GIT1 Mediates Src-dependent Activation of Phospholipase Cγ by Angiotensin II and Epidermal Growth Factor*

Critical events for vasoconstrictor and growth factor signal transduction include stimulation of phospholipase Cγ (PLCγ) and elevation of intracellular calcium. c-Src has been proposed as a common mediator for these signals activated by both G protein-coupled receptors (GPCRs) and tyrosine kinase-coupled receptors (TKRs). Here we show that the GPCR kinase-interacting protein-1 (GIT1) is a substrate for c-Src that undergoes tyrosine phosphorylation in response to angiotensin II (AngII) and EGF in vascular smooth muscle and 293 cells. GIT1 associates with PLCγ via the PLCγ Src homology 2 and 3 domains constitutively, and the interaction is unaltered by AngII and EGF. GIT1 interaction with PLCγ is required for PLCγ activation based on inhibition of tyrosine phosphorylation and calcium mobilization after GIT1 knockdown with antisense GIT1 oligonucleotides. GIT1 interacts with PLCγ via a novel Spa homology domain (SHD) and a coiled-coil domain. Deletion mutation analysis showed that GIT1(SHD) is required for AngII- and EGF-mediated PLCγ activation (measured by phosphorylation of Tyr783 and inositol 1,4,5-trisphosphate formation). We propose that GIT1 is a novel regulator of PLCγ function that mediates PLCγ activation by c-Src and integrates signal transduction by GPCRs and TKRs.

AngII-mediated growth effects in VSMC also require the rapid activation of phospholipase C␥ (PLC␥) and mobilization of intracellular calcium (12)(13)(14). Early growth response genes induced by AngII such as c-fos and c-myc require calcium mobilization and activation of AP-1-dependent transcription. A key role for tyrosine kinases in AngII-mediated increases in intracellular calcium has been established (7). There is evidence that binding of AngII to the AT1R stimulates tyrosine phosphorylation of several proteins including Shc and GRB-2 prior to activation of members of the Ras family (15). Rapid tyrosine phosphorylation is probably mediated by nonreceptor tyrosine kinases including Src family tyrosine kinases, JAK2 and PYK2 (16 -22). Activation of PLC␥ requires tyrosine phosphorylation as shown by inhibition of Ins(1,4,5)P 3 formation with inhibitors such as herbimycin A and genistein (13). A specific role for Src was suggested by demonstrating rapid activation of Src in VSMC, inhibition of PLC␥ activation with antibodies to Src, and impaired signal transduction in SrcϪ/Ϫ cells (12,16,23). It is well known that PLC␥1 and PLC␥2 are substrates for members of the Src kinase family (24,25). Specific to AngII is the fact that AngII-mediated transactivation of the EGF receptor may be mediated by c-Src (22), suggesting important roles for both EGFR and c-Src in AngII-mediated events.
To understand the role of Src kinases in AngII-mediated signal transduction, we have focused on PLC␥ tyrosine phosphorylation as a key rapid event. We showed previously that AngII stimulated tyrosine phosphorylation of a ϳ97-kDa protein (p97) that associated with PLC␥ in VSMC (26). AngIIinduced tyrosine phosphorylation of both PLC␥ and the p97 were dependent on Src (26). To determine the role of p97 in signal transduction mediated by Src and PLC␥, we purified, identified, and sequenced p97 that co-precipitated with PLC␥. Here, we demonstrate that p97 is the same as GIT1, originally characterized as a G protein-coupled receptor kinase-interacting protein (27).
Two GIT family members with numerous tissue-specific alternatively spliced isoforms have been identified in mammals and chicken. These include GIT1 (also termed Cat-1 for Cool (cloned out of library)-associated tyrosine-phosphorylated pro-tein (28) and p95-APP1 (29)) and GIT2 (30,31) (the product of the KIAA0148 gene, also termed paxillin kinase linker protein (32), CAT-2, and p95-AAP2 (33)). All GIT family members share a structure composed of an amino-terminal zinc fingerlike motif, an ADP-ribosylation factor (ARF) GTPase-activating protein (GAP) domain, three ankyrin repeats, and a conserved carboxyl-terminal region that interacts with paxillin. GIT proteins are active as GAPs for ARF1 and ARF6 (30) and bind GRK2. GIT1 affects the function of G protein-coupled receptors (and the epidermal growth factor receptor) that are internalized through the clathrin-coated pit pathway in a ␤-arrestin-and dynamin-sensitive manner (34). All GIT family members appear to bind a complex that includes the guanine nucleotide exchange factor PIX and the p21 GTPase-activated kinase PAK (28,31,32). GIT1 interacts with paxillin more robustly than GIT2 (31). In the present study, we show that GIT1 links the AT1R and the EGFR to PLC␥ activation by associating with PLC␥. Our results demonstrate a novel role for GIT1 as an integrator of c-Src-dependent signal transduction activated by GPCRs and TKRs.

MATERIALS AND METHODS
Mass Spectrometric Analysis-The band corresponding to the 97-kDa protein was cut out of a silver-stained SDS-polyacrylamide gel and digested with trypsin. Four tryptic peptides were analyzed by mass spectrometric microsequence analysis (two showed the same sequence) as described previously (35). The sequences derived from these four peptides showed absolute identity with GIT1.
Cells and Transfection-VSMC were obtained from rat aorta as described (36). HEK293 cells and VSMC were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, penicillin, and streptomycin at 37°C in 5% CO 2 . HEK293T cells were transfected by LipofectAMINE/plus (Invitrogen). VSMC were transfected using the FuGENE 6 Reagent (Roche Applied Science). For cotransfection with AT1R, a ratio of 3:1 was used. After allowing protein expression for 24 h, cells were serum-deprived for 16 h and stimulated with 10 ng/ml EGF or 200 nM AngII. Phosphorothiolated sense (S), scrambled, or antisense (AS) oligonucleotides corresponding to the GIT1 sequence (S-GIT1, 5Ј-CAACTTCATCTGGGAGCACTC-3Ј; AS-GIT1, 5Ј-CTGATGAACTCTGACTTGATGG-3Ј) were prepared and transfected at 1 M for 48 h as previously described by our laboratory (37) and then serum-deprived for 16 h and stimulated with 10 ng/ml EGF or 200 nM AngII.
Immunoprecipitation and Immunoblotting-Anti-glutathione S-transferase (GST) monoclonal antibody was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-GIT1 monoclonal antibody was from BD Transduction Laboratories (Lexington, KY). Anti-FLAG M2 and anti-HA monoclonal antibodies were from Sigma. Anti-Xpress antibody monoclonal antibody was from Invitrogen. Anti-phosphotyrosine antibody (4G10) was from Upstate Biotechnology, Inc. (Lake Placid, NY), and anti-HA antibody was from Roche Applied Science. Anti-Tyr(P) 783 PLC was from Cell Signaling (Beverly, MA). For immunoprecipitations, cells were lysed in radioimmune precipitation buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholic acid, 0.1% SDS, 50 mM Tris-HCl, pH 8.0). Analysis of autoradiograms after immunoblotting was performed by scanning densitometry and processing with NIH Image. Statistical analysis was performed with Student's t test.
Infusion of Rats-Rats (n ϭ 3 for each group) were infused with 100 nM/kg/min AngII for 5 min, and aorta were isolated as described (38).
Ins (1,4,5)P 3 Formation and Intracellular Ca 2ϩ -VSMC and HEK293 cells were transfected as described and stimulated with AngII or EGF. Cells were harvested by adding ice-cold 10% (v/v) HClO 4 . Ins(1,4,5)P 3 was extracted with Freon and tri-n-octylamine. The samples were then measured according to the manufacturer's protocol (Amersham Biosciences). Intracellular Ca 2ϩ was measured as previously described (39). In brief, cells were grown on glass coverslips. After cotransfection with cDNAs as described, cells were loaded for 30 min at room temperature with the fluorescent Ca 2ϩ indicator 1 M fura 2-AM (Molecular Probes, Inc., Eugene, OR). The coverslip with dye-loaded cells was mounted in a tissue chamber (Bellco). Cells were then stimulated with AngII and EGF and sequentially exposed to 340-and 380-nm wavelength light using two excitation monochromators at a switching frequency of 100 Hz controlled by an optical chopper.
Protein-Protein Interaction Assays-HEK293T cells were co-transfected with Xpress-GIT1(wt) or pcDNA3.1 by LipofectAMINE/Plus. Cells were harvested, lysed in radioimmune precipitation buffer, and sonicated at 40 kHz for 6 s, and cell lysates were centrifuged (maximum speed, 10 min, 4°C). Approximately 500 g of precleared lysates were immunoprecipitated by 2 g of anti-HA and mouse IgG and were separated and probed with anti-Xpress and anti-HA antibodies and then with a horseradish peroxidase-conjugated secondary antibodies (Amersham Biosciences) and visualized by the ECL technique.

A c-Src-dependent Phosphoprotein That Interacts with PLC␥
Specifically Is GIT1-To define Src kinase substrates that mediate GPCR signal transduction, we characterized proteins rapidly tyrosine-phosphorylated in an Src-dependent manner. Previous data from our laboratory suggested an essential role for c-Src in AT1R signal transduction, especially mediated by PLC␥, which is tyrosine-phosphorylated within 1 min of AngII binding (16,23,26,40). Therefore, we immunoprecipitated proteins from VSMC that associated with PLC␥ and characterized proteins that were dependent on Src using both pharmacologic inhibition and SrcϪ/Ϫ cells. This analysis yielded a 97-kDa protein (p97) that was tyrosine-phosphorylated in response to AngII (26). p97 was excised from a silver stained SDS-polyacrylamide gel, digested with trypsin, and analyzed by mass spectrometry and microsequence analysis (35). Comparison with the SWISS-PROT protein sequence data base entries yielded a complete match with GIT1 previously identified by Premont et al. (27) (Fig. 1A). In growth-arrested VSMC, there was a low level of GIT1 tyrosine phosphorylation, which rapidly and transiently increased by 2.2 Ϯ 0.2-fold in response to AngII after 1 min (Fig. 1, B and C, upper panels, n ϭ 6, p Ͻ 0.05). In the absence of AngII, GIT1 was bound to PLC␥ as shown by co-precipitation in GIT1 immunoprecipitates (Fig.  1B, bottom panels), and the interaction did not change with AngII. PLC␥ also exhibited a low level of tyrosine phosphorylation under control conditions, and AngII increased PLC␥ tyrosine phosphorylation by 2.5 Ϯ 0.5-fold at 1 min (Fig. 1B, second panel; n ϭ 6, p Ͻ 0.05). Similar results were obtained when PLC␥ was immunoprecipitated (Fig. 1C). There was no change in the amount of GIT1 that co-precipitated with PLC␥ in response to AngII (Fig. 1C, third panel), indicating that GIT1 protein, already associated with PLC␥ under basal conditions, was being phosphorylated. It is well established that AngII activates PLC␦ and PLC␤ as well as PLC␥ (41). Neither PLC␦ nor PLC␤ bound to GIT1 in VSMC under basal conditions or after AngII stimulation (not shown). The relative expression levels of PLC␥, PLC␦, and PLC␤ in VSMC were similar, as assessed by Western blot (not shown).
GIT1 Plays a Key Role in Stimulation of PLC␥ Tyrosine Phosphorylation and Calcium Mobilization in VSMC-To show a role for GIT1 in PLC␥ function, we used antisense GIT1 oligonucleotides to decrease GIT1 expression in VSMC as pre-FIG. 1. Identification and cloning of GIT1 and binding of GIT1 to PLC␥. A, the derived tryptic peptides from the 97-kDa protein that immunoprecipitated with PLC␥ are shown below the sequence of mouse GIT1. The peptides showed absolute identity with GIT1. B and C, AngII-induced tyrosine phosphorylation of GIT1 and PLC␥. VSMC were stimulated with 200 nM AngII for 1 min. Cell lysates were immunoprecipitated with either GIT1 (B) or PLC␥ antibody (C) (IP). Immunoblotting (IB) was performed with anti-phosphotyrosine antibody (4G10; upper two panels). Binding of GIT1 to PLC␥ as well as equal loading was assured by stripping and reprobing the membrane with PLC␥ or GIT1 antibody (lower two panels). D, sense and antisense GIT1 oligonucleotides were transfected into VSMC as described. Cell lysates were prepared, and GIT1 protein content was measured by Western blot (IB: GIT1). Tyrosine phosphorylation of PLC␥ was measured by 4G10 antibody. The specificity of the sense and antisense oligonucleotides was demonstrated by lack of effect on PLC␥ expression (IB: PLC␥). There was no effect of 200 nM AngII (1 min) on the interaction between GIT1 and PLC␥, although it is clear that tyrosine phosphorylation of GIT1 increased. viously described in our laboratory (37). Sense and antisense GIT1 oligonucleotides were designed based on unique sequences and ability to inhibit mRNA expression. Transfection of antisense GIT1 oligonucleotides significantly decreased GIT1 protein expression at 24 h (Fig. 1D, top panel), whereas sense and scrambled (not shown) GIT1 oligonucleotides had minimal effect. Treatment with sense or antisense GIT1 oligonucleotides had no significant effect on PLC␥ expression (Fig.  1D, lower panel). Treatment with antisense GIT1 oligonucleotides significantly decreased PLC␥ tyrosine phosphorylation induced by AngII (Fig. 1D, middle, 63% inhibition, n ϭ 3), whereas sense and scrambled (not shown) GIT1 oligonucleotides had no significant effect. The decrease in basal PLC␥ tyrosine phosphorylation observed with antisense GIT1 oligonucleotides was variable and not significant.
To assess the effect of GIT1 depletion on PLC␥ function, we measured calcium mobilization in VSMC using Fura-2. Cells were stimulated for 1 min with AngII, which rapidly increased intracellular calcium as previously reported (42). Peak intracellular calcium concentration in response to AngII was 223 Ϯ 40 nM, whereas in cells transfected with antisense GIT1 oligonucleotides, peak calcium was 130 Ϯ 73 nM. The percentage increase in calcium concentration stimulated by AngII was significantly different in antisense GIT1 oligonucleotide-transfected cells compared with sense and scrambled GIT1 oligonucleotide-transfected cells (83% inhibition, n ϭ 5). Thus, GIT1 is required for AngII-stimulated calcium mobilization.
GIT1 Plays a Key Role in Stimulation of PLC␥ Tyrosine Phosphorylation in VSMC-PLC␥ can be activated by several stimuli in VSMC including AngII, thrombin, PDGF, and EGF (43,44). We initially identified GIT1 as a 97-kDa protein bound to PLC␥ by analysis of VSMC after AngII stimulation. However, a more general role for GIT1 in G protein-coupled receptors and tyrosine kinase-coupled receptors has been proposed (34,45). As shown in Fig. 2A, AngII, thrombin, EGF, and PDGF all stimulated tyrosine phosphorylation of GIT1 in VSMC (from 1.5-to 2.5-fold). We do not believe that AngII-mediated GIT1 phosphorylation was due to transactivation of the PDGF and EGF receptors (46,47), because AngII stimulated phosphoryl-ation of GIT1 more rapidly and to a greater extent than PDGF and EGF. In addition, inhibiting PDGF and EGF receptors with AG1478 and AG1295 did not alter AngII-stimulated GIT1 phosphorylation, despite inhibiting transactivation (not shown).
To demonstrate that GIT1 is important for AngII function in vivo, we infused rats for 5 min with AngII and analyzed tyrosine phosphorylation of GIT1 (Fig. 2B). In rat aortas, we found that GIT1 bound to PLC␥ in both control and agonist-treated aortas. In response to AngII infusion, tyrosine phosphorylation of GIT1 was increased by ϳ2-fold in the aorta.
Structure-Function Analysis of GIT1 in HEK 293 Cells-To determine the mechanism by which GIT1 regulates receptormediated activation of PLC␥, we established a model system using 293 cells, since VSMC are difficult to transfect efficiently with plasmid cDNAs. When transfected with the AngII type 1 receptor (AT1R) cDNA, up to 90% of 293 cells express the AT1R (42) (data not shown). Importantly, in AT1R-expressing 293 cells, AngII (like EGF) stimulates tyrosine phosphorylation of many proteins. To study GIT1 regulation in 293 cells, we first cloned and sequenced the homologous mouse GIT1 cDNA (mGIT1) (Fig. 1A). Comparison of the human, rat, and mouse GIT1 sequences showed that mGIT1 had 89% homology to human GIT1 and 85% homology to rat GIT1. We transfected wild type Xpress-tagged GIT1 cDNA (Xpress-GIT1(wt)) into 293 cells and confirmed that PLC␥ bound to GIT1 under basal conditions (Fig. 3A). To verify that both PLC␥ and GIT1 were tyrosine-phosphorylated by AngII in 293 cells, we cotransfected Xpress GIT1(wt) and HA-tagged AT1R at a ratio of 3:1. After allowing protein expression for 24 h, the cells were serumdeprived for 16 h and then stimulated with 200 nM AngII for 1 min. In response to AngII, GIT1 and PLC␥ were tyrosinephosphorylated in 293 cells to a similar extent as in VSMC (Fig. 3B). EGF-mediated GIT1 and PLC␥ phosphorylation were used as positive controls, since 293 cells express high levels of endogenous EGFR. EGF (10 ng/ml for 5 min) stimulated tyrosine phosphorylation of PLC␥ and Xpress-GIT1(wt) in 293 cells to a similar extent as AngII (Fig. 3, C and D, left two lanes). To prove that GIT1 and PLC␥ phosphorylation were c-Src-dependent in 293 cells, we repeated the EGF experiment in the presence of a Src kinase inhibitor, 10 M 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo [3,4-d]pyrimidine (PP2). Activation of PLC␥ was assessed by measuring phosphorylation of Tyr 783 , which has been shown to correlate with PLC␥ activity (48). Treatment with protein phosphatase 2 significantly decreased PLC␥ (Fig. 3C) and GIT1 (Fig. 3D) tyrosine phosphorylation (92 Ϯ 6 and 95 Ϯ 5% inhibition, respectively, n ϭ 3), demonstrating an essential role for Src family kinases in 293 cells. There was no significant effect of 10 M protein phosphatase 2 on EGF-mediated EGFR tyrosine phosphorylation (Fig. 3E). These results are very similar to data previously published by our laboratory and others for protein phosphatase 1 in VSMC (23,26) and for protein phosphatase 2 in ECV304 cells (50), cardiac myocytes (51), and endothelial cells (52). We were unable to perform antisense GIT1 knockdown experiments in 293 cells because the endogenous level of GIT1 expression was too low to assay by Western blot (not shown).
Identification of the PLC␥ Binding Domains Required for Interaction with GIT1-We next investigated the domains of PLC␥ required for GIT1 interaction. Full-length PLC␥ contains two SH2 domains (amino-terminal (SH2n) and carboxyl-terminal (SH2c)) and a single SH3 domain (25). Using 293 cells transfected with Xpress-GIT1(wt) and a FLAG-tagged construct of PLC␥ that contained only the Src homology domains (PLC␥-SH2n-SH2c-SH3), we performed co-precipitation (Fig.  4A). Immunoprecipitation with anti-FLAG antibody precipi- nM AngII, and cell lysates were immunoprecipitated (IP) with PLC␥ antibody. Immunoblotting (IB) was performed using the indicated antibodies. Tyrosine phosphorylation of both GIT1 and PLC␥ by AngII was similar to results in VSMC (Fig. 2). C, 293 cells were transfected with PLC␥. Cells were stimulated with 10 ng/ml EGF after treatment for 30 min with 10 M protein phosphatase 2 or vehicle (Me 2 SO; DMSO). Cell lysates were immunoprecipitated with PLC␥ antibody. Immunoblotting was performed using the indicated antibodies. D, 293 cells were transfected with Xpress-GIT1(wt). Using the same experimental protocol as in C, cell lysates were immunoprecipitated with GIT1 antibody. Immunoblotting was performed using the indicated antibodies. E, cells were stimulated with 10 ng/ml EGF for the indicated times after treatment for 30 min with 10 M protein phosphatase 2 or vehicle (Me 2 SO). Cell lysates were immunoprecipitated with EGFR antibody. Immunoblotting was performed using the indicated antibodies (note that pY-EGFR (845) is specific for autophosphorylation). tated ϳ15% of Xpress-GIT1(wt), suggesting a significant interaction. No GIT1 precipitated with vector alone (P-Tag2C; Fig.  4A). Next, we performed a GST pull-down assay with commercially available GST fusion proteins containing different combinations of the SH2 and SH3 domains of PLC␥. We were readily able to co-precipitate GIT1 with GST-PLC␥ that contained all three domains (SH2n-SH2c-SH3 in Fig. 4B), indicating that there is probably a direct interaction between GIT1 and PLC␥. Further binding analysis showed that all three domains were required for maximal GIT1 binding (Fig. 4B). An essential role for the first SH2 domain of PLC␥ was apparent, since there was no binding of SH2c-SH3, whereas SH2n-SH3 and SH2n retained ϳ40% of the binding observed with SH2n-SH2c-SH3 (Fig. 4B).
Identification of the GIT1 Binding Domains Required for Interaction with PLC␥-GIT1 is composed of a zinc fingercontaining ARF GAP domain (GAP), ankyrin repeat region (ANK), two carboxyl paxillin-binding subdomains, a Spa2 homology domain (SHD), and three putative coiled coil (CC) domains (Fig. 5A). The ARF-GAP domain has been shown to be functionally important for endocytosis (45). Interestingly, only GIT1 (amino acids 250 -396) and yeast Spa2 family members contain SHD motifs (1). The CC regions include CC1 (residues 253-274), which overlaps with the SHD domain, CC2 (residues 424 -474), and CC3 (residues 649 -669) (53). The CC2 region is important for interactions with PAK (32), the CC3 (including 646 -770) is essential for interactions with paxillin (53), and the SHD region overlaps with binding sites defined for focal adhesion kinase and PIX (53). To identify the regions in GIT1 that interact with PLC␥, we made deletion mutations of these putative protein interacting domains as described (Fig. 5A).
Because the SHD domain and CC1 overlap, we first investigated the requirement for this domain in GIT1-PLC␥ binding. As shown in Fig. 5B, FLAG-GIT1(del-SHD), which lacks aa 250 -420, was unable to bind PLC␥ in transfected 293 cells. In contrast, GIT1 (aa 1-420) and GIT1 (aa 250 -770) were able to bind PLC␥ (Table I). These results suggested that critical binding motifs were present in aa 250 -420. To provide further evidence for binding via the SHD domain, we compared the ability of GST-GIT1 (aa 1-250) and GST-GIT1 (aa 250 -420) to precipitate PLC␥ in a pull-down assay. As expected, PLC␥ was pulled down by GST-GIT1 (aa 250 -420) but not by GST-GIT1 (aa 1-250) (Fig. 5C). To investigate the role of the CC2 domain we transfected cells with GIT1(del-CC2) and GIT1(wt). In com-  parison with GIT1(wt), precipitation of GIT1(del-CC2) by anti-PLC␥ antibody was significantly decreased (percentage decrease in GIT1(del-CC2) binding was 69 Ϯ 10%, n ϭ 3; Fig. 5D, Table I). These results indicate that the SHD and CC2 domains play important roles in GIT1 interactions with PLC␥.
To evaluate whether the requirement for GIT1 binding to PLC␥ via the SHD domain held true for AngII, we repeated the experiments in 293 cells cotransfected with the AT1R, PLC␥, and GIT1 constructs. As shown in Fig. 7A, AngII stimulated Tyr(P) 783 maximally at 5 min (2.5-fold increase) in cells transfected with vector control. Transfection of GIT1(wt) did not alter the time course but significantly increased Tyr(P) 783 compared with control (5.6-fold increase). In contrast, GIT1(del-SHD) was not significantly different from control at 0 min (2.1-fold increase) (Fig. 7B). However, at 5 min, pY783 was markedly reduced in cells transfected with GIT1(del-SHD), compared to cells transfected with GIT1 (wt).
Interaction of GIT1 with the AT1R and c-Src-An unanswered question is how AngII binding to the AT1R leads to tyrosine phosphorylation of PLC␥ and GIT1. To determine whether there were direct interactions of the AT1R and c-Src with either GIT1 or PLC␥, we performed co-immunoprecipitation experiments. Neither PLC␥ nor GIT1 directly associated with the AT1R based on this analysis, since immunoprecipitation of the AT1R did not co-precipitate PLC␥ or GIT1, nor was the AT1R present in PLC␥ and GIT1 immunoprecipitates from VSMC (data not shown). Similarly, c-Src was not present in these immunoprecipitates. To determine whether GIT1 interacted with these proteins when overexpressed, we cotransfected GIT1(wt) and the AT1R into 293 cells. GIT1(wt) did not directly associate with the AT1R or c-Src in 293 cells (data not shown). Thus, the interactions between these proteins are transient and/or other proteins are involved. DISCUSSION The major findings of this study are the identification of GIT1 as a novel mediator for activation of PLC␥ that appears to integrate signal transduction from both GPCRs and TKRs. Binding of GIT1 to PLC␥ requires the Spa2 homology domain (aa 250 -470), a domain whose function has only been defined in yeast. In addition, GIT1 appears to be a substrate for c-Src phosphorylation in response to activation of both GPCRs and TKRs and may explain, in part, the ability of AngII to activate PLC␥. Several independent experimental approaches support an essential role for GIT1 in PLC␥ function. Knockdown of GIT1 expression significantly decreased PLC␥ activation. Transfection of a GIT1 mutant that does not bind PLC␥, termed GIT1(del-SHD), prevented PLC␥ activation as shown by decreased PLC␥ Y783 phosphorylation and Ins(1,4,5)P3 formation. Treatment of cells with the c-Src inhibitor protein phosphatase 2 blocked AngII-and EGF-mediated phosphorylation of both GIT1 and PLC␥, indicating the importance of c-Src as a common mediator. These data support previous studies documenting an important role for Src-dependent tyrosine phosphorylation in PLC␥ activation and calcium mobilization in several cell types (50,54,55). In addition, results from our laboratory in SrcϪ/Ϫ cells and with c-Src inhibitors showed a correlation between decreased GIT1 tyrosine phosphorylation and activation of PLC␥ (23,26).
The central portion of GIT1 (aa 250 -470) encompassing the SHD, CC1, and CC2 domains is required for regulation of PLC␥. PLC␥ activation requires PLC␥ interaction with the SHD and CC1 (and to a lesser extent CC2) domains of GIT1 and binding of GIT1 to the SH2-SH3 domains of PLC␥. Future studies will be required to define the specific mechanisms that mediate activation of PLC␥ when bound to GIT1, since binding is constitutive. It is likely that c-Src-mediated phosphorylation of specific tyrosine residues in regions of GIT1 will be important based on our finding that tyrosine phosphorylation was required for agonist-mediated Ins(1,4,5)P 3 formation. Based on the present study, we propose a model to explain the role of GIT1 in activation of PLC␥ by AngII and EGF (Fig. 8). In serum-starved cells, GIT1 exists in a preassembled complex with PLC␥. Upon agonist binding (for both GPCRs and TKRs), c-Src is activated and phosphorylates critical tyrosine residues in GIT1. We propose that a conformational change in GIT1 now recruits additional signaling molecules (e.g. focal adhesion kinase) and/or enables c-Src to phosphorylate additional tyrosine residues (e.g. PLC␥ Tyr 783 ) that lead to activation of PLC␥. This pathway is separate from the transactivation pathway for GPCR-mediated tyrosine phosphorylation of the EGFR by generation of heparin-binding EGF.
The newly defined role of GIT1 in PLC␥ activation and Ins(1,4,5)P 3 formation described here complements the known roles of GIT family members. Previous studies indicate important roles for GIT proteins in receptor endocytosis (27,34,45) and cell motility (53,56). GIT1 (also termed Cat-1 for Cool (cloned out of library)-associated tyrosine-phosphorylated protein) was originally identified in a yeast two-hybrid screening for proteins that bind G protein-coupled receptor kinase 2 (27). Initial results demonstrated that GIT1 functions as a GAP for the ARF family of proteins, and GIT1 was shown to regulate both GPCR and TKR receptor internalization. GIT proteins are expressed ubiquitously in human tissues (27,31) so that GIT family members are likely to regulate many membrane receptors. Work from Zhao et al. (53) and from Turner and colleagues (32,56) supports an important role for GIT family members in focal adhesions and cell motility. Specifically, these authors FIG. 8. Model of PLC␥ regulation by GIT1. AngII binding to the AT1R (or EGF binding to the EGFR) activates c-Src. c-Src phosphorylates GIT1, which is bound to PLC␥ in a complex under basal conditions. Conformational changes in GIT1 induced by tyrosine phosphorylation lead to phosphorylation and activation of PLC␥. Activation of PLC␥ leads to increases in inositol 1,4,5-trisphosphate formation and calcium mobilization.
FIG. 7. GIT1 SHD domain is required for AngII-stimulated PLC␥ tyrosine phosphorylation. Cells were co-transfected as in Fig. 6, except that HA-tagged AT1R was also transfected in all cases. Results are representative of three experiments. To quantify results, the gels were scanned, and the densitometry of pY-PLC␥ was arbitrarily set to 1.0. Results shown are representative of three experiments.
propose that upon cell activation, Cdc42 recruitment of PAK and PIX drives the association of GIT1 with focal adhesions. This favors dissociation of paxillin from focal adhesions that become destabilized and promotes motility by decreasing cell adherence. The association of GIT proteins with PIX, PAK, and paxillin suggests a functional role for GIT proteins in regulation of the cytoskeleton (29,31,53). Because of the important role of calcium in cell motility and focal adhesion function (e.g. Pyk2 is a calcium-dependent focal adhesion kinase), the present study suggests that GIT1 may be involved in local regulation of calcium homeostasis. Future work will be required to define the possible roles of known GIT1-interacting proteins in regulation of PLC␥, especially at specific intracellular sites such as focal adhesions.
To date only one PLC scaffold has been described, the Drosophila photoreceptor PDZ scaffold protein INAD, which interacts with PLC-␤ (57). In this system, the association between PLC and INAD is important to terminate the light response via PLC functioning as a GAP. PLC␥ has been shown to bind several kinases including Tnk1, c-Abl, focal adhesion kinase, and TKRs (both EGFR and PDGF receptor) (55,58). However, these interactions result in direct tyrosine phosphorylation of PLC␥. In contrast, the mechanism by which c-Src phosphorylates and activates PLC␥ remains unknown. To our knowledge, GIT1 is the first mammalian protein proposed as a PLC␥ scaffold. Scaffolds have been shown to exhibit at least four characteristics as detailed by Chang and Karin (59). They 1) assemble specific modules; 2) exclude (or insulate) other molecules; 3) promote sequential activation of enzymes by physical interaction; and 4) provide feedback regulation of cell surface receptors. It seems logical that another characteristic of scaffolds is to direct subcellular localization of signaling events (60). For example, several MEK1-ERK1/2 scaffolds have been identified with specific subcellular roles: KSR1 to plasma membrane (61), SpaII to actin cytoskeleton (62), and MKP1 to nucleus (49). Because of the known interactions of GIT1 with proteins in both focal adhesions and actin cytoskeleton, we suggest that an important role for GIT1 may be to regulate PLC␥ activity in these locations.