Src-induced phosphorylation of caveolin-2 on tyrosine 19. Phospho-caveolin-2 (Tyr(P)19) is localized near focal adhesions, remains associated with lipid rafts/caveolae, but no longer forms a high molecular mass hetero-oligomer with caveolin-1.

Caveolin-2 is the least well studied member of the caveolin gene family. It is believed that caveolin-2 is an "accessory protein" that functions in conjunction with caveolin-1. At the level of the ER, caveolin-2 interacts with caveolin-1 to form a high molecular mass hetero-oligomeric complex that is targeted to lipid rafts and drives the formation of caveolae. However, caveolin-2 is not required for caveolae formation, implying that it may fulfill some unknown regulatory role. Here, we present the first evidence that caveolin-2 is a phosphoprotein. We show that caveolin-2 undergoes Src-induced phosphorylation on tyrosine 19. To study this phosphorylation event in vivo, we generated a novel phospho-specific antibody probe that only recognizes phosphocaveolin-2 (Tyr(P)(19)). We then used NIH-3T3 cells stably overexpressing c-Src to examine the localization and biochemical properties of phosphocaveolin-2 (Tyr(P)(19)). Our results indicate that phosphocaveolin-2 (Tyr(P)(19)) is localized near focal adhesions, remains associated with lipid rafts/caveolae, but no longer forms a high molecular mass hetero-oligomer with caveolin-1. Instead, phosphocaveolin-2 (Tyr(P)(19)) behaves as a monomer/dimer in velocity gradients. Thus, we conclude that the tyrosine phosphorylation of caveolin-2 (Tyr(P)(19)) may function as a signal that is recognized by the cellular machinery to induce the dissociation of caveolin-2 from caveolin-1 oligomers. We also demonstrate that (i) insulin-stimulation of adipocytes and (ii) integrin ligation of endothelial cells can both induce the tyrosine phosphorylation of caveolin-2 (Tyr(P)(19)). During integrin ligation, phosphocaveolin-2 (Tyr(P)(19)) co-localizes with activated FAK at focal adhesions. Thus, phosphocaveolin-2 (Tyr(P)(19)) may function as a docking site for Src homology domain-2 (SH2) domain containing proteins during signal transduction. In support of this notion, we identify several SH2 domain containing proteins, namely c-Src, NCK, and Ras-GAP, that interact with caveolin-2 in a phosphorylation-dependent manner. Furthermore, our co-immunoprecipitation experiments show that caveolin-2 and Ras-GAP are constitutively associated in c-Src expressing NIH-3T3 cells, but not in untransfected NIH-3T3 cells.

phosphorylation-dependent manner. Furthermore, our co-immunoprecipitation experiments show that caveolin-2 and Ras-GAP are constitutively associated in c-Src expressing NIH-3T3 cells, but not in untransfected NIH-3T3 cells.
Receptor tyrosine kinases (RTKs) 1 contain one or more autophosphorylation sites. Upon ligand binding, a specific receptor is activated and undergoes autophosphorylation on specific tyrosine residues. One known function for tyrosine phosphorylation is to confer docking sites for Src homology domain-2 (SH2)-containing proteins. Numerous downstream signaling molecules possessing either intrinsic kinase activity or an adapter function are then recruited to the plasma membrane and utilize their SH2 domains to form hetero-oligomeric signaling complexes with activated RTKs.
Numerous RTKs, such as epidermal growth factor receptor (EGF-R), platelet-derived growth factor receptor (PDGF-R), and insulin receptor (Ins-R), as well as non-receptor tyrosine kinases (NRTKs; Src-family tyrosine kinases) are now known to be concentrated within specialized plasma membrane-associated microdomains, known as lipid rafts and caveolae. It is believed that lipid rafts/caveolae serve as integrated platforms to recruit and concentrate signaling molecules, allowing them to form preassembled complexes. Caveolins, a family of 21-24-kDa integral membrane proteins, are known to play an active role in this process by directly interacting with and modulating the activity of many of these signaling molecules.
The caveolin gene family consists of caveolin-1, -2, and -3 (1-3). Caveolin-1 and -2 are co-expressed in most cell types and tightly interact with each other to form a high molecular mass hetero-oligomeric complex. In contrast, the expression of caveolin-3 is muscle-specific, and it forms a high molecular mass homo-oligomeric complex (4 -6).
Caveolin-2 is 38% identical and 56% similar to caveolin-1, being the most divergent member of the caveolin gene family (7). Unlike caveolin-1 and -3, caveolin-2 does not have the capacity to form caveolae by itself. In Cav-1(Ϫ/Ϫ) cells, caveolin-2 is retained intracellularly at the level of the Golgi complex where it undergoes proteasomal degradation (8). In light of these observations, caveolin-2 is considered to be an accessory protein for caveolin-1.
Historically, caveolin-1 was first identified as a tyrosinephosphorylated protein in v-Src-transformed chicken embryo fibroblasts (9). By using microsequence analysis and site-directed mutagenesis, we showed that tyrosine 14 is the principal site for Src-induced tyrosine phosphorylation of caveolin-1 (10). Tyrosine 14 and its flanking region closely resembles the known recognition motifs for tyrosine kinases (KYVD-SEGHLpY: (R/K/Q)X 2-4 (D/E)X 2-3 pY); therefore, we speculated that phosphocaveolin-1 (Tyr(P) 14 ) might have a functional role in signal transduction. To test this hypothesis, we generated and characterized a novel phospho-specific mAb probe that selectively recognizes phosphocaveolin-1 (Tyr(P) 14 ). Using this phospho-specific mAb, we and others have now shown that the tyrosine phosphorylation of caveolin-1 also occurs in normal cells, but in a tightly regulated fashion. For example, EGF, insulin, and integrin ligation, as well as osmotic shock, all have been shown to induce the phosphorylation of caveolin-1 on tyrosine 14 (11)(12)(13). Moreover, phosphocaveolin-1 (Tyr(P) 14 ) directly binds to the SH2 domain-containing protein Grb7, thereby augmenting both anchorage-independent growth and EGF-stimulated cell migration (12).
The physical association of caveolin-2 with caveolin-1 raises the possibility that caveolin-2 may also be a cellular target for tyrosine phosphorylation by Src family kinases. Several indirect lines of evidence support this hypothesis as: (i) caveolin-2 forms a high molecular mass hetero-oligomer through a direct interaction with caveolin-1, and (ii) like caveolin-1, the caveolin-2 protein sequence contains a highly conserved motif for recognition by tyrosine kinases at its extreme N terminus (centered around tyrosine 19; QLFMADDApY). Therefore, we speculated that caveolin-2 might be phosphorylated by Src family kinases, thus contributing to the recruitment of signaling molecules to the cytoplasmic face of lipid rafts/caveolae microdomains.
Anti-sera Development and IgG Purification-Briefly, a synthetic phospho-peptide was conjugated to maleimide-activated keyhole limpet hemocyanin through a free C-terminal cysteine residue. New Zealand White rabbits were inoculated subcutaneously with 200 g of immunogen and adjuvant on weeks 0, 4, 8, and 10. Anti-sera were collected on weeks 6, 10, and 12 and pooled for total IgG isolation. Specific IgGs were differentially purified by affinity chromatography against their respective phospho-and non-phosphopeptides coupled to Affi-Gel 15 (Bio-Rad).
IP/Western of Caveolin-2 (WT) and Caveolin-2 (Y19A)-COS-7 cells were transiently co-transfected with caveolin-2 (WT) or caveolin-2 (Y19A), alone or in combination with c-Src (WT). Thirty-six hours post-transfection, the cells were processed for immunoprecipitation using protein A-Sepharose CL-4B (Amersham Biosciences). Briefly, cells were lysed in IP buffer containing 10 mM Tris, pH 8.0, 150 mM NaCl, 1% Triton X-100, 60 mM octyl glucoside with phosphatase inhibitor (50 mM NaF, 30 mM sodium pyrophosphate, 100 M sodium orthovanadate) and protease inhibitors (Roche Molecular Biochemicals). Lysates were precleared by addition of 50 l of 1:1 slurry of protein A-Sepharose in TNET buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100) containing 1 mg/ml bovine serum albumin. After 30 min at 4°C, samples were centrifuged for 5 s at 15,000 ϫ g. The resulting supernatants were transferred to fresh tubes, and 50 l of protein A-Sepharose was added together with anti-Myc IgG (A-14; Santa Cruz Biotechnology, Inc.; a rabbit pAb). Samples were then incubated for an additional 3 h at 4°C. Immunoprecipitates were washed five times with IP buffer, and samples were separated by 12.5% SDS-PAGE and transferred to nitrocellulose. Blots were then probed with a well characterized mAb directed against phosphotyrosine (PY20; BD Transduction Laboratories).
Treatment of Cells with Ligands-For insulin stimulation, 3T3-L1 adipocytes (day 8) were incubated in serum-free medium overnight. Insulin (150 nM) was then added, and the cells were incubated at 37°C for 30 min. Cells were collected and processed for SDS-PAGE/Western blot. For the integrin ligation experiments, HMECs were grown to ϳ70% confluence and then incubated in serum-free medium overnight. Cells were then trypsinized, collected in serum-free media, and reseeded onto fibronectin (100 g/ml) or poly-L-lysine (10 g/ml) precoated coverslips. After seeding, the cells were incubated at 37°C for 20 min, fixed in 2% paraformaldehyde, and processed for immunofluorescence microscopy.
Immunofluorescence-All reactions were performed at room temperature. NIH 3T3 cells (untransfected or stably expressing c-Src) were briefly washed three times with PBS and fixed for 45 min in PBS containing 2% paraformaldehyde. Fixed cells were rinsed with PBS and treated with 50 mM NH 4 Cl in PBS for 10 min to quench free aldehyde groups. Cells were then permeabilized with 0.1% Triton X-100 for 10 min at room temperature and washed twice with PBS, 10 min each. Cells were then incubated with PBS, 0.2% bovine serum albumin containing: (i) a 1:200 dilution of anti-phosphocaveolin-2 IgG (pAb) and anti-caveolin-2 IgG (mAb 65) and (ii) lissamine rhodamine B sulfonylchloride-conjugated goat anti-mouse antibody (5 g/ml) and fluorescein isothiocyanate-conjugated donkey anti-rabbit antibody (5 g/ml). The primary antibody incubation was for 60 min, and the secondary antibody incubation was for 30 min. Cells were washed three times with PBS between incubations. Slides were mounted with Slow-Fade antifade reagent and observed under CCD camera. Other double labeling experiments were carried out in a similar fashion.
CCD Imaging and Deconvolution-Using an Olympus IX80 microscope with a 60ϫ Plan Neofluar objective and a Photometrics cooled CCD camera with a 35-mm shutter, the images were acquired and processed using IP Lab on a Power Mac 8500. For each sample, three to five two-dimensional images were acquired, deconvolved, and then combined into one image.
Immunoblotting with Anti-Phosphocaveolin-2 IgG-Cells were lysed in boiling sample buffer (21). Samples were then collected and boiled for a total of 5 min. Samples were sonicated briefly. Cellular proteins were resolved by SDS-PAGE (12.5% acrylamide) and transferred to nitrocellulose membranes (0.2 m). Blots were incubated for 2 h in TBST (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.2% Tween 20) containing 5% bovine serum albumin. After three washes with TBST, membranes were incubated for 2 h with the primary antibody (ϳ500-fold diluted in TBST with 5% bovine serum albumin) and for 1 h with horseradish peroxidaseconjugated goat anti-rabbit IgG (ϳ5,000-fold diluted). Proteins were detected using an ECL detection kit (Amersham Biosciences). For the peptide competition assay, peptides stocks were prepared in Me 2 SO. Each peptide (ϳ100 g/ml; a 100-fold molar excess of peptide) was mixed with the primary antibody solution prior to Western blotting. See Fig. 3A for a detailed description of the competing peptides that were used.
Preparation of Caveolae-enriched Membrane Fractions-NIH 3T3 cells (untransfected or stably expressing c-Src) were washed with PBS and lysed with 2 ml of ice-cold MES-buffered saline (MBS, 25 mM MES, pH 6.5, 0.15 M NaCl) containing 1% (v/v) Triton X-100 (11,14,15,(22)(23)(24)(25)(26)(27)(28)(29). Homogenization was carried out with 10 strokes of a loosefitting Dounce homogenizer. The homogenate was adjusted to 40% sucrose by the addition of 2 ml of 80% sucrose prepared in MBS and placed at the bottom of an ultracentrifuge tube. A 5%/30% discontinuous sucrose gradient was formed above the homogenate and centrifuged at 39,000 rpm for 16 h in a SW 41 rotor (Beckman Instruments). A light scattering band confined to the 5-30% interface was observed that contained caveolin-1, but excluded most other cellular proteins. From the top of each gradient, 1-ml gradient fractions were collected to yield a total of 12 fractions. An equal volume from each gradient fraction was separated by SDS-PAGE and subjected to immunoblot analysis.
Velocity Gradient Centrifugation-NIH 3T3 cells (untransfected or stably expressing c-Src) were dissociated in MES-buffered saline containing 60 mM octyl glucoside. Solubilized material was loaded atop a 5-40% linear sucrose gradient and centrifuged at 50,000 rpm for 10 h in a SW 60 rotor (4 -7). Gradient fractions were collected from above and subjected to immunoblot analysis. Molecular mass standards for velocity gradient centrifugation were as we described previously (4 -7).
In Vitro Phosphorylation-Caveolin-rich domains were purified from murine lung tissue, as described (15,24). Caveolin-rich membrane domains (ϳ5 g) were then resuspended in 20 l of kinase buffer (20 mM Hepes, pH 7.4, 1 mM MgCl 2 , 1 mM MnCl 2 ), and the reaction was initiated by addition of 1 mM ATP. After 10 min at room temperature, the reaction was halted by addition of 20 l of 2ϫ SDS sample buffer and boiling for 2 min. For the peptides competition assay, 2.5 l of peptides stocks or the vehicle alone control (Me 2 SO) were added and preincubated on ice for 30 min. Peptide stocks were prepared in Me 2 SO at a concentration of 5 mM. Each peptide was then diluted to the 10ϫ concentration (500 and 100 M) in kinase buffer. See Table I for a detailed description of the peptides that were used.

Detection of SH2 Domain Proteins That Bind
Tyrosine-phosphorylated Caveolin-2-Purified non-phosphorylated and tyrosine phosphorylated GST-Cav-2-(1-86) fusion proteins were immobilized on glutathione-agarose beads and incubated with lysates from normal NIH 3T3 cells. These lysates were generated with IP buffer containing a battery of phosphatase and protease inhibitors (10 mM Tris, pH 8.0, 150 mM NaCl, 1% Triton X-100, 60 mM octyl glucoside, 50 mM NaF, 30 mM sodium pyrophosphate, 100 M sodium orthovanadate, pepstatin A (1 g/ml), and 1 tablet of complete protease inhibitor mixture (Roche Molecular Biochemicals). After incubation rotating overnight at 4°C, the beads were washed with lysis buffer (five times), separated by 12.5% SDS-PAGE, and transferred to nitrocellulose membrane. Bound proteins were visualized by immunoblotting with a panel of antibodies directed against known SH2 domain containing proteins (Table II). These antibodies were purchased from BD Transduction Laboratories, Inc.
In Vivo Ras-GAP/Caveolin-2 Co-immunoprecipitation Experiments-NIH 3T3 cells (untransfected or stably expressing c-Src) were lysed in IP buffer containing phosphatase and protease inhibitors (detailed above) and subjected to immunoprecipitation with protein A-Sepharose CL-4B (Amersham Biosciences). Briefly, lysates were first precleared by addition of a 50 l of 1:1 slurry of protein A-Sepharose in TNET buffer (defined above) containing 1 mg/ml bovine serum albumin. After 30 min of preclearing at 4°C, samples were centrifuged for 5 s at 15,000 ϫ g, and the supernatants were transferred to fresh tubes. Then, 50 l of protein A-Sepharose was added together with an irrelevant IgG (2 g/ml) or Ras-GAP pAb IgG (2 g/ml; Santa Cruz Biotechnology, Inc.). After incubation for 3 h at 4°C, the immunoprecipitates were washed three times with IP buffer, and the samples were separated by 12.5% SDS-PAGE and transferred to nitrocellulose membranes. Blots were probed with a mAb to detect total caveolin-2 or pAb to detect tyrosine 19-phosphorylated caveolin-2.

Caveolin-2 Is Phosphorylated on Tyrosine 19 by c-Src-Orig-
inally, caveolin-1 was first identified as a major phosphoprotein in v-Src-transformed fibroblasts (9). This phosphorylation event is dependent on both the correct myristoylation and the enzymatic activity of c-Src, as myristoylation-deficient or kinase-dead mutants of c-Src are unable to phosphorylate caveolin-1 (12). Previously, our group mapped the phosphorylation site on caveolin-1 to tyrosine 14 (10). Using a mAb that selectively recognizes tyrosine 14 phosphorylated caveolin-1, we showed caveolin-1 undergoes phosphorylation by c-Src, and in response to a variety of stimuli, including treatment with insulin (12), EGF (12), and osmotic shock (13).
To test our hypothesis, we first examined whether other caveolins can be tyrosine-phosphorylated. Myc-tagged caveolin-2 and caveolin-3 were transiently transfected into COS-7, alone or in combination with c-Src. Cell lysates were then immunoprecipitated and subjected to immunoblotting with anti-phosphotyrosine IgG. As shown in Fig. 1A, caveolin-2 was dramatically tyrosine phosphorylated by c-Src, whereas caveolin-3 failed to undergo detectable tyrosine phosphorylation.
To examine whether caveolin-2 undergoes phosphorylation on Tyr 19 , we next mutated this site to Ala. Fig. 1B shows that the Cav-2 (Y19A) mutant is no longer tyrosine phosphorylated, as predicted. Thus, it appears that Tyr 19 is the major site of caveolin-2 tyrosine phosphorylation; alternatively, phosphorylation at tyrosine 19 may be required before other sites can be phosphorylated.
A Phospho-specific Antibody Probe That Only Recognizes Tyr 19 -phosphorylated Caveolin-2-To directly monitor the phosphorylation state of caveolin-2 in vivo, we next generated a phospho-specific antibody probe that only recognizes Tyr 19phosphorylated caveolin-2. For this purpose, a synthetic peptide containing phosphorylated Tyr 19 and its flanking sequences (MADDA(pY)SHHSGC, residues 14 -25) was used to immunize rabbits. Polyclonal antibodies were obtained from rabbit serum by affinity purification using the peptide immunogen (see "Experimental Procedures"). Fig. 2 shows the high selectivity of anti-phosphocaveolin-2 IgG. We reconstituted this phosphorylation event in vivo by co-transfecting COS-7 cells with caveolin-2 and c-Src and monitored this event by Western blotting using anti-phosphocaveolin-2 IgG ( Fig. 2A). Importantly, anti-phosphocaveolin-2 IgG specifically recognized wild-type caveolin-2 in a phosphorylationdependent manner; detection was strictly dependent on cotransfection with the c-Src kinase. In addition, the mutant, Cav-2 (Y19A), was not detectable with this antibody probe, even when co-transfected with c-Src. Also, note that non-phosphorylated caveolin-2 migrates at ϳ20 kDa, while tyrosinephosphorylated caveolin-2 migrated at ϳ24 -28 kDa. A similar upward shift in eletrophoretic mobility was previously observed with phospho-caveolin-1 (Tyr(P) 14 ) (12).
We further tested the specificity of anti-phosphocaveolin-2 IgG using NIH 3T3 cells stably expressing the c-Src tyrosine kinase. This would allow us to detect the phosphorylation of endogenous caveolin-2 that is quite well expressed in NIH 3T3 cells. As shown in Fig. 2B, the tyrosine phosphorylation of caveolin-2 was dramatically increased in NIH 3T3 cells stably expressing c-Src, whereas little or no phosphorylation of caveolin-2 was detected in normal untransfected NIH 3T3 cells.
It has been previously shown that the phosphorylation of caveolin-1 is tightly regulated. For example, we demonstrated that insulin stimulation of 3T3-L1 adipocytes induces the ty-rosine phosphorylation of caveolin-1 on tyrosine 14 (12). It is believed that this insulin-stimulated phosphorylation event is mediated by an endogenous member of the Src family of tyrosine kinases (11). However, it remains unknown whether caveolin-2 also undergoes insulin-stimulated tyrosine phosphorylation. Therefore, 3T3-L1 adipocytes were serum-starved and insulin-stimulated. Indeed, insulin treatment was sufficient to induce the tyrosine phosphorylation of caveolin-2 (Fig. 2C). Thus, endogenous caveolin-2 normally undergoes phosphorylation on tyrosine 19 in response to physiologically relevant hormonal stimuli.
To further examine the specificity of the anti-phosphocaveolin-2 IgG, we performed competition assays using various synthetic peptides derived from caveolin-1 and -2 phosphorylation

Src-induced phosphorylation of caveolin-2 on tyrosine 19 in vivo.
A, caveolin-2 versus caveolin-3. COS-7 cells were transfected with C-terminally Myc-tagged caveolin-2 or -3, in the presence or absence of c-Src. Thirty-six hours post-transfection, the cells were processed for immunoprecipitation. Immunoprecipitates were then probed with a pAb that recognizes phosphotyrosine. Note that caveolin-2 was dramatically tyrosine-phosphorylated by c-Src, whereas caveolin-3 failed to undergo detectable tyrosine phosphorylation. Note that equivalent amounts of caveolin-2 and caveolin-3 were expressed, as determined by immunoblot analysis with mAb 9E10 directed against the Myc-epitope. B, caveolin-2 (WT versus Y19A). To examine whether caveolin-2 undergoes phosphorylation on Tyr 19 , we next mutated this site to Ala. COS-7 cells were transiently transfected with caveolin-2 (WT or Y19A), alone or in combination with c-Src. Thirty-six hours post-transfection, the cells were processed for immunoprecipitation. Total caveolin-2 was retrieved using a pAb directed against the Cterminal Myc epitope. Immunoprecipitates were then probed with a mAb that recognizes phosphotyrosine (mAb PY20). Note that when tyrosine 19 is mutated to alanine (Y19A), the reactivity of caveolin-2 with PY20 is completely abolished. motifs (Fig. 3). Immunoblots were prepared using COS-7 cells co-transfected with c-Src and caveolin-2. Subsequently, these immunoblots were then cut into strips and incubated with anti-phosphocaveolin-2 IgG, alone or in combination with the following three competing peptides: Cav-2 (Tyr 19 ), Cav-2 (Tyr(P) 19 ), or Cav-1 (Tyr(P) 14 ) (Fig. 3A).
Co-localization of Phosphocaveolin-2 (Tyr(P) 19 ) with Phosphocaveolin-1 (Tyr(P) 14 ) at Focal Adhesions in c-Src Expressing NIH 3T3 Cells-Previously we showed that phosphocaveolin-1 (Tyr(P) 14 ) is localized at focal contacts or adhesions (12). To examine the subcellular localization of tyrosine-phosphorylated caveolin-2, we next doubly immunostained cells with anticaveolin-2 IgG (mouse mAb) and anti-phosphocaveolin-2 (Tyr(P) 19 ) (rabbit pAb). These bound primary antibodies were visualized by using distinctly tagged fluorescent secondary antibodies (see "Experimental Procedures"). Fig. 4A shows the cellular distribution of phosphocaveolin-2 (Tyr(P) 19 ). In NIH 3T3 cells stably expressing wild-type c-Src, immunostaining with anti-phosphocaveolin-2 (Tyr(P) 19 ) IgG appeared as large fluorescent dots in the center and along the cell periphery (right), a pattern reminiscent of focal adhesions. Immunostaining with anti-caveolin-2 also appeared punctate, but the dots were much smaller in size (left). These two labeling patterns appeared distinct, suggesting that tyrosine-phosphorylated caveolin-2 is present in a different region of the cell or that only a subpopulation of caveolin-2 molecules are tyrosinephosphorylated. It should be noted that immunostaining at focal contacts was also detectable to a small extent with anticaveolin-2 IgG on longer exposures (not shown).
In NIH 3T3 cells stably expressing Src, immunostaining with anti-phosphocaveolin-1 appeared as large dots at focal contacts, and phosphocaveolin-1 was strictly co-localized with paxillin, a known marker of focal adhesions (12). To investigate FIG. 2. Characterization of a rabbit pAb probe that only recognizes tyrosine 19 phosphorylated caveolin-2. A tyrosine-phosphorylated caveolin-2 peptide (residues 14 -25; MADDA(pY)SHHSGC) was used to immunize rabbit and generate a phosphocaveolin-2 (Tyr(P) 19 )-specific pAb probe. A, caveolin-2 Y19A mutant. COS-7 cells were transiently transfected with caveolin-2 (WT or Y19A) and c-Src, alone or in combination. Note that anti-phosphocaveolin-2 (Tyr(P) 19 ) IgG only recognizes caveolin-2 when it is co-expressed with c-Src. Importantly, when tyrosine 19 is mutated to alanine, this immunoreactivity is completely abolished. Immunoblotting with a mAb that recognizes total caveolin-2 is shown as a control for equal loading. B, NIH 3T3 cells stably expressing c-Src. Lysates were prepared from untransfected NIH 3T3 cells and NIH 3T3 cells stably expressing c-Src. After SDS-PAGE and transfer to nitrocellulose, these samples were subjected to immunoblot analysis. Note that anti-phosphocaveolin-2 (Tyr(P) 19 ) IgG clearly recognizes endogenous caveolin-2 in NIH 3T3 cells that stably express c-Src. In contrast, little or no immunoreactivity was observed in normal NIH 3T3 cells. Each lane contains equal amounts of total protein. Immunoblotting with a mAb that recognizes total caveolin-2 is shown for comparison. C, insulin-stimulation of 3T3-L1 adipocytes. Fully differentiated adipocytes (day 8) were incubated in serum-free medium overnight, and then stimulated with insulin (150 nM) for 30 min. Note that caveolin-2 undergoes tyrosine phosphorylation in an insulin-dependent manner, as we have previously observed for caveolin-1 (12). 19 ) IgG: peptide competition with the immunogen and structurally related peptides. A, synthetic peptides. To further examine the specificity of the anti-phosphocaveolin-2 (Tyr(P) 19 ) IgG, we performed competition assays using various synthetic peptides derived from caveolin-1 and caveolin-2 phosphorylation motifs: Cav-2 (Tyr 19 ), Cav-2 (Tyr(P) 19 ), or Cav-1 (Tyr(P) 14 ). The sequences of these three caveolin-based peptides are shown. B, peptide competition. COS-7 cells were co-transfected caveolin-2 and c-Src and subjected to preparative SDS-PAGE. After transfer to nitrocellulose, the sheet was cut into strips and incubated with anti-phosphocaveolin-2 (Tyr(P) 19 ) IgG, alone or in combination with peptides. Note that the unphosphorylated caveolin-2 peptide (Cav-2 (Tyr 19 )), and the tyrosine-phosphorylated caveolin-1 peptide (Cav-1 (Tyr(P) 14 )) had no effect on the immunoreactivity of anti-phosphocaveolin-2 IgG. However, immunoreactivity was almost completely abolished when anti-phosphocaveolin-2 IgG were co-incubated with the tyrosine-phosphorylated caveolin-2 peptide (Cav-2 (Tyr(P) 19 ), the immunogen).
Tyrosine-phosphorylated Caveolin-2 (Tyr(P) 19 ) Remains Confined to Lipid Rafts, but Is No Longer Physically Associated with High Molecular Mass Caveolin Hetero-oligomers-As tyrosine-phosphorylated caveolin-2 (Tyr(P) 19 ) was localized in close proximity to focal adhesions, we next assessed the biochemical properties of tyrosine phosphorylated caveolin-2 using established assay systems. These approaches have been used previously by our group to characterize the properties of total caveolin-1 and phosphocaveolin-1 (Tyr(P) 14 ) (12).
Caveolin-1 and -2 normally form a stable high molecular mass hetero-oligomeric complex (5). This oligomerization event occurs right after or during the synthesis of caveolins at the level of the ER. Although the exact mechanism of how these proteins form oligomeric complexes remains unknown, the interaction between the membrane spanning domains of caveolin-2 and caveolin-1 is known to be required for hetero-oligomer assembly (31). Once these hetero-oligomers reach the trans-Golgi, they become resistant to Triton X-100 solubilization due to their incorporation into lipid rafts (14). This characteristic detergent insolubility is thought to reflect the specialized lipid microenvironment of rafts, which are highly enriched in sphingolipids and cholesterol. Although caveolin-1 can form stable homo-oligomers, caveolin-2 requires hetero-oligomerization with caveolin-1 for exit from the Golgi and its proper targeting to plasmalemmal caveolae. In the absence of caveolin-1, caveolin-2 remains in a monomer-dimer form, and is retained within the Golgi compartment (4,32), where it is subsequently degraded by the proteasome (8). Therefore, we next examined the effect of tyrosine phosphorylation of caveolin-2 (Tyr(P) 19 ) on its oligomeric state and targeting to lipid rafts/caveolae using NIH 3T3 cells (untransfected or stably expressing c-Src).
In both normal and c-Src-expressing NIH 3T3 cells, total caveolin-2 formed hetero-oligomers with caveolin-1 and was targeted to caveolae-enriched membrane fractions, as expected (Fig. 5, A and B). Interestingly, virtually identical results were obtained with caveolin-2 phosphorylated on tyrosine 19, indicating that caveolin-2 remains lipid raft-associated after tyrosine phosphorylation in NIH 3T3 cells that stably express c-Src. Similar results were noted for tyrosine phosphorylated caveolin-1 (Tyr(P) 14 ), as demonstrated in our previous report (12). As the subcellular distribution of tyrosine phosphorylated caveolin-2 (Tyr(P) 19 ) coincides with phosphocaveolin-1 (Tyr(P) 14 ) at focal adhesions by fluorescence microscopy (Fig.  4), these biochemical results are consistent with the idea that lipid rafts/caveolae in close proximity to focal adhesions are preferentially targeted for tyrosine phosphorylation.
Surprisingly, however, we found that tyrosine-phosphorylated caveolin-2 (Tyr(P) 19 ) was not present in a high molecular mass complex, but instead migrated as a monomer/dimer in velocity gradients (Ͻ66 kDa). As shown in Fig. 5B, total caveolin-2 was still present within large oligomeric complexes (Ն150 kDa) in both untransfected and c-Src expressing NIH 3T3 cells. As we have previously shown that total caveolin-1 and phosphocaveolin-1 (Tyr(P) 14 ) migrate exclusively as high molecular mass complexes in this gradient system (12), we conclude that phosphocaveolin-2 (Tyr(P) 19 ) is no longer physically associated with high molecular mass caveolin-1 oligomers, but remains associated with lipid rafts (Fig. 5A). Exactly the same monomer-dimer migration pattern is observed when total caveolin-2 is expressed in the absence of caveolin-1 (4); however, in this case caveolin-2 remains Triton-soluble (i.e. is not lipid raft associated) and is trapped at the level of the Golgi complex.

FIG. 4. Tyrosine 19-phosphorylated caveolin-2 is localized in close proximity to focal adhesions in NIH 3T3 cells stably expressing c-Src.
Cells were doubly immunostained with two distinct primary antibodies as detailed below. Bound primary antibodies were visualized by using distinctly tagged fluorescent secondary antibodies. In A and B, images were acquired with an Olympus IX80 microscope with a 60ϫ Plan Neofluar objective and a Photometrics cooled CCD camera with a 35-mm shutter. A, anti-caveolin-2 (mAb) and anti-phosphocaveolin-2 (pAb). Note that staining with anti-phosphocaveolin-2 (Tyr(P) 19 ) appeared as large irregular dots, but these dots were confined to the cell periphery and appeared to coincide with focal contacts or adhesions (right). Immunostaining with anti-caveolin-2 also appeared punctate, but the dots were much smaller in size (left). These two labeling patterns are distinct, suggesting that tyrosine-phosphorylated caveolin-2 (Tyr(P) 19 ) is present in a different region of the cell or that only a subpopulation of caveolin-2 molecules are tyrosine-phosphorylated. B, anti-phospho-caveolin-1 (mAb) and anti-phospho-caveolin-2 (pAb). Note that tyrosine-phosphorylated caveolin-2 (Tyr(P) 19 ) is colocalized with tyrosine-phosphorylated caveolin-1 (Tyr(P) 14 ). In the case of caveolin-1, these sites of tyrosine phosphorylation activity are known to correspond to focal adhesions (12).
Caveolae-enriched domains were purified from murine lung tissue using an established protocol (24) and incubated in kinase reaction buffer in the presence or absence of exogenous ATP (1 mM). Fig. 6A (right top panel) shows that the addition of ATP induced phosphorylation of caveolin-2 on tyrosine 19. Thus, a kinase activity with the same specificity as is observed in vivo is present within purified caveolae membranes. Other tyrosine-phosphorylated proteins were visualized by immunoblotting with an anti-phosphotyrosine mAb (PY20) (left panel); the major tyrosine-phosphorylated proteins appeared in the ϳ50 -60 kDa and the ϳ200 kDa range, consistent with the autophosphorylation of Src (right lower panel) and Src family tyrosine kinases and receptor tyrosine kinases, respectively. It should be noted that a series of ϳ24 -28-kDa bands corresponding to phosphocaveolin-2 were detectable with PY20 as well.
We previously showed that the phosphorylation of caveolin-1 by c-Src was dependent on two Src "domains," i.e. the N-terminally attached myristate moiety and the adjacent N-terminal protein sequence proximal to the myristoylation site (12). The importance of these c-Src domains for caveolin-1 tyrosine phosphorylation was determined using a competition assay with synthetic peptides, thereby mimicking these Src domains. We utilized the same approach to determine whether c-Src phosphorylates caveolin-2 by a similar mechanism.
Myristoylated or non-myristoylated N-terminal c-Src peptides were preincubated with purified caveolae from mouse lung tissue prior to the initiation of the kinase reaction. The kinase reaction was then initiated by adding exogenous ATP. As shown in Fig. 6B, only the N-Myr-c-Src peptide greatly reduced the phosphorylation of caveolin-2 in a concentrationdependent manner. Importantly, the non-myristoylated c-Src peptide was unable to compete with endogenous Src for phosphorylation of caveolin-2 at the same concentration.
As a control for specificity, myristoylated and non-myristoylated N-terminal c-Yes peptides were also evaluated. However, they failed to compete with endogenous Src for phosphorylation of caveolin-2 at the same concentration. These results are consistent with the idea that caveolin-2 phosphorylation is induced by endogenous Src that is localized within caveolae.
It remains unknown whether tyrosine phosphorylation of caveolin-2 alone is sufficient to cause its dissociation from caveolin-1. Another more likely possibility is that tyrosine phosphorylation of caveolin-2 is simply recognized as a signal by the cellular machinery to catalyze the dissociation of caveolin-2 from caveolin-1. To distinguish between these possibilities, we took advantage of our ability to reconstitute the tyrosine phosphorylation of caveolin-2 in vitro.
Caveolae-enriched domains purified from murine lung tissue were incubated in kinase reaction buffer in the presence or absence of exogenous ATP (1 mM). Afterward, the caveolae were harvested, solubilized with octyl glucoside, and subjected to velocity gradient centrifugation. Interestingly, Fig. 6C shows that phosphocaveolin-2 (Tyr(P) 19 ) is oligomeric under these conditions, indicating that it remains associated with high molecular mass caveolin-1 oligomers. Thus, we favor the idea that tyrosine phosphorylation of caveolin-2 is recognized as a signal by the cellular machinery to catalyze the dissociation of caveolin-2 from caveolin-1.

Vanadate Treatment Induces the Accumulation of Tyrosine 19-phosphorylated Caveolin-2 in Normal NIH 3T3 Cells-It
remains unclear what ligands might stimulate tyrosine phosphorylation of caveolins in NIH 3T3 cells. To investigate whether caveolin-2 undergoes tyrosine phosphorylation in normal NIH 3T3 cells, these cells were treated with vanadate (100 M), a tyrosine phosphatase inhibitor. If caveolin-2 undergoes tyrosine phosphorylation, vanadate should prevent de-phosphorylation and allow detection of accumulated tyrosine-phosphorylated intermediates. Fig. 7 shows that treatment with vanadate for increasing time leads to the accumulation of tyrosine 19 phosphorylated caveolin-2.
Identification of SH2 Domain Containing Binding Partners for Phosphocaveolin-2 (Tyr(P) 19 ): Ras-GAP, c-Src, and Nck-One of the known functions of tyrosine phosphorylation is to serve as a binding site for the recruitment of SH2 domain containing proteins. To identify SH2-bearing proteins that specifically bind to phosphorylated caveolin-2, we used an estab-FIG. 5. Tyrosine-phosphorylated caveolin-2 (Tyr(P) 19 ) is targeted to lipid rafts/caveolae, but no longer forms a high molecular mass hetero-oligomer with caveolin-1. In both panels, tyrosine-phosphorylated caveolin-2 (Tyr(P) 19 ) was detected with a novel pAb probe, and total caveolin-2 was visualized using mAb 65. A, caveolar targeting. NIH 3T3 cells (untransfected or stably expressing c-Src) were homogenized in a buffer containing 1% Triton X-100 and subjected to sucrose density gradient centrifugation. Twelve 1-ml fractions were collected, and an aliquot of each fraction (ϳ20 l) was analyzed by SDS-PAGE/Western blotting. As expected, caveolin-2 is highly enriched in fractions 4 -5 that represent the lipid raft/caveolae-enriched membrane fractions. Note that both total caveolin-2 and tyrosine-phosphorylated caveolin-2 (Tyr(P) 19 ) are correctly targeted to the lipid raft/ caveolae-enriched membrane fractions in both normal and c-Src expressing NIH 3T3 cells. B, velocity gradient analysis. NIH 3T3 cells (untransfected or stably expressing c-Src) were solubilized, loaded atop a 5-40% sucrose density gradient, and subjected to centrifugation for 10 h. Twelve fractions were recovered, and a 20-l aliquot from each fraction was analyzed by SDS-PAGE/Western blotting. Arrows mark the positions of molecular mass standards. Note that phosphocaveolin-2 (Tyr(P) 19 ) behaves as monomer/dimer (Ͻ66 kDa) in these velocity gradients (upper panel). In contrast, total caveolin-2 migrates as a high molecular mass oligomer (Ͼ150 kDa; middle and lower panels), precisely coinciding with the distribution of caveolin-1 ((12); data not shown).
lished in vitro binding approach (12,34). Briefly, we purified GST fusion proteins containing either non-phosphorylated or tyrosine-phosphorylated caveolin-2 from two different bacterial strains ( Fig. 8A; and see "Experimental Procedures"). These purified GST-Cav-2-(1-86) fusion proteins were then incubated with lysates prepared from normal NIH 3T3 cells. After extensive washing, the bound material was subjected to Western blot analysis with a panel of antibodies directed against SH2 domain containing proteins (listed in Table II). Fig. 8B shows that among the over 20 proteins surveyed, only three proteins (Ras-GAP, c-Src, and Nck) showed binding activity toward tyrosine-phosphorylated caveolin-2. Importantly, these proteins bound only to tyrosine-phosphorylated GST-Cav-2-(1-86), but not to GST alone or non-phosphorylated GST-Cav-2- (1-86). Interestingly, Grb7, which binds to phosphorylated caveolin-1 (12), did not show any binding activity toward phospho-caveolin-2, clearly demonstrating the specificity of this binding event (lower panel).
To assess the possible relevance of these binding events in vivo, we next performed a series of well controlled co-immunoprecipitation experiments using NIH 3T3 (untransfected versus c-Src stably transfected), which endogenously express both Ras-GAP and caveolin-2. Lysates were subjected to immunoprecipitation with antibodies directed against either Ras-GAP or caveolin-2, followed by Western blot analysis with anticaveolin-2 IgG or anti-Ras-GAP IgG. Fig. 9 shows that Ras-GAP and caveolin-2 co-immunoprecipitate in NIH 3T3 cells that stably express c-Src, but not in untransfected NIH 3T3 cells, as predicted. Therefore, the interaction between caveolin-2 and Ras-GAP is strictly dependent on tyrosine phosphorylation of caveolin-2. These results indicate that the tyrosine phosphorylation of caveolin-2 recruits the binding of a specific subset of SH2 domain-containing proteins both in vitro and in vivo.
Integrin Ligation Induces the Tyrosine Phosphorylation of Caveolin-2 (Tyr(P) 19 ), Which Co-localizes with Activated FAK at Focal Adhesions-A variety of extracellular stimuli are known to induce the tyrosine phosphorylation of caveolin-1 (Tyr(P) 14 ), such as treatment with growth factions (insulin and EGF) or osmotic shock (13,(35)(36)(37). Recently, Mettouchi et al. (38) showed that the tyrosine phosphorylation of caveolin-1 (Tyr(P) 14 ) can also be stimulated by integrin ligation. By the use of immunostaining techniques on human umbilical vein endothelial cells, they showed that cellular adhesion to fibronectin led to the accumulation of phosphocaveolin-1 (Tyr(P) 14 ) at focal adhesions.
Based on our current observation that phosphocaveolin-2 (Tyr(P) 19 ) and phosphocaveolin-1 (Tyr(P) 14 ) co-localize at focal adhesions in NIH 3T3 cells stably expressing c-Src, we hypothesized that caveolin-2 might also undergo tyrosine phosphorylation in response to integrin ligation. For this purpose, we used human microvascular endothelial cells (HMECs), which are known to abundantly express caveolin-1 and -2 (19), and plated them on fibronectin (FN)-coated plates to determine whether integrin ligation induces the tyrosine phosphorylation of caveolin-2. HMECs plated on poly-L-lysine (PLL)-coated plates were FIG. 6. Purified caveolae-enriched membranes contain a kinase that phosphorylates caveolin-2 on tyrosine 19 in vitro. A, in vitro kinase assay. Caveolae-enriched domains were purified from murine lung tissue using an established protocol (24) and incubated in kinase reaction buffer in the presence or absence of exogenous ATP (1 mM). After 10 min at 25°C, the reaction was halted by the addition of SDS sample buffer/boiling, and samples were subjected to Western blotting. Note that addition of ATP dramatically induced phosphorylation of caveolin-2 on tyrosine 19 (right, upper panel). The total pattern of tyrosine-phosphorylated proteins was visualized by immunoblotting with an anti-phosphotyrosine mAb (PY20) (left panel). Immunoblotting with a mAb that recognizes total caveolin-2 is shown as a control for equal loading (right, middle panel). Note that Src is present within caveolae-enriched membranes, as we have shown previously (12,24), and is activated by the addition of ATP (right, lower panel). Activated Src was visualized with anti-c-Src (Tyr(P) 416 ) IgG (pAb; Cell Signaling Technology). B, Src family kinase peptide competition. Caveolae-enriched membrane domains were purified from murine lung tissue and incubated in kinase reaction buffer in the presence of exogenous ATP (1 mM). Different concentrations (10 or 50 M) of myristoylated or nonmyristoylated peptides (detailed in Table I) were added to the kinase reaction. Samples were then subjected to SDS-PAGE/Western blot analysis with anti-phosphocaveolin-2 (Tyr(P) 19 ) IgG (pAb; upper panel) or anti-caveolin-2 IgG (mAb; lower panel). Note that the N-Myr-c-Src peptide (50 M) completely abolished the phosphorylation of caveolin-2 on tyrosine 19 by an endogenous caveolar tyrosine kinase. In contrast, the N-Myr-c-Yes peptide had no effect. Similarly, the non-myristoylated c-Src peptide also had no effect. Thus, the inhibitory activity of the N-Myr-c-Src peptide is both myristoylation-dependent and sequencespecific. C, velocity gradient analysis. Caveolae-enriched membrane domains were purified from murine lung tissue and incubated in kinase reaction buffer in the presence of exogenous ATP (1 mM). After 10 min at 25°C, the reaction was halted by the addition of MBS buffer containing 60 mM octyl glucoside and phosphatae inhibitors, and the samples were subjected to velocity gradient cetrifugation, followed by SDS-PAGE and Western blotting. Note that phosphocaveolin-2 (Tyr(P) 19 ) behaves as a high molecular mass oligomer under these conditions (see Fig. 5B for comparison).
used as an appropriate negative control for these studies and were processed in parallel. Cells were examined by fluorescence microscopy ϳ20 min after plating.
In agreement with the work of Mettouchi et al. (38), tyrosinephosphorylated caveolin-1 (Tyr(P) 14 ) was easily detected in cells seeded on FN-coated plates and was localized along the cell periphery, as expected. However, little or no phosphorylated caveolin-1 (Tyr(P) 14 ) was detected in cells seeded on PLLcoated plates (Fig. 10A, upper panels). Importantly, tyrosinephosphorylated caveolin-2 (Tyr(P) 19 ) was also easily detected in cells seeded on FN-coated plates, but a significant intracellular perinuclear pool was present, in addition to the plasma membrane-associated pool. The plasma membrane-associated pool of phosphocaveolin-2 (Tyr(P) 19 ) clearly coincided with the distribution of phosphocaveolin-1 (Tyr(P) 14 ). Similarly, little or no phosphorylated caveolin-2 (Tyr(P) 19 ) was detected in cells seeded on PLL-coated plates (Fig. 10A, lower panels), indicating that the tyrosine phosphorylation of caveolin-2 is FN-dependent.
We next performed double labeling experiments with antibodies directed against total phosphotyrosine. Interestingly, the distribution of phosphocaveolin-2 (Tyr(P) 19 ) precisely colocalized with the major sites of tyrosine phosphorylation, i.e. focal adhesions, in cells plated on FN (Fig. 10, B and C). Virtually identical results were obtained with phosphocaveolin-1 (Tyr(P) 14 ) (Fig. 10, B and C).
To assess the activation state of FAK under these conditions, HMECs were next immunostained with a phospho-specific antibody that only recognizes the activated form of FAK (Tyr(P) 397 ). As shown in Fig. 10D, only HMECs plated on FN showed staining with anti-phospho-FAK (Tyr(P) 397 ); no staining was observed with cells plated on PLL, as expected. Note also that phospho-FAK (Tyr(P) 397 ) and phospho-caveolin-2 (Tyr(P) 19 ) show significant co-localization, suggesting that lipid rafts/caveolae in close proximity to focal adhesions are preferentially phosphorylated during integrin ligation. DISCUSSION Caveolin-1 and -2 are co-expressed and form a high molecular mass hetero-oligomeric complex in many cell types, particularly in fibroblasts, adipocytes, and endothelial cells (7). The exact physiological role of caveolin-2 remains unknown; how-ever, caveolin-2 is thought to function as an "accessory protein" in conjunction with caveolin-1 (4).
Caveolin-1 was originally identified as a major substrate of v-Src and it co-purifies with Src family tyrosine kinases (9,39). Previously, our group and others have shown that c-Src also induces the phosphorylation of caveolin-1 on tyrosine 14, and phosphocaveolin-1 (Tyr(P) 14 ) is preferentially localized within lipid rafts/caveolae near focal adhesions (12). Although the primary sequence of caveolin-2 is highly divergent from caveo- FIG. 7. Vanadate treatment induces the accumulation of tyrosine 19-phosphorylated caveolin-2 in normal NIH 3T3 cells. To investigate whether caveolin-2 undergoes tyrosine phosphorylation in normal NIH 3T3 cells, these cells were treated with vanadate (100 M), a tyrosine phosphatase inhibitor. Note that treatment with vanadate for increasing times leads to the accumulation of tyrosine 19-phosphorylated caveolin-2 (upper panel). Immunoblotting with a mAb that recognizes total caveolin-2 is shown as a control for equal loading (lower panel).

FIG. 8. Identification of SH2 domain-containing proteins that bind to caveolin-2 in a phosphorylation-dependent manner:
Ras-GAP, c-Src, and Nck. A, GST-caveolin-2 fusion proteins. A GST fusion protein carrying the N-terminal domain of caveolin-2-(1-86) was purified from normal bacteria (BL21) or from a bacterial strain harboring a tyrosine kinase (TKB1). Note that anti-phosphotyrosine IgG (mAb PY20) only recognized tyrosine-phosphorylated caveolin-2 produced in the TKB1 strain (middle panel), despite equal protein loading (upper panel). Anti-phosphocaveolin-2 (Tyr(P) 19 ) IgG only recognized GSTcaveolin-2-(1-86) produced in the TKB1 strain as well. These results indicate that GST-caveolin-2 (1-86) produced in the TKB1 strain is correctly phosphorylated on tyrosine 19. B, in vitro binding of SH2 domain containing proteins. Non-phosphorylated and tyrosine-phosphorylated GST-caveolin-2-(1-86) fusion proteins were prepared as in A and incubated with cell lysates derived from normal NIH 3T3 cells. After binding, washing, and elution, the eluates were subjected to immunoblot analysis with antibodies directed against 24 different SH2 domain containing proteins (detailed in Table II). Note that, of the proteins surveyed, only three SH2 domain-containing proteins showed binding activity toward caveolin-2. These proteins included Ras-GAP, c-Src, and Nck. Importantly, all three of SH2 domain-containing proteins bound to tyrosine-phosphorylated GST-caveolin-2-(1-86), but not to GST alone or non-phosphorylated GST-caveolin-2. We have previously shown the phosphocaveolin-1 (Tyr(P) 14 ) physically associates with the SH2 domain adapter protein, Grb7, using the same assay system (12). In contrast, phosphocaveolin-2 (Tyr(P) 19 ) did not show any binding activity toward Grb7.
lin-1 and -3, the caveolin-2 sequence still contains a well conserved motif for recognition by tyrosine kinases at its extreme N terminus (centered around tyrosine 19). Based on these observations, we hypothesized that caveolin-2 may undergo Src-induced phosphorylation on tyrosine 19.
Here, we studied the tyrosine phosphorylation of caveolin-2 in vivo by employing a novel phospho-specific polyclonal antibody that selectively recognizes caveolin-2 phosphorylated on tyrosine 19. First, we characterized the specificity of anti-phosphocaveolin-2 (Tyr(P) 19 ) IgG by reconstituting the tyrosine phosphorylation of caveolin-2 both in vivo and in vitro. Cotransfection of caveolin-2 with c-Src resulted in the tyrosine phosphorylation of caveolin-2; however, caveolin-3 failed to undergo tyrosine phosphorylation under identical conditions, further demonstrating the specificity of these interactions. When tyrosine 19 was replaced with alanine, this phosphorylation event was completely abolished as measured using antibodies that recognize generic phosphotyrosine, such as the mAb PY20. Therefore, we identified tyrosine 19 within caveolin-2 as the site of Src-induced tyrosine phosphorylation. NIH 3T3 cells stably expressing c-Src showed high levels of phosphocaveolin-2 (Tyr(P) 19 ) at steady state, which was clearly localized near focal contacts/adhesions, the major cellular sites of tyrosine kinase-mediated signal transduction. Interestingly, phosphocaveolin-2 (Tyr(P) 19 ) was strictly co-localized with phosphocaveolin-1 (Tyr(P) 14 ), but did not co-localize with total caveolin-2.
We also evaluated the biochemical properties of phosphocaveolin-2 (Tyr(P) 19 ). At the level of the ER, caveolin-1 and -2 are known to tightly interact and they form a high molecular mass hetero-oligomer, containing ϳ14 -16 individual caveolin molecules (5,40). At the level of the Golgi, these ϳ300ϳ350-kDa hetero-oligomer units are thought to undergo a second stage of oligomerization via homotypic C-terminal interactions, giving rise to the formation of a caveolae-sized vesicle that is targeted to the plasma membrane (41). Interestingly, we show here that phosphocaveolin-2 (p19) remains associated with lipid rafts/caveolae, as judged by its Triton insolubility and buoyancy in sucrose density gradients. Perhaps surprisingly, however, phosphocaveolin-2 (Tyr(P) 19 ) no longer forms a high molecular mass hetero-oligomeric complex with caveolin-1. Instead, phosphocaveolin-2 (Tyr(P) 19 ) migrates as a monomer/ dimer in velocity gradients and is clearly separated from high molecular mass caveolin-1/caveolin-2 hetero-oligomers. However, in vitro phosphorylation experiments using purified caveolae membranes indicated that the phosphorylation of caveolin-2 on tyrosine 19 is not sufficient to disrupt its interaction with caveolin-1. Thus, we conclude that the tyrosine phosphorylation of caveolin-2 (Tyr(P) 19 ) functions as a "signal" that is recognized by the cellular machinery to induce the dissociation of caveolin-2 from caveolin-1 oligomers.
It appears that the Src-mediated phosphorylation of caveolin-2 is regulated in a similar manner as we observed for caveolin-1. Purified caveolae membranes contain a kinase activity that phosphorylates caveolin-2 on tyrosine 19 in vitro. In addition, we show that the tyrosine phosphorylation of caveolin-2 could be inhibited by competition with a myristoylated peptide derived from the N terminus of c-Src. This inhibition was myristoylation-dependent and sequence-specific, as the corresponding non-myristoylated c-Src peptide and a myristoylated c-Yes peptide had no effect in this assay system. These data provide evidence that both lipid modification and the adjacent N-terminal sequence of c-Src are important to mediate the phosphorylation of caveolin-2 on tyrosine 19.
Interestingly, among the 20 phosphotyrosine-binding proteins we evaluated as potential binding partners for phosphocaveolin-2 (Tyr(P) 19 ), we showed that only three SH2 domaincontaining proteins bind to caveolin-2, namely c-Src, NCK, and Ras-GAP. It is important to note that binding was strictly dependent on the tyrosine phosphorylation of caveolin-2. In contrast, we have previously shown the phosphocaveolin-1 (Tyr(P) 14 ) physically associated with the SH2 domain adapter protein Grb7. However, phosphocaveolin-2 (Tyr(P) 19 ) did not show any binding activity toward Grb7. Thus, phosphocaveolin-1 (Tyr(P) 14 ) and phosphocaveolin-2 (Tyr(P) 19 ) may serve as binding partners for different SH2 domain containing proteins. As such, the tyrosine phosphorylation of caveolin family members would lead to the recruitment and compartmentalization of distinct signaling cascades.
Ras-GAP acts as an inhibitor of Ras signaling by interacting with activated GTP-bound Ras and stimulating its otherwise weak intrinsic GTPase activity, thereby catalyzing the conversion of GTP to GDP. Within its N-terminal regions, p120-Ras-GAP contains two SH2 domains and one intervening SH3domain. The Ras-GTPase stimulating activity has been localized to a C-terminal region. The interaction of phosphocaveolin-2 (Tyr(P) 19 ) with Ras-GAP may provide an additional mechanism by which Ras-GAP activity may be regulated. For example, the suppressor of cytokine signaling-3 (SOCS-3) has been shown to bind and regulate p120Ras-GAP in a phosphorylation-dependent manner. In response to growth factors, SOCS-3 undergoes phosphorylation and subsequently binds to FIG. 9. Ras-GAP binds phosphocaveolin-2 (Tyr(P) 19 ) in a phosphorylation-dependent manner in vivo. Cell lysates derived from normal and c-Src stably transfected NIH 3T3 cells were prepared and subjected to immunoprecipitation with antibodies directed against either Ras-GAP or caveolin-2. These immunoprecipitates were then subjected to Western blot analysis with anticaveolin-2 IgG or anti-Ras-GAP IgG (right panels). Note that Ras-GAP and caveolin-2 co-immunprecipitate in NIH 3T3 cells that stably express c-Src, but not in untransfected NIH 3T3 cells. These results indicate that this interaction is strictly dependent on co-expression with c-Src. Importantly, normal and c-Src stably transfected NIH 3T3 both express equivalent amounts of Ras-GAP and caveolin-2 (left panels).
p120Ras-GAP, leading to sustained Ras activity (42). Here, we show that caveolin-2 can also bind Ras-GAP in a phosphorylationdependent manner. Although further investigation is required to discover the exact function of phosphocaveolin-2, we speculate that phosphorylated caveolin-2 may inhibit Ras-GAP activity through its direct interaction with Ras-GAP.
Growth factors and integrins are known to activate the Ras-p42/44 MAP kinase pathway. At the cell periphery, these stimuli assemble focal adhesions to activate signaling through FAK and other molecules. Giancotti and his colleagues (37) have proposed that caveolin-1 participates in integrin signaling by recruiting Fyn, which results in activation of the Shc/Grb2/Ras-p42/44 MAP kinase cascade. In agreement with this proposed model, phospho-caveolin-2 (Tyr(P) 19 ) may contribute to Ras activation through its direct interaction with Ras-GAP. We previously showed that phosphorylated caveolin-1 (Tyr(P) 14 ) is localized near focal adhesion and binds to Grb7, a molecule that interacts directly with FAK (43,44). Similarly, phosphorylated caveolin-2 (Tyr(P) 19 ) may serve to link signaling molecules to the focal adhesion machinery in concert with phosphorylated FIG. 10. FN-mediated integrin ligation induces the tyrosine phosphorylation of caveolin-2 near focal adhesions. Based on our current observation that phosphocaveolin-2 (Tyr(P) 19 ) and phosphocaveolin-1 (Tyr(P) 14 ) co-localize at focal adhesions in NIH 3T3 cells stably expressing c-Src, we hypothesized that caveolin-2 might also undergo tyrosine phosphorylation in response to integrin ligation. For this purpose, we used HMECs that are known to abundantly express caveolin-1 and -2 (19) and plated them on FN-coated plates. HMECs plated on PLL-coated plates were used as an appropriate negative control for these studies and were processed in parallel. Cells were examined by fluorescence microscopy 20 min after plating. A, phosphocaveolin-1 (Tyr(P) 14 ; mAb) and phosphocaveolin-2 (Tyr(P) 19 ; pAb). Tyrosine-phosphorylated caveolin-1 (Tyr(P) 14 ) was easily detected in cells seeded on FN-coated plates and was localized along the cell periphery, as expected. However, little or no phosphorylated caveolin-1 (Tyr(P) 14 ) was detected in cells seeded on PLL-coated plates (upper panels). Importantly, tyrosine-phosphorylated caveolin-2 (Tyr(P) 19 ) was also easily detected in cells seeded on FN-coated plates, but a significant intracellular peri-nuclear pool was present, in addition to the plasma membrane-associated pool. The plasma membrane-associated pool of phosphocaveolin-2 (Tyr(P) 19 ) clearly coincided with the distribution of phosphocaveolin-1 (Tyr(P) 14 ). Similarly, little or no phosphorylated caveolin-2 (Tyr(P) 19 ) was detected in cells seeded on PLL-coated plates (lower panels), indicating that the tyrosine phosphorylation of caveolin-2 is FN-dependent. B, phosphocaveolin-1 (Tyr(P) 14 ; mAb) and anti-phospho-tyrosine (pAb). We next performed double labeling experiments with antibodies directed against total phosphotyrosine. Interestingly, the distribution of phosphocaveolin-1 (Tyr(P) 14 ) precisely co-localized with the major sites of tyrosine phosphorylation in cells plated on FN. C, phosphocaveolin-2 (Tyr(P) 19 ; pAb) and anti-phosphotyrosine (PY20; mAb). The distribution of phosphocaveolin-2 (Tyr(P) 19 ) also precisely colocalized with the major sites of tyrosine phosphorylation in cells plated on FN. D, phospho-FAK (Tyr(P) 397 ; mAb) and phosphocaveolin-2 (Tyr(P) 19 ; pAb). To assess the activation state of FAK under these conditions, HMECs were next immunostained with a phospho-specific antibody that only recognizes the activated form of FAK (Tyr(P) 397 ). Note that only HMECs plated on FN show staining with anti-phospho-FAK (Tyr(P) 397 ); no staining was observed with cells plated on PLL, as expected. Note also that phospho-FAK (Tyr(P) 397 ) and phosphocaveolin-2 (Tyr(P) 19 ) show significant co-localization, suggesting that lipid rafts/caveolae in close proximity to focal adhesions are preferentially phosphorylated during integrin ligation. caveolin-1 (Tyr(P) 14 ). In support of this hypothesis, phosphocaveolin-2 (Tyr(P) 19 ) and phosphocaveolin-1 (Tyr(P) 14 ) are both co-localized to a subpopulation of lipid rafts/caveolae that are proximal to focal adhesions. In addition, we showed here that integrin ligation (using fibronectin-coated plates) specifically induces the tyrosine phosphorylation of both caveolin-1 (Tyr(P) 14 ) and caveolin-2 (Tyr(P) 19 ), which co-localize with activated FAK (Tyr(P) 397 ) at focal adhesions in endothelial cells.