interaction between RAFTK kinase activity and the tyrosine residue at 402 present in different RAFTK molecules would mediate the autophosphorylation of RAFTK. However, if the autophosphorylation is mediated by a cis-acting mechanism, the two components, RAFTK kinase activity and the tyrosine residue at 402, should be present in a single RAFTK molecule. To investigate this hypothesis, we utilized co-expression of two RAFTK constructs differentially fused with either FLAG epitope (an eight-amino acid peptide, DYKDDDDK) or GFP. As shown in Fig. 2A, a GFP-tagged kinase mutant RAFTK (KM-GFP) was phosphorylated at Tyr 402 by co-expressed FLAG-tagged wild type RAFTK (WT-FLAG), suggesting the existence of trans-acting autophosphorylation of RAFTK. Confirming the trans-autophosphorylation of RAFTK, co-expression of Y402F-FLAG with KM-GFP showed that intact kinase domain-containing Y402F-FLAG phosphorylated KM-GFP ( Fig. 2A), although neither of the constructs could be autophosphorylated alone (Fig. 1B). Based on the tyrosine phosphorylation level of KM-GFP, the transautophosphorylation activity of WT-FLAG was much greater than that of Y402F-FLAG ( Fig. 2A).
Next, we examined whether trans-autophosphorylation is the only autophosphorylation mechanism or whether cis-autophosphorylation also plays a role in RAFTK autophosphorylation. We reasoned that if the autophosphorylation of RAFTK was mediated by a trans-acting mechanism, overexpression of KM-GFP would inhibit the autophosphorylation of WT-FLAG through competition. Therefore, the level of WT-FLAG autophosphorylation could be determined by the relative amounts of the kinase-inactive molecule (KM-GFP) upon their co-expression. However, if the autophosphorylation occurs in a single molecule through a cis-acting mechanism, the autophosphorylation of WT-FLAG would be preserved at a certain level of phosphorylation regardless of KM-GFP co-expression. To this end, we co-transfected WT-FLAG and KM-GFP in various doses to examine the potential competition between the two constructs. As shown in Fig. 2B, KM-GFP inhibited the autophosphorylation of WT-FLAG at Tyr 402 in a dose-dependent manner, and overexpression of KM-GFP almost completely abolished the autophosphorylation of WT-FLAG, indicating that RAFTK autophosphorylation is a trans-acting event not a cis-acting event. As a control, we examined the effect of the del N-FLAG RAFTK construct on WT-FLAG autophosphorylation upon their co-expression. As shown in Fig. 2C, del N-FLAG did not inhibit the autophosphorylation of WT-FLAG at Tyr 402 , indicating that inhibition of WT-FLAG autophosphorylation was a specific process dependent on the kinase-defective KM-GFP. To confirm that the inhibition of RAFTK wild type autophosphorylation by the RAFTK kinase mutant was through a trans-acting interaction and not through inhibition of RAFTK subcellular localization by the focal adhesion targeting domain at its C terminus, we examined the effects of KM-GFP on the autophosphorylation of del C-FLAG that lacks a C terminus region (Fig. 1A). In support of the trans-acting interaction of  (Fig. 2D), suggesting that the KM-mediated inhibition of autophosphorylation is not dependent on the C terminus, which controls the subcellular localization of RAFTK to focal adhesion sites. Although we as well as other groups have reported that the RAFTK kinase mutant worked as a dominant negative (2, 8, 18 -20), its detailed mechanism of action is unknown. The intermolecular trans-phosphorylation of RAFTK provides a molecular basis for the observation that the RAFTK kinase mutant dominant-negatively inhibits wild type RAFTK autophosphorylation. Overexpression of exogenous RAFTK might produce clustering of the protein independent of subcellular localization. It may also explain why overexpression of RAFTK by transient transfection or viral infection leads to a high level of RAFTK phosphorylation. However, under physiological conditions, extracellular stimuli may provide a signal that results in the clustering of RAFTK and trans-autophosphorylation. For example, potassium depolarization of neuroendocrine PC12 cells induced a distinct redistribution of RAFTK at point contacts, which serve as activation centers for the recruitment of various signaling molecules, such as Src, paxillin, and Grb2. 2 Trans-acting Autophosphorylation of Endogenous RAFTK upon Potassium Depolarization in PC12 Cells-To investigate trans-autophosphorylation of endogenous RAFTK, we utilized rat pheochromocytoma PC12 cells that express high levels of endogenous RAFTK (2). The interaction between exogenous GFP-tagged kinase-mutated RAFTK (KM-GFP) and endogenous RAFTK was monitored by examining autophosphorylation of KM-GFP upon potassium depolarization. As shown in Fig. 3A, immunoblotting with 4G10 and anti-Tyr(P) 402 -RAFTK antibodies revealed that autophosphorylation of KM-GFP was induced in the presence of endogenous RAFTK upon KCl stimulation. GFP-tagged wild type RAFTK (WT-GFP) showed constitutively autophosphorylated RAFTK (Fig. 3A). Although en-2 S.-Y. Park and S. Avraham, unpublished data.

FIG. 2. RAFTK auto-phosphorylation was mediated by transautophosphorylation.
A, HEK 293 cells were co-transfected with 1 g of GFP-tagged kinase-mutated K457A (KM-GFP) RAFTK and 2 g of FLAG-tagged wild type (WT-FLAG) or Y402F mutant (Y402F-FLAG) RAFTK, followed by immunoprecipitation (IP) and immunoblotting (IB) with specific antibodies as indicated. B and C, KM-GFP inhibited the autophosphorylation of WT-FLAG in a dose-dependent manner. HEK 293 cells were co-transfected with 0.2 g of FLAG-tagged wild type (WT-FLAG) RAFTK plus various amounts of GFP-tagged kinase-mutated K457A (KM-GFP) RAFTK (B) or FLAG-tagged N terminus deletion mutant RAFTK (del N-FLAG) (C), followed by immunoprecipitation with mouse anti-FLAG (M2; Sigma) or goat anti-RAFTK (N terminus-specific) (N-19; Santa Cruz Biotechnology.). After resolving on gel electrophoresis, the proteins were immunoblotted with specific antibodies, as indicated. Immunoblotting of GFP antibody with whole cell lysates (WCL) was performed to show the expression of GFP-tagged KM-GFP. D, KM-GFP inhibited the autophosphorylation of del C-FLAG in a dose-dependent manner. HEK 293 cells were co-transfected with 0.2 g of the FLAG-tagged C-terminal deletion mutant of RAFTK (del C-FLAG) plus various amounts of GFP-tagged kinase-mutated K457A (KM-GFP) RAFTK, followed by immunoprecipitation and immunoblotting with specific antibodies, as indicated. The phosphotyrosine ratio (pY/RAFTK or Tyr(P) 402 /RAFTK) demonstrates the relative amounts of phosphorylation at each tyrosine-or Tyr 402 -specific phosphorylation in the samples.

FIG. 3. Trans-acting autophosphorylation of endogenous RAFTK upon potassium depolarization in PC12 cells. A, PC12
cells were transfected with GFP-tagged WT RAFTK or KM RAFTK cDNA (4 g) using LipofectAMINE 2000 (Invitrogen). Empty vector pEGFP-C3 (V-GFP) cDNA was used as a control. After a 24-h incubation, the cells were unstimulated or stimulated with KCl (60 mM, 5 min). The cells were harvested with RIPA buffer followed by immunoprecipitation (IP) with mouse anti-GFP antibodies (MBL) and immunoblotting (IB) with specific antibodies, as indicated. The phosphorylation of endogenous RAFTK was examined by immunoprecipitation with rabbit anti-RAFTK antibody and immunoblotting with specific antibodies, as indicated. B, HEK 293 cells were transfected with GFPtagged WT RAFTK or KM RAFTK cDNA (2 g) using the calcium phosphate method. Empty vector pEGFP-C3 (V-GFP) cDNA was used as a control. After 24 h of incubation, the cells were unstimulated or stimulated with KCl (60 mM, 5 min). Whole cell lysate preparation, immunoprecipitation, and immunoblotting procedures were the same as described above (for A). The phosphotyrosine ratio (pY/RAFTK or Tyr(P) 402 /RAFTK) demonstrates the relative amounts of phosphorylation at each tyrosine-or Tyr 402 -specific phosphorylation in the samples.
dogenous RAFTK showed induction of phosphorylation upon KCl stimulation in PC12 cells transfected with either vector-GFP (V-GFP) or KM-GFP, endogenous RAFTK showed induction of phosphorylation in cells transfected with WT-GFP in the absence of KCl stimulation (Fig. 3A). These results demonstrate that trans-acting autophosphorylation of RAFTK also occurs between exogenous and endogenous RAFTK. As a control, we expressed GFP-tagged WT or KM RAFTK in 293 cells that have no endogenous RAFTK expression. We found that exogenous GFP-tagged kinase-mutated RAFTK (KM-GFP) did not induce any autophosphorylation either in the presence or absence of KCl stimuli in the RAFTK-deficient HEK 293 cells (Fig. 3B), whereas the GFP-tagged wild type RAFTK (WT-GFP) showed constitutive autophosphorylation (Fig. 3B). These data support the notion that the autophosphorylation of kinase-mutated RAFTK requires endogenous RAFTK activity.
Association between Differentially Tagged RAFTKs-Next, to further characterize the molecular interaction of RAFTK in trans-autophosphorylation, we investigated association between differentially tagged RAFTKs. FLAG-tagged wild type RAFTK (WT-FLAG) and GFP-tagged wild type RAFTK (WT-GFP) were co-immunoprecipitated in a dose-dependent manner upon their co-expression (Fig. 4A). This suggests that physical association of two RAFTK molecules can be detected during trans-autophosphorylation. To confirm the association between the differentially tagged RAFTKs, we analyzed this association by an in vitro pull-down assay using affinity chromatography with anti-FLAG antibody-conjugated agarose. As shown in Fig.  4B, WT-GFP was specifically associated and eluted with WT-FLAG. This supports the co-immunoprecipitation result (Fig.  4A). The relative amount of co-associated WT-GFP with WT-FLAG was very small, suggesting that the self-association of RAFTK was weak and transient.
Analysis of RAFTK Regions Mediating Self-association-To further investigate the self-association region(s) in RAFTK, we analyzed the association of WT-GFP with various deletion mutants using immunoprecipitation. Both the N-and C-terminal deletion mutants associated with wild type RAFTK (WT-GFP), albeit at a significantly decreased level as compared with the WT-FLAG and KM-FLAG, indicating that the N-or C-terminal region was partially responsible for the association with WT-GFP (Fig. 5). Even deletion of the N terminus and Kinase domain (del NK) or of the kinase domain alone (del NC) showed some association. This suggests the possibility of multiple association regions. Because many receptor-type tyrosine kinases demonstrate a formation of stable dimers and trans-acting autophosphorylation (21), we attempted to detect the formation of stable dimers using nondenatured native gel electrophoresis, a sucrose density gradient, and chemical coupling with various cross-linkers. However, we did not observe the formation of a stable complex using any of these approaches (data not shown). Although RAFTK induces trans-acting autophosphorylation as shown in receptor-type tyrosine kinases (21), our data indicate that the trans-interaction of RAFTK does not form stable dimers. Thus, this suggests that RAFTK self-association is a transient event with more dynamic spatial and temporal responses. The Trans-autophosphorylation of RAFTK Is a Src Kinaseindependent Process-Autophosphorylated RAFTK induces the activation of Src kinase through the association of RAFTK phosphotyrosine at 402 and the Src-SH2 domain (8). Activated Src, in turn, mediates the phosphorylation of RAFTK at several tyrosine residues in the kinase domain and the C terminus region of RAFTK, which leads to increased RAFTK kinase activity and the recruitment of signaling molecules, such as Grb2 (6,8). To analyze whether Src also plays a role in RAFTK trans-autophosphorylation through either increased RAFTK kinase activity or direct phosphorylation of Tyr 402 , we utilized two cell lines: a murine embryonic fibroblast cell line lacking Src family kinases (SYF) and a c-Src-reintroduced cell line (SYF ϩ Src). First, we examined the autophosphorylation of various RAFTK constructs upon single transfection. As shown in Fig. 6A (first panel with 4G10 antibody), in SYF cells only the WT-FLAG and del C-FLAG showed the tyrosine phosphorylation, whereas in SYF ϩ Src cells the KM-GFP, Y402F-FLAG, and del N-FLAG as well as WT-FLAG and del C-FLAG showed significant tyrosine phosphorylation. However, Src did not affect the autophosphorylation level of RAFTK at Tyr 402 (Fig. 6A, second panel with anti-Tyr(P) 402 -RAFTK-specific antibody), although Src significantly increased the tyrosine phosphorylation of RAFTK. These data suggest that Src does not affect RAFTK autophosphorylation either directly or indirectly. To examine the expression of Src in these two cell lines, whole cell lysates were analyzed by immunoblotting with Src antibody. Lack of Src expression was observed in the SYF cells, and positive expression of Src was found in the SYF ϩ Src cells (Fig.  6A, fifth panel). Interestingly, immunoblotting with anti-phosphospecific Tyr(P) 418 -Src antibody against the active form of Src kinase demonstrated that Src is constitutively activated in the SYF ϩ Src cells (Fig. 6A, sixth panel). Because c-Srcreintroduced SYF ϩ Src cells were immortalized by simian virus 40 large T antigen from murine embryonic fibroblasts, SV40 large T antigen might activate c-Src and lead to the phosphorylation of RAFTK in the absence of interactions with Tyr 402 of RAFTK. This possibility is consistent with another report investigating the interaction of v-Src with FAK, which showed that the FAK Tyr 397 autophosphorylation site was not required for v-Src to stably interact with and phosphorylate FAK (22). In comparison, because HEK 293 cells do not have constitutively activated Src, only autophosphorylated RAFTK can activate Src, which is followed by Src-mediated phosphorylation of the Tyr 579/580 residues (Fig. 1B). It is important to note that in the SYF ϩ Src cells, overexpression of Src induced the tyrosine phosphorylation of KM-GFP, Y402F-FLAG, and del N-FLAG other than at site Tyr 402 (compare first and second After 48 h, the cells were harvested with RIPA buffer, followed by immunoprecipitation (IP) and immunoblotting (IB) with specific antibodies, as indicated. A, autophosphorylation of various RAFTK constructs in SYF or SYF ϩ Src cells. The cell lysates (500 g) were immunoprecipitated with either goat anti-FLAG antibody (for del C-FLAG) or rabbit anti-RAFTK antibody (for all others), followed by gel electrophoresis and immunoblotting as indicated. The same membrane was used for multiple probings after membrane deprobing. To check the expression and activation of Src kinase, whole cell lysates (WCL) were resolved by gel electrophoresis, followed by immunoblotting with rabbit anti-Src antibody (Santa Cruz) and rabbit anti-Tyr(P) 418 -Src antibody (BioSource International, Inc). B, trans-autophosphorylation of KM-GFP by co-transfection of WT-FLAG or del C-FLAG. KM-GFP proteins were immunoprecipitated with anti-GFP antibodies, followed by immunoblotting with 4G10 antibody (Upstate Biotechnology Inc.) to check tyrosine phosphorylation, anti-Tyr(P) 402 -RAFTK-specific antibody (Upstate Biotechnology Inc.) to check Tyr 402 phosphorylation, and mouse anti-RAFTK antibody (Transduction Laboratory) to check KM-GFP protein levels. C, co-transfection of Y402F-FLAG with various FLAG-or GFP-tagged RAFTK cDNA constructs. Y402F-FLAG proteins were immunoprecipitated with anti-FLAG antibodies, followed by immunoblotting with 4G10 antibody to check tyrosine phosphorylation and mouse anti-RAFTK antibody (Transduction Laboratory) to check Y402F-FLAG protein levels. The percentage of phosphotyrosine ratio (Tyr(P) 402 /pY, %) demonstrates the relative amounts of Tyr 402 -specific phosphorylation to pan tyrosine phosphorylation in the samples. panels of Fig. 6A showing pan phosphotyrosine 4G10 antibody and Tyr(P) 402 phosphospecific antibody). Next, we examined the trans-autophosphorylation of RAFTK upon co-transfection of KM-GFP with various RAFTK constructs. As shown in Fig.  6B (first panel with 4G10 antibody), in SYF cells the KM-GFP was phosphorylated at a low level by co-transfection of WT-FLAG, Y402F-FLAG, del C-FLAG, and del N-FLAG, whereas in SYF ϩ Src cells the KM-GFP was significantly increased in phosphorylation by co-transfection of WT-FLAG, Y402F-FLAG, del C-FLAG, and del N-FLAG. However, Src did not significantly affect the trans-autophosphorylation of KM-GFP at Tyr 402 (Fig. 6B, second panel with anti-Tyr(P) 402 -RAFTK-specific antibody), although Src did significantly increase the tyrosine phosphorylation of KM-GFP. These data suggest that Src does not affect RAFTK autophosphorylation either directly or indirectly.
Trans-autophosphorylation of RAFTK Occurred Only at Tyr 402 -Although it is well established that RAFTK is autophosphorylated upon stimulation, it is unclear to date whether the autophosphorylation occurs only at Tyr 402 or at other additional sites. Lakkakorpi et al. (14) reported that RAFTK-Y402F showed autophosphorylation activities based on an in vitro kinase assay with immunoprecipitated RAFTK-Y402F, suggesting the existence of additional autophosphorylation sites in addition to Tyr 402 . However, this result cannot exclude the possibility that a co-immunoprecipitated kinase, such as Src, might mediate RAFTK phosphorylation at other sites. Moreover, there is no report yet showing the autophosphorylation of RAFTK other than at the Tyr 402 site in vivo. Therefore, we investigated whether the trans-autophosphorylation of RAFTK occurs in vivo at tyrosine residues other than Tyr 402 using co-transfection of Y402F-FLAG with various RAFTK constructs. To this end, we utilized Src family kinase-deficient SYF cells, as compared with SYF ϩ Src cells, to block the possible Src-mediated tyrosine phosphorylation of RAFTK. As shown in Fig. 6C, upon co-transfection of the indicated constructs, none of the RAFTK constructs phosphorylated Y402F-FLAG in the Src family-deficient SYF cells, whereas RAFTK-Y402F was phosphorylated in the c-Src reintroduced SYF ϩ Src cells, indicating that RAFTK trans-autophosphorylation occurs only at residue Tyr 402 . Thus, RAFTK tyrosine residue 402 is a unique trans-autophosphorylation site for RAFTK activation. Although Lakkakorpi et al. (14) reported the phosphorylation of RAFTK-KM at Tyr 402 , Tyr 579/580 , and Tyr 881 in primary osteoclast-like cells, suggesting the possible involvement of another tyrosine kinase(s), given our data showing that activated Src phosphorylated RAFTK independent of the Tyr 402 site (Fig. 6B) and that endogenous RAFTK can phosphorylate RAFTK-KM in a trans-acting mode, the suggestion that another kinase(s) may phosphorylate RAFTK Tyr 402 is not convincing.
Role of Src in RAFTK-mediated Paxillin Phosphorylation-Because RAFTK can exhibit several different levels of phosphorylation status, it is of interest to examine which state is critical for the phosphorylation of substrates, such as paxillin. Paxillin is a focal adhesion molecule and is well known as an in vivo substrate of RAFTK. We investigated whether autophosphorylation of RAFTK would be sufficient to phosphorylate paxillin or whether Src-mediated RAFTK phosphorylation would be required. We examined RAFTK-mediated paxillin phosphorylation in SYF as compared with SYF ϩ Src cell lines using co-expression of various RAFTK constructs with paxillin. As shown in Fig. 7, RAFTK-mediated paxillin phosphorylation was significantly enhanced in the SYF ϩ Src cells. These data suggest that autophosphorylation of RAFTK is not required for paxillin phosphorylation and that Src plays an important role in RAFTK-mediated paxillin phosphorylation. KM-GFP showed some paxillin phosphorylation in the SYF ϩ Src cells, which might be mediated by Src activity (Fig. 7). Interestingly, in both cell lines, paxillin was significantly phosphorylated by del C-FLAG that lacks the RAFTK C-terminal region including the paxillin-binding focal adhesion targeting domain (Fig. 7), suggesting an alternative pathway for RAFTK to phosphorylate paxillin. Although RAFTK autophosphorylation is an initial step in RAFTK activation leading to the recruitment and activation of Src, activated Src, in turn, plays an important role in RAFTK-mediated paxillin phosphorylation as shown in Fig.  7. This result suggests an interdependent loop-like activation mechanism involving RAFTK and Src. This concerted activation mechanism of RAFTK and Src may provide a crucial step for tight regulation upon physiological stimuli.
Based on our study, the activation of RAFTK occurs in several steps: 1) upon its stimulation, RAFTK trans-phosphorylates the autophosphorylation site 402, which induces RAFTK activation; 2) the RAFTK phosphotyrosine 402 then recruits and activates Src kinase, which in turn phosphorylates the kinase domain of RAFTK and enhances its kinase activity; and 3) the enhanced RAFTK activity induces the activation of downstream signaling pathways.
Our results may provide important clues for the pharmacological application of RAFTK. Because RAFTK activity and downstream signaling heavily depend on RAFTK autophosphorylation of 402 through Src recruitment, blocking the Tyr 402 autophosphorylation of RAFTK can be a good therapeutic target to inhibit RAFTK function.