Src Tyrosine Kinases, G a Subunits, and H-Ras Share a Common Membrane-anchored Scaffolding Protein, Caveolin

Caveolae are plasma membrane specializations present in most cell types. Caveolin, a 22-kDa integral membrane protein, is a principal structural and regulatory component of caveolae membranes. Previous studies have demonstrated that caveolin co-purifies with lipid modified signaling molecules, including Gα subunits, H-Ras, c-Src, and other related Src family tyrosine kinases. In addition, it has been shown that caveolin interacts directly with Gα subunits and H-Ras, preferentially recognizing the inactive conformation of these molecules. However, it is not known whether caveolin interacts directly or indirectly with Src family tyrosine kinases. Here, we examine the structural and functional interaction of caveolin with Src family tyrosine kinases. Caveolin was recombinantly expressed as a glutathione S-transferase fusion. Using an established in vitro binding assay, we find that caveolin interacts with wild-type Src (c-Src) but does not form a stable complex with mutationally activated Src (v-Src). Thus, it appears that caveolin prefers the inactive conformation of Src. Deletion mutagenesis indicates that the Src-interacting domain of caveolin is located within residues 82-101, a cytosolic membrane-proximal region of caveolin. A caveolin peptide derived from this region (residues 82-101) functionally suppressed the auto-activation of purified recombinant c-Src tyrosine kinase and Fyn, a related Src family tyrosine kinase. We further analyzed the effect of caveolin on c-Src activity in vivo by transiently co-expressing full-length caveolin and c-Src tyrosine kinase in 293T cells. Co-expression with caveolin dramatically suppressed the tyrosine kinase activity of c-Src as measured via an immune complex kinase assay. Thus, it appears that caveolin structurally and functionally interacts with wild-type c-Src via caveolin residues 82-101. Besides interacting with Src family kinases, this cytosolic caveolin domain (residues 82-101) has the following unique features. First, it is required to form multivalent homo-oligomers of caveolin. Second, it interacts with G-protein α-subunits and down-regulates their GTPase activity. Third, it binds to wild-type H-Ras. Fourth, it is membrane-proximal, suggesting that it may be involved in other potential protein-protein interactions. Thus, we have termed this 20-amino acid stretch of caveolin residues the caveolin scaffolding domain.

tions. Thus, we have termed this 20-amino acid stretch of caveolin residues the caveolin scaffolding domain.
Caveolae are small bulb-shaped invaginations located at or near the cell surface (1). They represent a micro-domain of the plasma membrane (1,2). Although caveolae are present in most cells, they are most abundant in terminally differentiated cell types: endothelia, adipocytes, muscle cells (skeletal, cardiac, and smooth), and type I pneumoctyes (reviewed in Refs. 3 and 4). For example, in adipocytes they may occupy up to 20% of the total plasma membrane surface area (5). In striking contrast, caveolin and caveolae are reduced or absent in fibroblasts transformed by certain activated oncogenes (such as v-Abl or H-Ras (G12V)) (6).
In this regard, several independent lines of evidence suggest that caveolin may function as a scaffolding protein within caveolae membranes. (i) Both the N-terminal and C-terminal domains of caveolin face the cytoplasm, allowing them to freely interact with cytosolic molecules (22)(23)(24). In accordance with this membrane topology, caveolin remains inaccessible to biotinylation probes that have been used to efficiently label proteins that face the extracellular environment (25). (ii) Caveolin undergoes two stages of oligomerization. First, caveolin monomers assemble into discrete multivalent oligomers containing ϳ14 -16 monomers per oligomer (25,26). Subsequently, these individual caveolin homo-oligomers (4 -6-nm particles) can interact with each other to form caveolae-like structures in vitro (25-50 nm clusters) (25). (iii) A cytosolic membrane proximal domain of caveolin (residues 82-101) interacts directly with G ␣ subunits and H-Ras (19,27). This caveolin region preferentially recognizes the inactive conformation of these molecules, because mutationally activated G ␣ subunits (G s ; Q227L) and H-Ras (G12V) fail to interact with caveolin (19,27). (iv) Interaction of caveolin with purified heterotrimeric G-proteins functionally suppresses their GTPase activity, holding the G-pro-tein in the inactive conformation (27). Thus, caveolin may organize the formation of caveolae microdomains and orchestrate caveolae-related signaling events.
Here, we examine the structural and functional interaction of caveolin with Src family tyrosine kinases. We find that caveolin interacts directly with wild-type Src (c-Src) but fails to stably interact with a mutationally activated form of Src (v-Src). The Src-interacting domain of caveolin was localized to caveolin residues 82-101, and a peptide encoding this sequence dose-dependently suppressed the auto-activation of purified Src kinases (c-Src and Fyn). Furthermore, transient co-expression of c-Src with the full-length caveolin in 293T cells inhibited the kinase activity of c-Src.
Thus, it appears that three distinct classes of lipid-modified signaling molecules (G ␣ subunits (27), H-Ras (19), and c-Src (this report)) that normally co-purify with caveolin all recognize the same cytosolic membrane-proximal region of caveolin (residues 82-101). In addition, caveolin prefers the inactive conformation of these molecules, and caveolin binding can hold these molecules in the inactive conformation. As such, caveolin may function as a common membrane-anchored scaffolding protein for these and other cytoplasmic signaling molecules.
Our current observations are analogous to another family of scaffolding proteins (28), the AKAPs (A-kinase anchor proteins), which preferentially recognize the inactive conformation of protein kinase A, protein phosphatase 2B (calcineurin), and protein kinase C (␣and ␤-isoforms) (29,30). As we find with caveolin, each of these enzymes is inhibited when bound to AKAP-79 (29,30).
Cell Culture-Insect Sf21 cells were provided by Dr. Takashi Okamoto (Massachusetts General Hospital/Harvard Medical School). Sf21 cells were grown in Ex-cell 400 medium, 10% fetal bovine serum, and 1% penicillin-streptomycin at 27°C. For double transfection experiments, 293T cells (gift of Dr. Kunxin Luo, Whitehead Institute, and Dr. Anthony J. Koleske, MIT) were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, glutamine, and antibiotics, as described.
Construction of Recombinant Baculovirus-expressed c-Src and v-Src-Baculovirus constructs encoding either c-Src or v-Src (from chicken) were the generous gifts of Drs. Zhou Songyang and Lewis Cantley (Beth Israel Hospital/Harvard Medical School) and were constructed as described previously (33). Briefly, the cDNA encoding pp60 c-src was subcloned into a transfer plasmid vector pAC373 (34) using the BamHI site. A mixture of 2 g of recombinant plasmid pAC373(c-Src) DNA and 1 g of purified wild-type baculoviral DNA was transfected into insect Sf9 cells. Four days later, culture supernatants were removed and centrifuged at 1,000 rpm for 10 min. Clarified supernatants containing wild-type and recombinant viruses were plaque assayed on a monolayer of Sf9 cells. Occlusion negative plaques were picked and seeded onto 2.5 ϫ 10 6 cells. After 3 days of incubation, cells and culture supernatants were removed and centrifuged at 1,000 rpm for 10 min. The cell pellets were analyzed by both in vitro kinase assay and immunoblotting using anti-Src antibody. Those plaques testing positive for the presence of kinase-active pp60 c-src were selected for three rounds of plaque purification (35) The selected plaques were used as virus stock for producing c-Src protein by infecting insect Sf21 cells. Similarly, v-Src was constructed as a recombinant baculovirus (33).
Interaction of Recombinant Baculovirus-expressed Src with GST-Caveolin Fusion Proteins-The interaction of GST-caveolin fusion proteins with baculovirus-expressed Src was evaluated essentially as we described for the interaction of caveolin with baculovirus-expressed heterotrimeric G-protein ␣subunits (27) and H-Ras (19). Briefly, GST or GST-caveolin fusion proteins bound on glutathione-agarose beads were extensively washed first with PBS (1ϫ) and lysis buffer containing protease inhibitors (3ϫ). These beads contained ϳ100 pmol of a given fusion protein/100 l of packed volume. Approximately 100 l of this material was incubated with 1 ml of precleared Src lysates by rotating overnight at 4°C. After binding, the beads were extensively washed (6 -8ϫ) with wash buffer containing 50 mM Hepes, pH 7.5, 120 mM NaCl, 1 mM EDTA, 0.5% CHAPS, and protease inhibitors. Finally, associated proteins were eluted with 100 l of elution buffer containing 50 mM Tris, pH 8.0, 1 mM EDTA, 1% Triton X-100, 10 mM reduced glutathione, and protease inhibitors. The eluate was mixed 1:1 with 2 ϫ sample buffer and subjected to SDS-PAGE (10% acrylamide). After transfer to nitrocellulose, Western blot analysis was performed with an anti-Src mAb probe (1:1000 dilution, Oncogene Sciences). Horseradish peroxidase-conjugated secondary antibodies (1:5000 dilution, Amersham Corp.) were used to visualized bound primary antibodies by an enhanced chemiluminescence assay (ECL) (Amersham Corp.).
Caveolin-derived Synthetic Peptides-Caveolin peptides were designed based on the protein sequence of the N-terminal domain of canine caveolin. Peptides were synthesized using standard methodology and subjected to amino acid analysis and mass spectroscopy (Massachusetts Intitute of Technology Biopolymers Laboratory) to confirm their composition. The following four peptides were utilized: peptide 1 (NRDPKHLNDDVVKIDFEDVIAEPEGTHSF, caveolin residues 53-81); peptide 2 (DGIWKASFTTFTVTKYWFYR, caveolin residues 82-101); peptide 3 (IWKASFTTF, caveolin residues 84 -92); and peptide 4 (TVTKYWFYR, caveolin residues 93-101). Note that peptides 3 and 4 correspond to the N-terminal and C-terminal halves of peptide 2, respectively.
In Vitro Auto-phosphorylation of Src Family Tyrosine Kinases-Two to six units of either purified recombinant c-Src or Fyn (44 -132 ng/ml; Upstate Biotechnology, Inc) were incubated with caveolin peptides at concentrations of 0, 0.3, 1, 3, and 10 M in the kinase reaction buffer. Reactions were performed in a total volume of 50 l of kinase reaction buffer (20 mM Hepes, pH 7.4, 5 mM MgCl 2 , 1 mM MnCl 2 ). The reaction was initiated by addition of 15 Ci of [␥-32 P]ATP. After incubation for 15 min at 25°C, the reaction was stopped by addition of 2 ϫ SDS-PAGE sample buffer and boiling for 2 min. Phosphorylated proteins were visualized by autoradiography using an intensifying screen. Control samples omitting either either [␥-32 P]ATP or Src family tyrosine kinases showed no activity. The activity of purified c-Src and Fyn was defined by the manufacturer (UBI, Inc.) as follows: in kinase assays that use only [␥-32 P]ATP one enzyme unit transfers 0.1 nmol of 32 P/min to a peptide substrate, i.e. cdc2 (residues 6 -20). Also, purified c-Src had a specific activity of 900,000 units/mg.
Co-expression of c-Src and Caveolin in Mammalian 293T Cells-Untagged caveolin and c-Src were co-expressed in 293T cells by cotransfection using a modified calcium phosphate precipitation procedure as described previously (19). Briefly, 293T cells were plated in 10-cm culture dishes at ϳ1 ϫ 10 6 cell/dish in Dulbecco's modified Eagle's medium containing 10% fetal calf serum (complete medium) and cultured until cells reached ϳ80% confluency. Just prior to transfection, the medium was removed and replaced with 4 ml of the fresh complete medium. For each 10-cm dish, the transfection mixture was prepared by adding 6 g of either c-Src and caveolin plasmid DNA (10 g of DNA for single transfection) to sterile H 2 O such that the final volume is 438 l. After addition of 62 l of 2 M CaCl 2 to the DNA/H 2 O solution, 500 l of 2 ϫ Hepes-buffered saline (pH 7.05) was added dropwise to the mixture with gentle agitation. Within 1-2 min, this mixture was added to the cells; dishes were gently agitated to ensure uniform mixing. 10 h post-transfection, the culture medium was replaced with fresh complete medium. 48 h post-tranfection, 293T cells were harvested and used for immune complex kinase assays. For transient expression, caveolin was subcloned into the MCS (HindIII-BamHI) of the vector pCB7 containing the hygro R marker (The vector was a gift from Dr. James E. Casanova, Massachusetts General Hospital) (24). A construct encoding the chicken c-Src cDNA for mammalian cell expression (pMHHB5) was a gift from Dr. David Shalloway (Cornell University).
Immunoprecipitation and Western Blotting of Caveolin and Src-48 h post-transfection, one or two 10-cm dishes of 293T cells (ϳ1-2 ϫ 10 6 cells/dish) co-transfected with c-Src and caveolin were washed three times with PBS (1ϫ), collected, and incubated in lysis buffer (10 mM Tris, pH 7.5, 50 mM NaCl, 1% Triton X-100, 60 mM octyl glucoside) for 30 min at 4°C. Lysates were clarified by centrifugation at 15,000 ϫ g for 15 min and precleared by incubation with 0.1% albumin-coated protein A-Sepharose for 6 h at 4°C. After preclearing, supernatants were transferred to tubes containing mouse monoclonal anti-Src IgG prebound to protein A-Sepharose. After incubation rotating overnight at 4°C, immunoprecipitates were washed three times with lysis buffer and divided into three equal aliquots. One aliquot was subjected to an immune complex kinase assay (see below). The two other aliquots were subjected to immunoblot analysis with either a c-Src mAb probe or a caveolin mAb probe (clone 2297).
Immune Complex Kinase Assays-Anti-Src-immunoprecipitates from 293T cells co-transfected with c-Src and caveolin were washed twice with kinase reaction buffer (20 mM Hepes, pH 7.4, 5 mM MgCl 2 , 1 mM MnCl 2 ). The kinase reaction was initiated by the addition of 2 Ci of [␥-32 P]ATP. Kinase reactions were performed in a total of 50 l at 25°C. After incubation for 10 min at 25°C, immunoprecipitates were washed twice with lysis buffer at 4°C and eluted by addition of 2ϫ SDS-PAGE sample buffer and boiling for 2 min. Proteins were separated by SDS-PAGE and visualized by autoradiography with an intensifying screen.
Scanning Densitometry-Quantitation was performed as detailed previously (36). Briefly, phosphorylated protein bands were digitized by high resolution optical scanning; volumetric integration of signal intensities was carried out using ImageQuant Software (Fast Scan; Molecular Dynamics).
Immunofluorescence-All reactions were performed at room temperature. Transfected 293T cells were briefly washed three times with PBS and fixed for 20 min in PBS containing 4% paraformaldehyde. Fixed cells were rinsed with PBS and treated with 25 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 and washed with PBS three times for 10 min each time. For double labeling, the cells were then successively incubated with PBS/3% bovine serum albumin containing: (i) 50 g/ml each of normal goat and donkey IgGs; (ii) a 1:400 dilution of mAb anti-c-Src and 40 g/ml anti-caveolin polyclonal IgG; and (iii) LRSC (lissamine rhodamine B sulfonyl chloride)-conjugated goat anti-mouse antibody (5 g/ml) and fluorescein isothiocyanate-conjugated donkey anti-rabbit antibody (5 g/ml). The first incubation was 30 min, whereas primary and secondary antibody reactions were 60 min each. Cells were washed three time with PBS between incubations. Slides were mounted with Slow-Fade anti-fade reagent and observed under a Bio-Rad MR600 confocal fluorescence microscope.
To test this hypothesis, we assessed the interaction of caveolin with Src family tyrosine kinases in vitro using (i) baculovirus-expressed Src proteins (c-Src and v-Src) and (ii) bacterially expressed caveolin fusion proteins. As shown in Fig. 1, both c-Src and v-Src were expressed well in baculovirus-infected Sf21 cells, and both migrated as a ϳ60-kDa band as expected. It should be noted that this band was not detectable in uninfected cells or in cells infected with vector alone (not shown). GST-caveolin fusion proteins were expressed and purified as we described previously (24,25,27).
As a first step, we examined the interaction of Src-tyrosine kinases with full-length caveolin affixed to glutathione-agarose beads (GST-FL-Cav, residues 1-178); GST alone served as a control for nonspecific binding. After binding, extensive washing, and elution with reduced glutathione, bound Src was visualized by immunoblot analysis. Fig. 1B shows that c-Src bound specifically to full-length caveolin but not to GST alone. In addition, v-Src failed to form a stable complex with caveolin under these binding conditions. This may reflect differences in the protein sequences of these two Src kinase, which diverge most extensively at their extreme C termini, resulting in the constitutive auto-activation of v-Src.
In order to define a minimal region of caveolin that can functionally support an association with c-Src, GST-caveolin fusion proteins encoding distinct portions of the N-or C-terminal domains of caveolin were used as the substrate for c-Src binding. Fig. 2 shows that only the GST fusion bearing caveolin residues 61-101 retained c-Src binding activity. However, similar fusion containing caveolin residues 1-81 did not show any interaction. This suggests that caveolin residues 82-101 are most critical for interaction with c-Src. This 20-amino acid stretch of caveolin residues is located within a membraneproximal region of the cytosolic N-terminal domain of caveolin. Similarly, this region of caveolin (residues 82-101) has been previously implicated in the direct interaction of caveolin with G ␣ subunits and H-Ras (19,27).
Caveolin Residues 82-101 Functionally Inhibit the Auto-ac-  (Table I). Using purified recombinant c-Src tyrosine kinase, we evaluated the effects of these caveolin peptides on the functional auto-activation of c-Src kinase in vitro.
Auto-activation of c-Src occurs through auto-phosphorylation of tyrosine 416. Tyrosine 416 is located within a conserved "activation loop" that is located within the kinase domain of many distinct families of protein kinases (37,38). Tyrosine 416 is not phosphorylated when c-Src is inactive and is phosphorylated when c-Src is active (38,39). In addition, the Y416F mutation in activated forms of c-Src is sufficient to inhibit their transforming ability (40 -43).
Tyrosine 416 is also the major site of auto-phosphorylation within c-Src in vitro (39,44). Thus, in vitro auto-phosphorylation of c-Src serves as a measure of its auto-activation. Fig. 3 illustrates the effect of these caveolin peptides on the functional auto-activation of purified recombinant c-Src kinase. Peptide 1 (Fig. 3A) had no effect at any concentration, whereas peptide 2 (Fig. 3B) dose-dependently suppressed the auto-activation of c-Src kinase. Approximately 85% inhibition was observed at a peptide concentration of 300 nM and was completely abolished at 3 M. It should be noted that peptide 2 corresponds to the caveolin region (residues 82-101) that we have implicated in the direct interaction of caveolin with c-Src using caveolin expressed as a GST fusion protein (see Fig. 2 above).
Interestingly, when peptide 2 was divided into two peptides (peptides 3 and 4; Fig. 3C), no inhibition of c-Src auto-activation was observed, even at a peptide concentration of 10 M. Quantitation of these experiments (Fig. 3, A-C) is presented in Fig. 3D. These observations suggest that the intact domain (caveolin residues 82-101) is required for both structural and functional interaction with c-Src tyrosine kinase.
To test the generality of this phenomenon, we next assessed the effect of caveolin peptides on the auto-phosphorylation of a related Src family tyrosine kinase, Fyn. Fig. 4 shows that similar results were obtained with Fyn kinase. The inhibitory effect of peptide 2 was slightly less potent on Fyn tyrosine kinase. Approximately 90% inhibition was observed at a peptide concentration of 3 M; this is about 10-fold less potent than the effect we observed with c-Src. As with c-Src, peptide 1 had no effect on Fyn auto-phosphorylation, even at a peptide concentration of 10 M. The observed differences in the potency of peptide 2 with different Src family tyrosine kinases suggests that (i) the interaction of caveolin with Src family tyrosine kinases is highly specific and (ii) that caveolin may preferentially interact with a select subset of Src family members in vivo.
Co-expression with Caveolin Suppresses the Auto-activation of c-Src in Vivo-The experiments described above indicate that caveolin residues 82-101 interact directly with c-Src and functionally suppress auto-activation of c-Src in vitro. However, it is not clear from our experiments whether full-length caveolin can perform this function in vivo.
To test this idea under more physiological conditions, we transiently co-expressed c-Src with full-length caveolin in 293T cells. These cells express extremely low levels of endogenous caveolin; little or no immunoreactivity was observed with anticaveolin IgG by Western blotting (not shown). After expression, c-Src kinase was isolated by immunoprecipitation and subjected to the immune complex kinase assay (Fig. 5). Although immunoblot analysis indicates that these immunoprecipitates contain equivalent amounts of c-Src kinase, little or no kinase activity was associated with c-Src when co-expressed with caveolin in vivo. Thus, co-expression of caveolin with c-Src can abolish its capacity for auto-phosphorylation, as predicted based on our experiments with caveolin peptides (see Fig. 3).

FIG. 2. Defining a region of caveolin that contains c-Src binding activity.
A, diagrammatic summary of each GST-caveolin fusion protein relative to a complete caveolin molecule. The numbers at end points reflect their exact amino acid position within caveolin. These fusion proteins correspond to the full-length caveolin protein (residues 1-178), the C-terminal caveolin domain (residues 135-178), and various portions of the N-terminal domain of caveolin (residues 1-21, 1-41, 1-61, 1-81, or 61-101). These GST-caveolin fusion proteins were constructed and characterized as we described previously (24,25,27). B, molecular mapping of a caveolin region that interacts with c-Src. Using a panel of GST-caveolin fusion proteins enumerated in A, we systematically identified a 41-amino acid region of caveolin (residues 61-101) that is functionally sufficient to interact with c-Src. After binding, extensive washing, and elution with reduced glutathione, bound Src was visualized by immunoblot analysis as in Fig. 1B. Detergent extracts of insect cells recombinantly expressing c-Src were prepared from from a total of four t25 flasks containing ϳ1 ϫ 10 6 cells each. Equivalent amounts of GST and GST-caveolin fusion proteins were used in these binding experiments. These results imply that the interaction of caveolin with Src family kinases may serve to negatively regulate their functional activity. In addition, double labeling of 293T cells co-transfected with c-Src and caveolin revealed significant co-localization of these two distinct gene products (Fig. 6). This is consistent with results demonstrating co-immunoprecipitation of caveolin with c-Src using antibodies directed against c-Src (Fig. 5A, bottom  panel).
Is Tyrosine Phosphorylation of Caveolin Required for the

Caveolin-mediated Inhibition of Src Family Kinases?-We
do not yet know the mechanism by which caveolin-derived peptides inhibit Src family kinases. One possibility is that this caveolin-mediated inhibition is related to tyrosine phosphorylation of caveolin or caveolin-derived peptides, resulting in a form of competitive substrate inhibition of Src family kinases. In support of this possibility, peptide 2 (DGIWKASFTTVTKY-WFYR; residues 82-101) contains both inhibitory activity and two tyrosine residues.
To directly examine the possible requirement for tyrosine  Table I, were examined for their effect on the autophosphorylation of purified recombinant c-Src kinase in vitro. The effects of peptide 1 (A), peptide 2 (B), and peptides 3 and 4 (C) are shown. Note that peptide 1 had no effect, whereas peptide 2 dose-dependently suppressed the auto-phosphorylation of c-Src. Peptide 2 encodes residues 82-101 of caveolin that contain caveolin binding activity as we have shown in Fig. 2B using GST-caveolin fusion proteins. Peptides 3 and 4 correspond to the N-terminal and C-terminal halves of peptide 2. Note that peptides 3 and 4 have no effect, indicating that the complete 82-101 region of caveolin (peptide 2) is required for inhibiting the auto-activation of c-Src. Quantitation of these experiments is provided in D. Cumulative data are shown as the means Ϯ S.D. These experiments were performed at least three times independently in duplicate. phosphorylation in this event, we generated two mutated caveolin peptides in which both of these tyrosine residues were changed to alanine (DGIWKASFTTVTKAWFAR; termed Y 3 A) or phenylalanine (DGIWKASFTTVTKFWFFR; termed Y 3 F). Fig. 7 shows that both mutated caveolin peptides lacking tyrosine (Y 3 A; Y 3 F) were as effective or more effective than the wild-type peptide sequence in suppressing the auto-phosphorylation of c-Src. In addition, the mutant Y 3 F peptide was approximately twice as potent as the wild-type caveolin peptide. These results demonstrate that tyrosine phosphorylation of this caveolin sequence (residues 82-101) is not required for its inhibitory activity toward c-Src.
In accordance with these results, the Src-binding region we have defined here within caveolin (caveolin residues 61-101) is not a substrate for tyrosine phosphorylation by c-Src, as we have published previously (see Fig. 2 within Ref. 45). Also, the site of caveolin phosphorylation by c-Src occurs at a single tyrosine residue (tyrosine 14; Ref. 45), and this residue is outside caveolin's Src-binding region (caveolin residues 61-101). Thus, these experimental observations also argue against competitive substrate inhibition via tyrosine phosphorylation.
What about the Functional Activity of Other Caveolin Family   FIG. 4. Effects of caveolin peptides on the auto-activation of Fyn, a closely related Src family tyrosine kinase. Caveolin peptides were examined for their effect on the autophosphorylation of purified recombinant Fyn kinase in vitro. The effects of peptide 1 (A) and peptide 2 (B) are shown. Note that peptide 1 had no effect, whereas peptide 2 dose-dependently suppressed the auto-activation of Fyn. Quantitation of these experiments is provided in C. Cumulative data are shown as the means Ϯ S.D. These experiments were performed at least three times independently in duplicate.

FIG. 5. In vivo effect of the full-length caveolin molecule on the auto-activation of c-Src tyrosine kinase revealed by co-expression of caveolin and c-Src in 293T cells.
A, 293T cells (ϳ1-2 ϫ 10 6 cells/10-cm dish) were transfected with c-Src alone or co-transfected with c-Src plus caveolin. 48 h post-transfection, cells were washed and collected in lysis buffer. Cell lysates were immunoprecipitated with anti-Src IgG bound to protein A-Sepharose. These immunoprecipitates were subjected to immunoblot analysis with a Src mAb probe (top panel), an immune complex kinase assay to detect Src auto-phosphorylation (middle panel), and immunoblot analysis with a caveolin mAb probe (bottom panel). Note that although both immunoprecipitates contain equivalent amounts of c-Src (top panel), co-expression with caveolin prevents the auto-phosphorylation of c-Src (middle panel). Also, caveolin co-immunoprecipitates with c-Src when using antibodies directed against c-Src (bottom panel). One 10-cm dish was used per immunoprecipitation. B, quantitation of A (middle panel). The autophosphorylation of c-Src tyrosine kinase is expressed in arbitrary units.
Members?-Recently, two novel caveolin-related proteins have been identified and cloned. These proteins, termed caveolin-2 and caveolin-3, are the products of separate caveolin genes (46 -48). Thus, caveolin (also known as caveolin-1) is the first member of a multi-gene family (46).
Caveolins-1, -2, and -3 are structurally homologous proteins but are immunologically distinct molecules; they have different but overlapping tissue distributions (20, 46 -48). For example, the expression of caveolin-3 is absolutely muscle-specific (skeletal and cardiac muscle cells) (20,47). Caveolin-1 is not expressed in these striated muscle tissues, but smooth muscle cells co-express caveolins-1 and -3 (20). Furthermore, caveolin-1 and -2 are co-expressed in adipocytes and share the same overlapping tissue distribution (46). Thus, a given mammalian cell, such as smooth muscle cells or fibroblasts, may co-express up to three or four immunologically distinct caveolin protein products.
As a consequence of these findings, we also evaluated the effects of peptides derived from caveolins-2 and -3 that correspond to the 82-101 region in caveolin-1. Peptides derived from all three caveolins contain two conserved tyrosine residues (marked by arrows in Fig. 8A) and are extremely homologous. Fig. 8 shows that only the peptides derived from caveolins-1 and -3 exhibited inhibitory effects; the caveolin-2-derived peptide had no inhibitory effect.
This differential effect was also observed when these three peptides were used to evaluate their effects on the functional activity of hetero-trimeric G-proteins. More specifically, peptides derived from caveolins-1 and -3 both dose-dependently suppressed the GTPase activity of purified heterotrimeric Gproteins, whereas the caveolin-2-derived peptide did not possess this inhibitory activity (27,46,47). Thus, our previous results with G-proteins directly parallel our current results with Src family tyrosine kinases. DISCUSSION Protein-tyrosine kinases play an essential role in the regulation of cellular growth control and differentiation. Two general classes of tyrosine kinases have been defined: transmembrane receptors and nonreceptor cytoplasmically oriented kinases (49). Receptor tyrosine kinases are activated by the binding of ligand to the extracellular protein domain, whereas non-receptor tyrosine kinases must be activated indirectly. Both classes of tyrosine kinases can function as transforming oncogenes when mutationally activated (50,51).
c-Src is the prototype of a family of non-receptor tyrosine kinases. v-Src, the transforming component of the Rous sarcoma virus, is derived from c-Src, a normal cellular gene. These two gene products differ primarily in their extreme C-terminal regions (38,52,53). Mutationally activated c-Src has been implicated in the pathogenesis of carcinomas, including breast tumors (54). In contrast, recombinant overexpression of wildtype c-Src does not result in efficient cell transformation or increased tyrosine phosphorylation, indicating that the tyrosine kinase activity of c-Src is normally repressed (37,55). However, the tightly regulated tyrosine kinase activity of c-Src can be activated transiently by a number of growth factors (38,56).
To date, a total of nine distinct members of the Src family of tyrosine kinases have been identified and cloned (57). c-Src and all other family members contain five conserved functional domains (38). (i) The extreme N terminus contains a consensus sequence for lipid modification by addition of myrisate, palmitate, or both fatty acyl moieties (58 -60). (ii) and (iii) The SH2 and SH3 domains function in protein-protein interactions. The SH2 domain recognizes specific phosphotyrosine based motifs (61), whereas the SH3 domain binds proline-rich peptide sequences with the consensus PXXP (62)(63)(64). (iv) The SH1 domain (or catalytic domain) functions in the recognition and tyrosine phosphorylation of specific substrates (65). (v) Finally, the extreme C terminus contains a site of negative regulation, Tyr 527 (37,55). As a consequence of C-terminal divergence, v-Src does not contain Tyr 527 , which when phosphorylated can act to functionally inactivate c-Src kinase (38).
It is now well established that tyrosine phosphorylation of c-Src itself plays a major role in controlling its intrinsic kinase activity. Autophosphorylation of Tyr 416 located within the "activation domain" of the kinase domain results in activation of the enzyme, whereas phosphorylation of Tyr 527 results in kinase inactivation (38,66). Phosphorylation of Tyr 527 is thought to be mediated by a C-terminal Src kinase (Csk) (67)(68)(69)(70). After phosphorylation of Tyr 527 , c-Src is thought to "fold up" by the intramolecular interaction of Tyr 527 with the SH2-domain, resulting in an inactive enzyme conformation (37, although this model remains speculative. Because elevated Src tyrosine kinase activity is associated with cellular transformation, many investigators have FIG. 6. Localization of c-Src and caveolin within a single cell. 293T cells were co-transfected with c-Src and untagged mammalian caveolin. c-Src expression was detected with a mouse mAb that recognizes a conserved N-terminal epitope; caveolin expression was detected using rabbit polyclonal IgG directed against caveolin. Control experiments using singly transfected populations of cells confirmed the specificity of these antibodies; no cross-reaction was observed (not shown). Transfected cells expressing both c-Src and caveolin were selected for imaging by laser confocal fluorescence microscopy. Primary antibodies were detected using distinctly tagged fluorescent secondary antibodies: A, rhodamineconjugated for c-Src; B, fluorescein-conjugated for caveolin; and C, superposition of the fluorescent images of A and B. Note that co-localization of c-Src and caveolin generates a yellow fluorescent image.
searched for mechanisms to inactivate the kinase activity of c-Src. For example, Shoelson and colleagues have shown that high concentrations of peptide substrates inhibited the autophosphorylation of c-Src kinase (71). Similarly, Shalloway and colleagues found that auto-phosphorylation of the c-Src is an intermolecular process using phosphopeptides as a probe (72). However, in both reports, the investigators used much higher concentrations of peptides, such as 0.24 mM or even 8.5 mM. In contrast, here we have found that only 300 nM of caveolin peptide 2 (caveolin residues 81-101) strongly inhibited c-Src auto-phosphorylation; similar results were obtained with Fyn, a related Src-like kinase. Thus, our current studies imply that caveolin is a potent inhibitor of c-Src and perhaps other Src family tyrosine kinases. In addition, our results indicate that caveolin preferentially interacts with c-Src rather than v-Src. These observations suggest that caveolin better recognizes the inactive conformation of c-Src, and thus, c-Src tyrosine kinase activity could be normally repressed by an interaction with caveolin.
How does caveolin binding negatively regulate c-Src autoactivation? The functional existence of such a negative regulator or suppressor of c-Src has been suggested by earlier studies (37). Because mutations in the extreme C terminus, the kinase domain, the SH2 domain, or the SH3 domain can constitutively activate c-Src, this activation may be mediated a conformational change in one of these regions (38). Such a conformational change could also be induced normally by unknown allosteric effectors. Putative allosteric activators could stabilize c-Src in the activated state, whereas putative allosteric inhibitors could hold c-Src in an inactive conformation (37). Regarding allosteric inhibitors, they could act by shifting the equilibrium in favor of the inactive conformation, perhaps by masking Tyr 416 from kinases or Tyr 527 from phosphatases (37,38). Be-cause caveolin preferentially interacts with inactive c-Src (but not v-Src) and a caveolin-derived peptide or caveolin co-expression can inhibit auto-activation of c-Src, caveolin could act as an allosteric inhibitor of c-Src and other Src family kinases.
The subcellular distribution of several signaling molecules is restricted by association with scaffolding proteins (28). These scaffolding proteins, Ste5p (73,74), AKAP-79 (29,30,75,76), and 14 -3-3 (77-80) may simultaneously associate with distinct classes of signaling proteins to form a signaling pathway or module. So far, the accumulated evidence suggests that caveolin possesses all the qualities of a "classic" scaffolding protein because caveolin forms multivalent homo-oligomers and each caveolin-interacting protein binds to the same cytosolic membrane-proximal region of caveolin. As such, caveolin may provide a selective framework that segregates one group of signaling events from the next by preventing cross-talk between functionally unrelated signaling modules while facilitating cross-talk between related signaling modules.
Because a discrete cytosolic domain of caveolin (residues 82-101) (i) directly interacts with wild-type G ␣ subunits, H-Ras and c-Src; (ii) fails to interact with mutationally activated G ␣ subunits (G s ; Q227L), H-Ras (G12V), and Src (v-Src); (iii) functionally suppresses the activity of both G-proteins and Src family tyrosine kinases, holding them in the inactive conformation; and (iv) is involved in the formation of multivalent caveolin homo-oligomers, we propose the term caveolin-scaffolding domain be used when describing this region of caveolin. FIG. 7. Effects of mutated caveolin peptides on the auto-activation of c-Src tyrosine kinase. Wild-type and mutant caveolin peptides detailed in A were examined for their effect on the autophosphorylation of purified recombinant c-Src kinase in vitro. The effects of these peptides are shown in B. All peptides were added at a concentration of 1 M. Note that both mutated caveolin peptides lacking tyrosine residues (Y3 A and Y3 F) were as effective or more effective than the wild-type (WT; peptide 2, residues 82-101) peptide sequence in suppressing the auto-phosphorylation of c-Src. The Y 3 F peptide was approximately twice as potent as the wild-type caveolin peptide.
FIG. 8. Differential effect of peptides derived from caveolins-1, -2, and -3 on the auto-activation of c-Src tyrosine kinase. Recently, two caveolin-related proteins (caveolin-2 and caveolin-3) have been identified and cloned; caveolin has been retermed caveolin-1. As a consequence, we also evaluated the effects of peptides derived from caveolins-2 and -3 that correspond to the 82-101 region in caveolin-1; the sequences of these peptides are detailed in A. All three peptides contain two conserved tyrosine residues (marked by arrows) and are extremely homologous. The effects of these peptides are shown in B. All peptides were added at a concentration of 1 M. Note that only peptides derived from caveolins-1 and -3 exhibited inhibitory effects; the caveolin-2-derived peptide had no inhibitory effect. dren's Hospital/Harvard Medical School) for critical reading of the manuscript.