Caveolin Interacts with Trk A and p75NTR and Regulates Neurotrophin Signaling Pathways*

Neurotrophins signal through Trk tyrosine kinase receptors and the low-affinity neurotrophin receptor p75NTR. We have shown previously that activation of Trk A tyrosine kinase activity can inhibit p75NTR-dependent sphingomyelin hydrolysis, that caveolae are a localized site for p75NTR signaling, and that caveolin can directly interact with p75NTR. The ability of caveolin to also interact with tyrosine kinase receptors and inhibit their activity led us to hypothesize that caveolin expression may modulate interactions between neurotrophin signaling pathways. PC12 cells were transfected with caveolin that was expressed efficiently and targeted to the appropriate membrane domains. Upon exposure to nerve growth factor (NGF), caveolin-PC12 cells were unable to develop extensive neuritic processes. Caveolin expression in PC12 cells was found to diminish the magnitude and duration of Trk A activationin vivo. This inhibition may be due to a direct interaction of caveolin with Trk A, because Trk A co-immunoprecipitated with caveolin from Cav-Trk A-PC12 cells, and a glutathioneS-transferase-caveolin fusion protein bound to Trk A and inhibited NGF-induced autophosphorylation in vitro. Furthermore, the in vivo kinetics of the inhibition of Trk A tyrosine kinase activity by caveolin expression correlated with an increased ability of NGF to induce sphingomyelin hydrolysis through p75NTR. In summary, our results suggest that the interaction of caveolin with neurotrophin receptors may have functional consequences in regulating signaling through p75NTR and Trk A in neuronal and glial cell populations.

Neurotrophins are a family of growth factors that mediate the survival, development, and death of specific populations of neurons and glial cells (1). Many of the classic trophic signals elicited by neurotrophins (nerve growth factor (NGF), 1 brainderived neurotrophic factor, and neurotrophin-3) require the presence of a specific member of the Trk tyrosine kinase receptor family (Trk A, Trk B, and Trk C, respectively (2)(3)(4)(5)(6)(7)(8)). After neurotrophin binding, Trk receptors undergo dimerization and transphosphorylation on specific tyrosine residues, a requisite step for activation of their intrinsic tyrosine kinase activity and coupling to downstream signaling pathways (9,10). In contrast, neurotrophin-induced death signals (11,12) may be mediated in specific cell types through the interaction of NGF with a transmembrane protein known as the low-affinity neurotrophin receptor, p75 NTR (13,14). The pro-apoptotic effect of p75 NTR may be mediated by generation of the bioactive lipid metabolite ceramide (11) that derives from the hydrolysis of sphingomyelin (SM) (15)(16)(17).
An emerging theme in neurotrophin signaling is that reciprocal interactions between Trk and p75 NTR signaling pathways can dictate cellular responses to neurotrophins (14). Although it is well appreciated that p75 NTR can regulate Trk-dependent responses (14, 18 -20), the role of Trk in regulating p75 NTR -dependent signaling is not as well documented; nonetheless, it is supported by both biochemical and biologic observations. Although brain-derived neurotrophic factor and neurotrophin-3 effectively induced SM hydrolysis in PC12 cells, which do not express endogenous Trk B or Trk C receptors, NGF was ineffective (16). Significantly, the inhibition of Trk A tyrosine kinase activity with K252a enabled NGF to stimulate SM hydrolysis. These studies provided the first biochemical clue that Trk A activation can inhibit some aspects of p75 NTR signaling. Importantly, this biochemical observation has functional correlates relevant to the effect of neurotrophins on cell survival. An elegant study by Bamji et al. (21) demonstrated clearly that Trk A activation inhibited brain-derived neurotrophic factorinduced p75 NTR -dependent death of sympathetic neurons. Van der Zee et al. (22) reported that expression of Trk A was required to counteract the death-inducing effects of p75 NTR in medial septum and neostriatal cholinergic neurons. Lastly, retroviral expression of Trk A in mature oligodendrocytes inhibited NGF-induced p75 NTR -dependent apoptosis and blocked the stimulation of stress-activated protein kinases (23). These observations suggest that endogenous factors that regulate Trk tyrosine kinase activity would be expected to modulate interactions between neurotrophin signaling pathways.
Caveolin-1 is a putative transformation suppressor protein that has emerged as a candidate molecule that may regulate the activity of cell growth pathways. Caveolin-1 is a member of a multigene family that includes caveolin-2 and caveolin-3 and is a key structural protein in the morphologic formation of caveolae (24,25); for simplicity, our use of caveolin will be synonymous with caveolin-1. Caveolae are invaginations of the plasma membrane that are involved in the cellular transport of small molecules and are also considered as potential localized centers for signal transduction (24), especially through tyrosine kinase receptors (26 -28), sphingolipid (17,29), and phosphati-dylinositol signaling pathways (30,31). Additionally, most if not all cells contain caveolae-related membrane domains (CRDs). CRDs share a similar lipid constitution with caveolae but lack caveolin.
Caveolin can associate with G-proteins (32), ras (33), protein kinase C isoforms (34), Src (35), nitric oxide synthase (36), and tyrosine kinase receptors (37,38) through direct protein-protein interactions. Importantly, the interaction of these proteins with caveolin has functional consequences on signaling because caveolin can inhibit the activity of nitric oxide synthase (36,39), protein kinase C isoforms (34), and tyrosine kinases (37). Therefore, we hypothesized that the interaction of caveolin with Trk may also have functional consequences on regulating cross-talk between the Trk A and p75 NTR signaling pathways. We demonstrate that overexpression of caveolin in PC12 cells prevents neurite outgrowth and blocks prolonged Trk A autophosphorylation in response to NGF. Trk A interacts specifically with caveolin, and this interaction dramatically inhibits NGF-induced autophosphorylation of Trk A. Moreover, similar to pharmacologic inhibition of Trk A (16), caveolin inhibition of Trk A signaling enables NGF to induce SM hydrolysis through p75 NTR . These data raise the possibility that caveolin or caveolin-related proteins may regulate signaling through Trk and p75 NTR receptors in specific neuronal populations.

EXPERIMENTAL PROCEDURES
Materials-Mouse 2.5S NGF was obtained from Harlan Bioproducts for Science. Polyclonal antibodies against antiphosphotyrosine and the C terminus of caveolin were purchased from Signal Transduction Laboratories. Polyclonal antibodies raised against the N terminus of caveolin and against the intracellular domain of Trk (pan-Trk and C-14) were obtained from Santa Cruz Biotechnology. Polyclonal antibodies against the Trk A extracellular domain were obtained from Upstate Biotechnology.
Cell Lines-Mock-transfected NIH-3T3 fibroblasts were maintained as described previously (17). 615-Trk A-PC12 cells were a kind gift from Dr. B. Hempstead (Cornell University, New York, NY) and were maintained in RPMI 1640 medium containing 10% horse serum, 5% fetal calf serum, and 0.2 mg/ml Geneticin. Wild type PC12 cells were grown in the same medium lacking Geneticin. PC12 and Trk A-PC12 cells were transfected with a C-terminal myc-tagged form of caveolin-1 (33) using LipofectAMINE. To obtain stable clones expressing caveolin, the transfected cells were grown in the presence of 0.5 mg/ml hygromycin B. Antibiotic-resistant colonies were isolated and expanded, and the expression of caveolin was determined by immunoblot analysis. Several PC12 colonies expressing caveolin (Cav-PC12) were isolated and maintained in RPMI 1640 medium containing 10% horse serum and 5% fetal calf serum with 0.2 mg/ml hygromycin B. Caveolin expression was found to slow cell growth and gradually diminish as the cells were passaged. In some experiments, Trk A-PC12 cells were transiently transfected with caveolin as described above and used 48 -72 h after transfection.
Isolation of Caveolae-enriched Membranes (CEMs)-CEMs were isolated by nondetergent extraction in sodium carbonate and centrifugation over discontinuous sucrose gradients (17,33) or by using the Opti-Prep method (40). After centrifugation in sucrose gradients, 15 ϫ 0.8 ml fractions were removed, starting from the top of the gradient. To concentrate the membranes within different regions of the gradient, CEMs (fractions 4 -7), NCM-1 (fractions 8 -11), and NCM-2 (fractions 12-15) were pooled, diluted into MBS (25 mM 2-(N-morpholino)ethanesulfonic acid, 150 mM NaCl, and 5 mM EDTA), and spun at 100,000 ϫ g for 30 min at 4°C. The membrane proteins were used for immunoprecipitation or analyzed directly by SDS-PAGE. After transferring the proteins to nitrocellulose, caveolin was detected by immunoblot analysis and identified by co-migration with caveolin from a human endothelial cell lysate.
Co-Immunoprecipitation of Trk A with Caveolin-One 15-cm dish of Trk A-PC12 cells or Cav-Trk A-PC12 cells was washed twice with ice-cold PBS, and the cells were resuspended in 0.3 ml of immunoprecipitation (IP) buffer (10 mM Tris-HCl, pH 7.4, 10 mM EDTA, 1% Nonidet P-40, 60 mM ␤-octylglucoside, 1 mM phenylmethylsulfonyl fluoride, and 10 g/ml each of leupeptin, aprotinin, and bestatin). The cells were broken by sonication with three 15-s bursts, and the lysate was centrifuged at 21,000 ϫ g for 15 min. Samples containing 1.5 mg of protein from the supernatant were precleared with protein A-Sepharose, the supernatant was incubated for 2 h at 4°C with 2 g of polyclonal caveolin antibody, and caveolin was immunoprecipitated with the addition of PAS beads. The PAS beads were washed three times with 1 ml of ice-cold MBST (MBS plus 1% Triton X-100), and the bound proteins were solubilized by boiling in sample buffer for 5 min. After SDS-PAGE, the proteins were transferred to nitrocellulose membranes, and the membranes were probed with a pan-Trk antibody. Immunoreactive proteins were visualized using a horseradish peroxidase-conjugated secondary antibody and enhanced chemiluminescence.
Association of GST-Caveolin Fusion Proteins with Trk A and p75 NTR -Characterization of the fusion proteins used in these experiments and the methods used for their expression and purification have been described previously (17,32,41). In some experiments, caveolin fusion proteins were isolated with agarose-conjugated glutathione beads, and caveolin was released from the GST affinity handle by thrombin cleavage. After cleavage, the agarose-conjugated GST was removed by centrifugation, and the concentration of the released recombinant caveolin present in the supernatant was determined using the Bio-Rad reagent with bovine serum albumin as the standard. To examine the interaction of p75 NTR with caveolin, an agarose-conjugated GST-p75 NTR cytoplasmic domain fusion protein was incubated with equal amounts of the indicated recombinant caveolin protein (freed from the affinity handle), and association was determined in the cosedimentation assay that we have described previously (17).
To analyze the association of Trk A with the GST-caveolin fusion proteins, CEMs, NCM-1, and NCM-2 membrane fractions from Trk A-PC12 cells were resuspended in binding buffer (17). Equal amounts of protein from each fraction were then incubated with the indicated amounts of agarose-conjugated GST or GST-caveolin. After rotating the samples for 2 h at 4°C, the beads were sedimented by a brief centrifugation and washed with 3 ϫ 1 ml of ice-cold MBST. The bound proteins were eluted with 30 l of freshly prepared 10 mM reduced glutathione (pH 8.0), and the eluate was boiled in sample buffer. The samples were analyzed for the presence of Trk A as described above.
NGF Response in PC12 and Cav-PC12 Cells-PC12 and Cav-PC12 cells were cultured on polylysine or collagen-coated 35-mm culture dishes and plated at a density of ϳ5 ϫ 10 5 cells/dish in RPMI 1640 medium-1% horse serum in the presence of 0 or 100 ng/ml NGF for 4 days. Fresh NGF was added after 2 days, and the cells were photographed on day 4.
Autophosphorylation of Trk A-PC12 and Cav-PC12 cells were grown on 10-cm dishes to 70% confluence. The cells were washed twice with ice-cold PBS and placed in 10 ml of 1% horse serum for 4 h before the addition of 100 ng/ml NGF. NGF was added for 0 -30 min, and Trk A was immunoprecipitated as described after the addition of ice-cold IP buffer (16). Equal amounts of protein were immunoprecipitated using 2 g of the polyclonal pan-Trk antibody. The immunoprecipitates were analyzed by immunoblotting with a horseradish peroxidase-conjugated antiphosphotyrosine antibody.
In Vitro Inhibition of Trk A Autophosphorylation by GST-Caveolin-Two 15-cm dishes of Trk A-PC12 cells were washed twice with ice-cold PBS and resuspended in 1 ml of ice-cold IP buffer. The cells were incubated on a rotating platform at 4°C for 20 min and sonicated for 45 s on ice, and the lysate was centrifuged at 21,000 ϫ g for 15 min at 4°C. The supernatant was precleared with PAS beads and subsequently incubated for 2 h at 4°C with 5 g of a polyclonal antibody against the Trk A extracellular domain. PAS beads were added to bind the Trk A antibody; after 1 h at 4°C, the PAS beads were sedimented by a brief centrifugation, washed three times with MBST, and washed three times with kinase buffer (10 mM Tris, pH 7.4, and 10 mM MnCl 2 ), and the beads were resuspended in a final volume of 400 l of kinase buffer. 0 -120 g of GST, GST-caveolin, or an equal volume of the glutathione elution buffer (10 mM glutathione in kinase buffer) were added to 100-l aliquots of the Trk A immunoprecipitate. This mixture was incubated at 4°C for 1 h to permit association of the proteins. The kinase reaction was initiated by the addition of 40 M ATP and 100 ng/ml NGF, and the reaction proceeded for 15 min at 37°C. The PAS beads were washed three times with ice-cold MBST and boiled in sample buffer. Immunoprecipitated proteins were resolved by SDS-PAGE, and after transferring to nitrocellulose, Trk A autophosphorylation was examined using an antiphosphotyrosine antibody.
Measurement of Sphingomyelin Levels in the CEM Pool-Sphingomyelin levels were measured in Cav-PC12 cells as described previously (17,42). The amount of SM hydrolyzed was normalized to the total protein present in the sample.

RESULTS
Caveolin Overexpression Inhibits NGF-induced Differentiation of PC12 Cells-To determine whether caveolin could modulate neurotrophin signaling, we stably transfected PC12 cells with caveolin-1 cDNA and examined the cells for caveolin expression using our standard fractionation and immunoprecipitation protocols. As controls, mock-transfected PC12 and NIH-3T3 fibroblast cells were also fractionated, and the three membrane pools were isolated. Similar to our previous results, caveolin was not detectable in the untransfected PC12 membranes using a polyclonal antibody raised against the C terminus of caveolin ( Fig. 1A) (17). However, significant amounts of caveolin were observed in the Cav-PC12 cells, and it had the same apparent molecular mass as caveolin from a human endothelial cell lysate prepared as a caveolin standard (Fig. 1B). Importantly, the distribution of caveolin in Cav-PC12 cells was similar to that seen in fibroblasts (Fig. 1C), which express caveolin endogenously. Indeed, there was little difference in the percentage distribution of caveolin between the membrane pools from NIH-3T3 fibroblasts and the Cav-PC12 cells (Fig.  1D). These results indicate that PC12 cells efficiently express caveolin and target it to the appropriate membrane domains.
Because the flotation of membranes through a 38% sucrose cushion is not an overly stringent criterion for the isolation of caveolar membranes, we also fractionated Cav-PC12 cells using a second detergent-free method as described by Smart et al. (40). Using this procedure, which may provide a more homogenous caveolae preparation, caveolin was readily detectable in Cav-PC12 cells and PCNA cells, which are an L1 fibroblast cell line used as a positive control (Fig. 1E). However, because we have not yet examined whether caveolin expression in PC12 cells results in the morphologic formation of caveolae, these domains will still be referred to as CRDs. Interestingly, we observed a small amount of caveolin immunoreactivity in wild type PC12 cell membranes isolated using the Opti-Prep protocol (40), but only when using a polyclonal antibody raised against the N terminus of caveolin (43). These results suggest that undifferentiated PC12 cells express low levels of caveolin endogenously. Densitometric analysis of these bands indicates that the Cav-PC12 cells express approximately 3-fold more caveolin than wild type cells.
PC12 cells are a well-accepted model for studying neurotrophin signaling, and NGF induces a well-characterized morphologic change of these cells to a sympathetic neuron-like phenotype (44). Interestingly, caveolin overexpression had a dramatic effect on NGF-induced differentiation of the transfected cells. After 4 days of exposure to 100 ng/ml NGF, wild type PC12 cells developed extensive neurites ( Fig. 2A). In contrast, the Cav-PC12 cells showed significantly fewer processes with greater than 95% of the cells showing no apparent neuritic response to NGF (Fig. 2B).
Caveolin Diminishes Trk Autophosphorylation-Trk A has been demonstrated to be critical to NGF-induced differentiation of PC12 cells (45), suggesting that the presence of caveolin disrupts the signaling pathway through Trk A leading to differentiation. Therefore, we examined the effect of caveolin on Trk A autophosphorylation. PC12 and Cav-PC12 cells were maintained in RPMI 1640 medium containing 1% horse serum for 4 h and then treated with 100 ng/ml NGF for 0 -30 min. The cells were scraped into lysis buffer, and equal amounts of total protein were immunoprecipitated using a polyclonal pan-Trk antibody. The immunoprecipitates were subjected to SDS-PAGE, and Trk autophosphorylation was determined by immunoblot analysis using an antiphosphotyrosine antibody. Both PC12 and Cav-PC12 cells showed little Trk autophosphorylation at time 0. However, Trk autophosphorylation significantly increased after exposure to NGF for 15 min in both cell types (Fig. 3). Interestingly, whereas PC12 cells continued to exhibit tyrosine phosphorylation of Trk A for up to 30 min after the addition of NGF, tyrosine phosphorylation of Trk A was rapidly abrogated after 15 min of NGF stimulation in Cav-PC12 cells. Importantly, the lack of Trk A autophosphorylation was not due to significant changes in the amount of Trk A immunoprecipitated at any time point. These observations suggest that the expression of caveolin in PC12 cells may inhibit Trk A autophosphorylation and disrupt the duration of signaling necessary to induce cell differentiation (10).
Trk A Associates with Caveolin-We next sought to determine whether caveolin may directly inhibit Trk tyrosine kinase activity. We hypothesized that if a direct inhibition occurred, then caveolin and Trk A may undergo a specific protein-protein interaction. To examine this interaction, Trk A-PC12 cells were transiently transfected with caveolin. We chose this approach because Trk A-PC12 cells express up to a 20-fold greater level of Trk relative to wild type cells (46). Cav-Trk A-PC12 cells, Trk A-PC12 cells, and fibroblasts were lysed in IP buffer, and the supernatants were immunoprecipitated using the polyclonal caveolin antibody raised against the C terminus of caveolin. The immunoprecipitates were washed with MBST and solubilized in sample buffer, and the proteins were resolved by SDS-PAGE. Immunoblot analysis revealed that there was a significant Trk A band that co-immunoprecipitated with caveolin from Cav-Trk A-PC12 cells (Fig. 4). In contrast, only minor amounts of Trk A were detected in the mock-transfected Trk A-PC12 cells. As expected, no Trk immunoreactivity was detected in immunoprecipitates from the fibroblasts. The co-immunoprecipitation of Trk from the mock-transfected Trk-PC12 cells may have been due to the immunoprecipitation of some endogenous caveolin by the C terminus caveolin antibody (Fig.  1E). However, because the immunoprecipitation of caveolin from any of our PC12 cell lines using this antibody gave very equivocal results, the co-immunoprecipitation of any Trk with the C terminus caveolin antibody raised the possibility that , or NIH-3T3 fibroblasts (C) were immunoprecipitated using a polyclonal antibody against the C terminus of caveolin. Immunoblot analysis was performed using a monoclonal caveolin antibody. The arrow indicates the migration of caveolin relative to caveolin from a human endothelial (HE) cell lysate. The results shown are from one representative clonal population of Cav-PC12 cells. D, the bands from each membrane pool in B and C were quantified by densitometry, and the intensity was expressed as a percentage of the total intensity of the three caveolin bands. E, caveolin was immunoprecipitated from the caveolin-enriched fraction of the second Opti-Prep gradient using an antibody against the N terminus of caveolin, and immunoblot analysis was performed with the same antibody.
Trk was interacting nonspecifically with the antibody and/or to the protein A-Sepharose beads.
To further examine the specificity of the interaction of caveolin with Trk A, CRDs, NCM-1, and NCM-2 fractions were isolated from wild type Trk A-PC12 cells (which had no exogenous caveolin expression) and incubated with increasing concentrations of an agarose-conjugated GST-caveolin fusion protein. This fusion protein expressed the full-length form of caveolin comprising amino acids 1-178 (17,47). The membrane fractions were incubated with 10, 30, and 60 g of an agaroseconjugated GST-caveolin fusion protein or agarose-conjugated GST. After the incubation, the beads were sedimented and extensively washed, and the bound proteins were eluted with glutathione. The eluates were subjected to SDS-PAGE, and the presence of Trk A was determined by immunoblot analysis. Fig.  5 shows that Trk A interacted specifically with the GST-caveolin fusion protein and did not associate with GST alone. The GST-caveolin fusion protein bound approximately 29% of the Trk A present in the CRDs from Trk A-PC12 cells and 20% of the Trk A from CRDs isolated from wild type PC12 cells (data not shown).
Interestingly, we have identified that an interaction of Trk with caveolin could only be demonstrated with the full-length caveolin and not with fusion proteins that span the oligomerization and scaffolding domain 1 of caveolin (Fig. 6). In contrast, the interaction of p75 NTR with caveolin is mediated through amino acids in the caveolin scaffolding domain 1 (amino acids 82-101) (Fig. 6). These results suggest that both neurotrophin receptors interact with caveolin, but that these interactions may be regulated by distinct domains of caveolin.
GST-Caveolin Prevents Trk A Autophosphorylation-The experiments described above support that there is a specific interaction between caveolin and Trk A. However, these results did not establish whether this interaction affected kinase activity. To more directly address this question, the effect of recombinant caveolin on the activation of Trk A was assessed in an in vitro kinase assay. Trk A was immunoprecipitated from a Trk A-PC12 lysate using an antibody directed against the extracellular domain of Trk A. After incubation with PAS for 1 h at 4°C, the immunoprecipitate was washed and resuspended in kinase buffer. The indicated amount of GST-caveolin or GST (both eluted from the beads) or an equal volume of glutathione elution buffer was added to the reaction, and the proteins were allowed to associate for 1 h at 4°C. The kinase reaction was then initiated with the addition of ATP and NGF. After incubation at 37°C, the PAS beads were sedimented and washed with MBST, and the bound proteins were solubilized with sample buffer. After SDS-PAGE, the activation of Trk A was evaluated using immunoblot analysis for antiphosphotyrosine. Fig. 7 shows that there was a marked activation of Trk A in samples treated with either GST or elution buffer alone, indicating that NGF was effectively inducing receptor autophosphorylation. In contrast, the addition of 60 or 120 g of GST-caveolin totally abolished autophosphorylation, strongly

FIG. 3. Trk A activation is inhibited in Cav-PC12 cells.
Mocktransfected PC12 cells and Cav-PC12 cells were grown to 70% confluence on 10-cm dishes. The cells were placed in RPMI 1640 medium containing 1% horse serum for 4 h and then treated for the indicated times with 100 ng/ml NGF. The cells were lysed in IP buffer and immunoprecipitated using an anti-pan-Trk antibody. Trk autophosphorylation was determined by immunoblot analysis using an antiphosphotyrosine antibody.

FIG. 4. Trk A co-immunoprecipitates with caveolin in Trk A-PC12 cells transfected with caveolin.
One 15-cm dish of 70% confluent mock-transfected fibroblasts (PCMV), Trk A-PC12 cells, or Cav-Trk-PC12 cells was immunoprecipitated using an anti-caveolin antibody raised against the C terminus of caveolin. Immunoblot analysis was performed using an anti-pan-Trk antibody. Data are representative of similar results obtained in at least three experiments.
supporting that caveolin was specifically inhibiting NGF-induced Trk autophosphorylation.
Caveolin Modulates Trk A Inhibition of NGF-induced Sphingomyelin Hydrolysis-We have observed previously that the activation of Trk A by NGF inhibits p75 NTR -dependent SM hydrolysis (16). Significantly, the inhibition of Trk activation with the relatively specific Trk A inhibitor K252a enabled NGF to signal through p75 NTR . Because caveolin was essentially mimicking the inhibitory effect of K252a on Trk A tyrosine kinase activity, this suggested that caveolin may also enable NGF to signal through p75 NTR in the Cav-PC12 cells. Cav-PC12 cells were labeled with [ 3 H]choline for 3 days and placed in RPMI 1640 medium containing 1% horse serum. After 4 h, the cells were treated with either 100 ng/ml NGF or vehicle for 0 -30 min. CRDs were isolated and analyzed for SM content. Interestingly, in the Cav-PC12 cells, NGF-induced SM hydrolysis was observed at nearly all time points except for the 15 min time point (Fig. 8), when Trk A phosphorylation was found to be maximal (compare time course with Fig. 3). Importantly, essentially no NGF-induced SM hydrolysis was observed in CRDs from wild type PC12 cells (16). These observations suggest that caveolin may affect cross-talk between Trk A and p75 NTR by regulating the extent of Trk activation. DISCUSSION We have determined that caveolin can interact with both p75 NTR and Trk A, and that this interaction may regulate neurotrophin responses. In this regard, Trk A co-immunoprecipitated with caveolin in Trk A-PC12 cells that were transfected with caveolin. Trk A also bound specifically and cosedimented with a GST-caveolin fusion protein expressing the full-length form of caveolin. Caveolin expression resulted in decreased ligand-induced activation of Trk A tyrosine kinase activity that markedly blunted the NGF-induced differentiation of PC12 cells transfected with caveolin. Furthermore, we have demonstrated that caveolin can directly block NGF-induced Trk A autophosphorylation in vitro. Thus, our data support that the effect of caveolin expression on NGF-induced differentiation may be due to changes in the magnitude and duration of ligand-induced receptor autophosphorylation mediated through a direct and inhibitory interaction of caveolin with Trk A.
The lack of NGF-induced differentiation in Cav-PC12 cells cannot be explained by simple variability in the amount of Trk A immunoprecipitated from the Cav-PC12 cells, because Trk A levels were very comparable at all time points examined (Fig.  3). Additionally, because we observed some in vivo autophosphorylation of Trk A in the Cav-PC12 cells, it is unlikely that caveolin is preventing the binding of NGF to Trk A. However, we cannot rule out the possibility that caveolin may interfere with or destabilize the formation of stable Trk A dimers, leading to transient phosphorylation in vivo and blocking activation in vitro (9). Additionally, we cannot rule out the possibility that, in vivo, caveolin may recruit tyrosine phosphatases that rapidly dephosphorylate Trk A. However, the ability of the GST-caveolin fusion protein to prevent the activation of immunoprecipitated Trk A would suggest that a tyrosine phosphatase is not necessary for the inhibition of ligand-induced receptor autophosphorylation.
A previous study has demonstrated that caveolin inhibited epidermal growth factor receptor autophosphorylation, likely through a direct interaction of an N-terminal region of caveolin termed scaffolding domain 1 with the epidermal growth factor receptor catalytic domain (37). Couet et al. (37) have identified that scaffolding domain 1 of caveolin interacts with a specific caveolin binding domain present in the conserved kinase domain of many receptor-linked tyrosine kinases including Trk A. However, we were not able to demonstrate a strong interaction between Trk A and a caveolin fusion protein containing the scaffolding domain 1 of caveolin (less than 1% of the added Trk A bound to the GST-caveolin 61-101 fusion protein in contrast to a 29% binding to full-length caveolin). Thus, although tyrosine kinase receptors share a very similar sequence in this putative caveolin binding domain (37), differences may exist in how caveolin interacts with these receptors and affects kinase activity. Indeed, the idiosyncratic nature of the interaction of caveolin with tyrosine kinase receptors is exemplified by the association of the epidermal growth factor receptor and the insulin receptor with caveolin. Although the interaction of both of these receptors with caveolin is mediated by the interaction of each receptor's caveolin binding domain with scaffolding domain 1 of caveolin, caveolin inhibits epidermal growth factor receptor kinase activity (37) but stimulates insulin receptor kinase activity (38).
Surprisingly, p75 NTR , but not Trk A, was found to interact with caveolin through scaffolding domain 1. It is interesting that the cytoplasmic domain of p75 NTR does not contain what would be considered a strong match to sequences that were identified as putative caveolin binding domains (41), although it does possess a related juxtamembrane sequence that may be responsible for this interaction. However, a peptide of this region was unable to totally compete the binding of p75 NTR to full-length caveolin (data not shown), suggesting that other sequences within p75 NTR may regulate the interaction of this receptor with scaffolding domain 1 of caveolin (Fig. 6). Thus, p75 NTR and Trk receptors may interact with distinct domains of caveolin.
The identification that both classes of neurotrophin receptors may localize to the CRDs of synaptosomal membranes (48) and that caveolin and caveolin-related proteins are expressed in specific neuronal populations (49,50) suggests that the interactions of these proteins may be relevant to neurotrophin signaling pathways. Indeed, similar to the pharmacologic inhibition of Trk A with K252a, caveolin expression modulated the inhibition of p75 NTR signaling. In the absence of K252a, Cav-PC12 cells exhibited significant SM hydrolysis at all time points except 15 min after treatment with NGF (Fig. 8). This was the same time at which we observed a brief period of Trk A activation in Cav-PC12 cells. This pattern suggests that Trk A activation may quickly turn off SM hydrolysis in cells, and that caveolin, in turn, can regulate the ability of Trk A to inhibit p75 NTR signaling. These results provide the first evidence that caveolin may be an endogenous modulator of crosstalk between tyrosine kinase and sphingolipid signaling pathways coupled to Trk receptors and p75 NTR , respectively.
It is important to point out that there is little or no data that support an overlapping distribution between caveolin (49,50) and Trk A (51) in cells of the central nervous system. Therefore, FIG. 5. Trk A co-sediments with a GST-caveolin fusion protein.
The indicated amounts of the agarose-conjugated GST or GST-caveolin were incubated with 30 g of protein from CRDs, NCM-1, and NCM-2 membranes in IP buffer at 4°C for 2 h. The beads were washed with MBST, and the fusion proteins were eluted from the beads with 10 mM glutathione. Immunoblot analysis was performed using an anti-pan-Trk antibody. The results presented are from one experiment performed twice with a similar outcome. the physiologic consequence of the inhibition of Trk A by caveolin is uncertain but may serve as a model for the interaction of caveolin with other Trk family members. However, while this article was under review, it was reported that PC12 cells may express caveolin endogenously, and that caveolin expression increases during NGF-induced differentiation of PC12 cells (43). In agreement with this report, we were able to detect some caveolin in wild type PC12 cells, but only when using an antibody raised against the N terminus of caveolin (Fig. 1E). At first, these results would seem incongruous with our findings because if PC12 cells express caveolin endogenously and caveolin inhibits Trk A tyrosine kinase activity, how do PC12 cells undergo differentiation in response to NGF? It is intriguing that Galbiati et al. (43) reported that caveolin expression was up-regulated 2-3-fold by day 4 of the NGF-induced differentiation of PC12 cells. We achieved an increase of similar magnitude by stable overexpression of caveolin in undifferentiated PC12 cells (Fig. 1E). These results would be consistent with the hypothesis that increased expression of caveolin or caveolinrelated molecules in terminally differentiating neurons may help to regulate signals originating from Trk receptors. Because the process of differentiation in PC12 cells is likely to involve multiple pathways, it is conceivable that caveolin overexpression in undifferentiated PC12 cells may not permit Trk A to become sufficiently activated to couple effectively to the various downstream signaling molecules involved in differentiation (14). However, in undifferentiated wild type cells, endogenous caveolin levels may be insufficient to interfere with the prolonged Trk A activation induced by NGF addition. As FIG. 6. Trk A and p75 NTR show differential binding to caveolin scaffolding domain 1. A, an agarose-conjugated GST-p75 NTR cytoplasmic domain fusion protein (30 g) was incubated with equal amounts of recombinant full-length caveolin (1-178), caveolin (1-81), or caveolin (61-101). p75 NTR -associated proteins were isolated and subjected to SDS-PAGE. After transferring the proteins to nitrocellulose, the membrane was stained with Ponceau red and cut just below the migration of the GST-p75 NTR fusion protein, and the top half of the blot was probed with anti-p75 NTR antiserum, and the lower half was probed with the C terminus polyclonal caveolin antibody. The schematic shows the full-length and truncated caveolin proteins with the caveolin transmembrane region depicted in black. B, CRDs were isolated from Trk A-PC12 cells, and 30 g of protein were incubated with an equal amount of agarose-conjugated GST or GST-caveolin fusion protein. After 1 h at 4°C, the caveolin-associated proteins were sedimented by centrifugation, the beads were washed, and the proteins were subjected to SDS-PAGE. After the transfer to nitrocellulose, the presence of Trk A was determined using a pan-Trk antibody. The results are representative of those obtained from at least three experiments.

FIG. 7. GST-caveolin inhibits Trk A autophosphorylation in an
in vitro kinase assay. Trk A was immunoprecipitated from Trk A-PC12 cells, and the final PAS beads were resuspended in 400 l of kinase buffer. 100-l aliquots were then incubated for 1 h at 4°C with the indicated amounts of GST, GST-caveolin, or glutathione elution buffer. The reaction was initiated by the addition of 40 M ATP and 100 ng/ml NGF and stopped after 15 min at 37°C. The PAS beads containing the immunoprecipitated Trk A were washed with ice-cold MBST, and the proteins were subjected to SDS-PAGE. Trk autophosphorylation was determined by immunoblot analysis using an antiphosphotyrosine antibody. Data are from a representative experiment performed three times with identical results.

FIG. 8. Caveolin prevents Trk A inhibition of NGF-induced SM
hydrolysis. Cellular SM in wild type and Cav-PC12 cells was labeled with [ 3 H]choline for 3 days. The cells were washed with ice-cold PBS and placed in RPMI 1640 medium containing 1% horse serum for 4 h. Cells were then incubated for the indicated times with 100 ng/ml NGF, washed with ice-cold PBS, and lysed in 0.5 M Na 2 CO 3 ; the CRDs were isolated, and the protein was content measured. Total lipids from the CRDs were then extracted, and the amount of [ 3 H]SM was determined. SM levels were normalized to total protein concentration. The results shown are from a representative experiment performed twice. caveolin levels increase during the process of differentiation, the interaction of caveolin with Trk A may increase. Our data suggest that this interaction would promote more a transient phosphorylation of Trk A by NGF (Fig. 3) that may be sufficient to maintain survival. This would be consistent with the ongoing requirement for NGF activation of Trk A for the survival of differentiated PC12 cells (44).
In summary, the neurotrophin/p75 NTR /Trk ligand-receptor cassette is an excellent model system for elucidating the mechanisms whereby interactions between multiple ligands and receptors dictate cellular responses. This ligand-receptor cassette is unique from other signaling systems in that both receptor components can signal individually or in pair in response to multiple ligands and direct very distinct biologic outcomes. These complex interactions necessitate a high degree of organization and regulation. The mutual interaction of these receptors with a caveolar structural protein that organizes, sequesters, and regulates the activity of critical signaling molecules adds an additional level of complexity to this system. These interactions may have important functions in dictating neurotrophin responses in specific cell populations, i.e. regulating the rate of receptor clustering, which may affect the activation of both tyrosine kinase and sphingomyelinase. Elucidating the biochemical mechanisms affecting the interactions between the sphingolipid and tyrosine kinase signaling pathways and their biological consequences represents a novel and emerging area of cell regulation.