(Received for publication, November 13, 1996, and in revised form, January 27, 1997)
From the Department of Pharmacology and Toxicology, University of Kansas, Lawrence, Kansas 66045
Caveolae are plasma membrane microdomains that are enriched in caveolin, the structural protein of caveolae, sphingomyelin, and other signaling molecules. We previously suggested that neurotrophin-induced p75NTR-dependent sphingomyelin hydrolysis may be localized to the plasma membrane. Therefore, we examined if caveolae were a major site of p75NTR-dependent sphingomyelin hydrolysis in p75NTR-NIH 3T3 fibroblasts. Caveolin-enriched membranes (CEMs) were prepared by either detergent or detergent-free extraction and separated from noncaveolar membranes by centrifugation through sucrose gradients. Immunoblot analysis of the individual gradient fractions indicated that caveolin and p75NTR were enriched in CEMs. The localization of p75NTR to CEMs was not an artifact of receptor overexpression in the fibroblasts because a similar distribution of p75NTR was evident from PC12 cells, which endogenously express p75NTR. In the p75NTR fibroblasts, nerve growth factor induced a time-dependent hydrolysis of sphingomyelin only in CEMs with no hydrolysis detected in noncaveolar membranes. Intriguingly, endogenous p75NTR was found to co-immunoprecipitate with caveolin, suggesting that p75NTR may associate with caveolin in vivo. This interaction was confirmed in vitro by the co-immunoprecipitation of a glutathione S-transferase fusion protein expressing the cytoplasmic domain of p75NTR with caveolin. Collectively, these results demonstrate that neurotrophin-induced p75NTR-dependent sphingomyelin hydrolysis localizes to CEMs and suggest that the interaction of p75NTR with caveolin may affect signaling through p75NTR.
The neurotrophins are a family of growth factors involved in the survival, development, and death of specific populations of neurons and glial cells (1). The signal transduction systems that mediate these divergent biologies are initiated by the interaction of neurotrophins with two classes of cell surface receptors (2). Many of the trophic signals elicited by neurotrophins are initiated by the binding of these molecules to various Trk tyrosine kinase receptors (3-9). In contrast, recent data suggest that neurotrophin-mediated death signals are generated through the interaction of nerve growth factor (NGF)1 with a transmembrane protein known as the low affinity neurotrophin receptor, p75NTR (10-12). Although the signaling mechanisms used by p75NTR are still uncertain, we have demonstrated that p75NTR couples to sphingomyelin (SM) hydrolysis, producing the bioactive sphingolipid metabolite ceramide (13, 14).
Ceramide production following ligand binding of p75NTR has been implicated in antiproliferative responses in glial cells (15). For example, exogenous ceramide mimicked the effect of NGF on cell growth inhibition and differentiation of rat T9 glioma cells (13). Additionally, NGF-induced p75NTR-dependent apoptosis correlates with an early and prolonged increase in cellular ceramide levels in mature oligodendrocytes (10). Further, NGF-induced p75NTR-dependent ceramide production has been implicated in regulating dopamine release in primary cultures of mesencephalic neurons (16).
The coupling of p75NTR to ceramide production and the emerging role of this receptor-effector system in glial and neuronal biologies necessitates the identification of the signal-sensitive pool of lipid, localization of the specific compartment for signal generation/regeneration, and defining the site of interaction with potential downstream effector molecules.
Previous studies have suggested that ligand-induced SM hydrolysis may
use a plasma membrane pool of SM, which is not effected by treatments
that inhibit receptor internalization (14, 17, 18). Caveolae are
dynamic microdomains of the plasma membrane that are enriched in SM and
remain associated with the cell surface (19), suggesting that caveolae
may be localized sites for ligand-induced SM hydrolysis. Indeed,
interleukin-1 treatment of dermal fibroblasts induced SM hydrolysis
and ceramide generation within caveolae (20). Moreover, caveolae have
been implicated as sites for signal transduction via the epidermal
growth factor and platelet-derived growth factor receptors (21, 22).
Thus, caveolae are emerging as important signaling compartments within
the plasma membrane (23).
Caveolin is the major protein component of caveolae and is necessary for formation of intact caveolae (23). Caveolin serves as a scaffolding protein (24) involved in sequestering pools of G-proteins (25), Ras (26), and other effector molecules (21, 22, 27). Moreover, caveolin can interact with proteins in a very specific manner such that only inactive forms of G-proteins are found associated with caveolin (25). Thus, caveolin may be an important component of receptor/effector systems localized within caveolae.
The abundance of SM within caveolae and the role of caveolin in sequestering signaling molecules led us to examine if p75NTR-dependent SM hydrolysis localized to caveolae and if this receptor interacted with caveolin. We demonstrate herein that p75NTR and caveolin co-localize and that p75NTR-dependent SM hydrolysis occurs within CEMs. Further, p75NTR was found to co-immunoprecipitate with caveolin from CEMs. Collectively, these results suggest that caveolae may provide a highly localized site for neurotrophin-induced SM hydrolysis and that the association of p75NTR with caveolin may have a functional role in neurotrophin signaling through p75NTR.
Mouse 2.5 S NGF was obtained from Harlan Bioproducts for Science. [3H]Choline chloride (86 Ci/mmol) was purchased from American Radiolabeled Chemicals. Polyclonal anti-caveolin antibodies were products of Signal Transduction Labs. Anti-GST monoclonal antibody was obtained from Santa Cruz Biotechnology. Anti-p75NTR antibody 9991 (13) and the GST-p75NTR fusion protein expressing the cytoplasmic domain of p75NTR were generously provided by Dr. Moses Chao. Protein A-Sepharose and glutathione-Sepharose were from Pharmacia Biotech Inc. All tissue culture media and geneticin were products of Life Technologies, Inc.
Cell LinesThe production and characterization of the p75NTR-NIH 3T3 cells has been previously described (29). p75NTR-NIH 3T3 and PC12 cells were cultured as described (14). NIH 3T3 cells stably transfected with a pCMV vector lacking the p75NTR cDNA insert were maintained in Dulbecco's modified Eagle's medium/10% fetal calf serum containing 0.2 mg/ml geneticin. All cells were maintained at 37 °C in a humidified atmosphere containing 5% CO2.
Isolation of CEMsCEMs were prepared by either detergent or nondetergent extraction of cell proteins and centrifugation over discontinuous sucrose gradients essentially as described (26, 30). Cells were grown in 15-cm tissue culture dishes, and two dishes of cells were used to prepare the CEMs. The medium was aspirated, and the cells were washed with 2 × 10 ml of ice-cold phosphate-buffered saline. The cells were scraped into 2 ml of ice-cold MBST buffer (25 mM MES, 150 mM NaCl, 1% Triton X-100, 1 mM Pefabloc) and homogenized with a loose fitting Dounce homogenizer (20 strokes). The extract was adjusted to 40% sucrose by the addition of 2 ml of 80% sucrose in MBS lacking Triton X-100 and placed in the bottom of an ultracentrifuge tube. A discontinuous sucrose gradient was formed by overlaying this solution with 4 ml of 38% sucrose and 4 ml of 5% sucrose (both in MBS). The tubes were centrifuged at 31,000 rpm in a SW41 rotor for 16-18 h at 4 °C and 15 × 0.8 ml fractions were collected manually from the top of the gradient. Aliquots of each fraction were used for the determination of protein content (31).
To determine the distribution of p75NTR and caveolin within the gradient, 0.075 ml of each fraction were subjected to SDS-PAGE on 10 or 12% acrylamide gels followed by immunoblotting. Following SDS-PAGE, the proteins were transferred to nitrocellulose, and the membrane was stained with 0.5% Ponceau S to visualize total protein. As described previously (30), the bulk of the cellular proteins remained in fractions 10-15 of the gradient (data not shown). The membrane was typically cut in half at about the 44-kDa molecular mass marker, and the upper half of the blot was used to probe for p75NTR (75 kDa), and the lower half of the blot was used to probe for caveolin (22-24 kDa). Proteins were visualized using enhanced chemiluminescence (Amersham Corp.).
In some experiments, CEMs were prepared by nondetergent extraction with sodium carbonate (26). Detergent-free extraction of the cells was performed as above with the following modifications. The cells were scraped into 2 ml of 0.5 M sodium carbonate, pH 11, and homogenization was carried out sequentially with a loose fitting Dounce homogenizer (20 strokes), a Polytron tissue grinder (3 × 10-s bursts), and a sonicator (3 × 20-s bursts). The extract was brought to 45% sucrose (in MBS), and the discontinuous gradient was formed as above in sucrose solutions prepared in MBS containing 0.25 M sodium carbonate (26). Following centrifugation, individual fractions were collected as above. In some experiments, the individual gradient fractions were pooled into CEM (fractions 4-7), NCM-1 (fractions 8-11), and NCM-2 (fractions 12-15). The membranes from the pooled gradient fractions were isolated by centrifugation at 100,000 × g for 0.5 h at 4 °C. The isolated membrane fractions were used for immunoprecipitation or to measure lipid content.
Co-immunoprecipitation of p75NTR with CaveolinThe
membrane pellets from the CEM, NCM-1, and NCM-2 were resuspended in 0.3 ml of IP buffer (10 mM Tris-HCl, pH 7.4, 10 mM EDTA, 1% Nonidet P-40, 0.4% deoxycholate, 60 mM
-octylglucoside, 1 mM phenylmethylsulfonyl fluoride, and
10 µg/ml each of leupeptin, aprotinin, and bestatin). The samples
were precleared with 2 µg of IgG for 0.5 h at 4 °C. Following
removal of IgG with protein A-Sepharose, 2 µg of anti-caveolin
polyclonal antibody was added, and the samples were incubated for
2 h at 4 °C. Immune complexes were formed by the addition of
protein A-Sepharose and rotation for 1 h at 4 °C. The immune
complexes were sedimented by centrifugation at 14,000 rpm in a
microfuge and washed with 3 × 1 ml of ice-cold MBST. The bound
proteins were solubilized in sample buffer, and p75NTR and
caveolin were detected by immunoblotting as described above.
In experiments utilizing GST-p75NTR, the fusion protein was purified from bacterial lysates by affinity chromatography with glutathione-Sepharose (25). The purified fusion protein was released from the glutathione-Sepharose beads, and 0-30 µg of the fusion protein incubated directly with CEMs solubilized in IP buffer. Caveolin was immunoprecipitated, and the GST-p75NTR fusion protein was detected by immunoblotting with p75NTR antiserum or a monoclonal antibody against the GST affinity handle.
Metabolic Labeling and Sphingomyelin MeasurementsCellular SM pools were labeled with [3H]choline, and SM was quantitated as described previously (14).
To determine if a
pool of p75NTR may localize to CEMs, we subjected
p75NTR-NIH 3T3 cells to sodium carbonate extraction and
analyzed the individual gradient fractions for the distribution of
p75NTR and caveolin. Immunoblot analysis for p75NTR and
caveolin revealed that a large pool of p75NTR co-migrated with
caveolin (Fig. 1).
Because fibroblasts do not endogenously express p75NTR, it was
possible that the presence of p75NTR in CEMs was an artifact of
receptor overexpression in the fibroblasts. Therefore, we determined
whether p75NTR co-localized with CEMs isolated from PC12 cells,
which endogenously express p75NTR. Fractionation of PC12 cells
under the above conditions resulted in a similar distribution of
p75NTR within the gradient (Fig. 2). However,
immunoblot analysis did not detect caveolin, indicating that PC12 cells
either do not express caveolin or that they express an isoform not
recognized by the polyclonal antibody used in these experiments. The
sedimentation of the membranes within this area of the sucrose gradient
is indicative of a high buoyant density, suggesting that a pool of
p75NTR associates with areas of the membrane enriched in
glyco-sphingolipids but not necessarily enriched in caveolin (33).
Interestingly, detergent extraction of either the p75NTR-NIH 3T3 fibroblasts or the PC12 cells resulted in less p75NTR migrating in fractions 4-7 than was apparent under conditions of detergent-free extraction (data not shown). However, in the fibroblasts, caveolin still primarily migrated to fractions 4-7 due to the marked insolubility of this protein in Triton X-100 (33). Differences in the amount of p75NTR co-migrating with CEMs under conditions of detergent versus detergent-free extraction may be due to detergent solubilization of a palmitoylated form of p75NTR (35). Previous studies have identified that detergent extraction results in the loss of other lipid-modified proteins from CEMs (26). In contrast, detergent-free extraction of the cells with sodium carbonate facilitates the recovery of lipid-modified proteins within CEMs (26, 34). Indeed, labeling of the p75NTR fibroblasts with [3H]palmitate followed by detergent extraction resulted in the majority of palmitoylated p75NTR migrating in fractions 9-15, suggesting that detergent extraction may affect the partitioning of p75NTR between CEMs and NCMs (data not shown).
Neurotrophins Induce Sphingomyelin Hydrolysis within CEMsWe
have previously reported that neurotrophins hydrolyze an internal pool
of SM (14). Because the CEMs are enriched in SM (19, 20) and contain a
significant pool of p75NTR, we asked whether
neurotrophin-induced SM hydrolysis localized to this region of the
plasma membrane. We treated p75NTR-NIH 3T3 cells with
phosphate-buffered saline or NGF, detergent-extracted the cells, and
fractionated the membranes as above. The majority of the SM (47%)
migrated in fractions 4-7, with the remainder primarily distributed to
fractions 12-15 (35%). However, analysis of the amount of SM within
each fraction revealed that NGF hydrolyzed SM solely in the CEMs,
fractions 4-7 (Fig. 3).
To determine if the method of preparation of CEMs had any effect on the distribution of the neurotrophin sensitive pool of SM, we treated [3H]choline-labeled cells with phosphate-buffered saline or NGF and subjected them to detergent-free extraction. As expected, about 95% of the total SM recovered in the gradient fractions migrated within the CEMs (fractions 4-7) following sodium carbonate extraction (Table I). NGF increased SM hydrolysis in the CEMs in a time-dependent manner such that up to 19% of the SM within this fraction was hydrolyzed within 12 min (Table I). We conclude that although the extraction method used for the preparation of CEMs effects the distribution of p75NTR and SM within the sucrose gradient, the primary signal-sensitive pool of SM still localizes to CEMs. Intriguingly, neurotrophin-3 also induced SM hydrolysis primarily in fractions 4-7 from membranes isolated from PC12 cells (data not shown). Because the PC12 cells lacked caveolin, these results suggest that the presence of caveolin is not necessary for the coupling of p75NTR to SM hydrolysis.
|
Caveolin can associate with the endothelin Type A receptor (28), bind inactive forms of some signaling proteins, i.e. Ras and G-proteins (25, 26, 34), and form a high molecular weight complex composed of caveolin homooligomers (24). These observations, together with the results described above, led us to investigate whether p75NTR may directly interact with caveolin.
The p75NTR-NIH 3T3 fibroblasts were subjected to detergent-free extraction, and the membranes were separated over the sucrose gradient as described. Based upon the fractionation pattern of p75NTR shown in Fig. 1, the individual fractions were combined into CEMs (fractions 4-7), NCM-1 (fractions 8-11), and NCM-2 (fractions 12-15). Fractions 1-3 were discarded because they were found to contain minimal protein and no p75NTR or caveolin. The membranes from the pooled fractions were sedimented by centrifugation and then subjected to immunoprecipitation with a polyclonal anti-caveolin antibody.
Fig. 4A shows that the caveolin antibody
immunoprecipitated caveolin from the CEMs as well as some caveolin
present in the concentrated NCMs. However, the majority of caveolin was
observed in the CEMs. The presence of two caveolin bands in the CEMs
represent the and
isoforms of this protein (26). Although no
caveolin was detected in the individual fractions comprising NCM-1 and NCM-2 (Fig. 1), concentrating the membrane fractions may help immunoprecipitate caveolin being processed and transported to the
caveolae.
Intriguingly, p75NTR only co-immunoprecipitated with caveolin from CEMs, despite the presence of significant amounts of p75NTR within the other gradient fractions, especially NCM-2 (fractions 12-15, Fig. 1). These results suggest that p75NTR may specifically interact in vivo with organized forms of caveolin that are only present within caveolae (24).
Importantly, the caveolin antibody did not immunoprecipitate any caveolin nor p75NTR from membranes isolated from fractions 4-7 from PC12 cells (Fig. 4A). Because PC12 cells do not express any proteins recognized by the caveolin antibody, these results indicate that p75NTR was not nonspecifically associating with the antibody. Additionally, the caveolin antibody did not co-immunoprecipitate p75NTR from CEMs nor NCMs isolated from mock transfected fibroblasts, strongly supporting that caveolin can specifically associate with p75NTR in cells expressing both caveolin and p75NTR (Fig. 4B).
To further confirm that p75NTR could associate with caveolin,
we examined the ability of a GST-p75NTR fusion protein
expressing the full-length cytoplasmic domain of p75NTR to
associate with caveolin in vitro. CEMs isolated from mock transfected NIH 3T3 fibroblasts were incubated with 0 or 30 µg of the
GST-p75NTR fusion protein. Subsequently, caveolin was
immunoprecipitated with a polyclonal caveolin antibody. After formation
of the immune complexes with protein A-Sepharose, the bound proteins
were solubilized and resolved by SDS-PAGE. The proteins were
transferred to nitrocellulose, the blots were cut in half just above
the 32-kDa molecular marker, and the upper half of the blot was probed
with either the p75NTR antiserum (Fig.
5A) or a monoclonal antibody against the GST affinity handle (Fig. 5B). In both panels, the lower portion
of the blots were probed with the anti-caveolin antibody. Fig. 5 (A and B) shows that the caveolin antibody
co-immunoprecipitated the GST-p75NTR fusion protein as detected
using either the p75NTR antiserum or the GST antibody. As
expected, caveolin was equally immunoprecipitated in all instances.
Importantly, the caveolin antibody did not co-immunoprecipitate any
p75NTR or GST immunoreactive bands from CEMs incubated with GST
only (Fig. 5C). Thus, co-immunoprecipitation of the
GST-p75NTR fusion protein with caveolin was not mediated by an
interaction between caveolin and the GST affinity handle. Collectively,
these results strongly support that caveolin can directly associate with a region within the cytoplasmic domain of p75NTR.
Caveolae are emerging as highly localized sites for lipid signaling events. Indeed, recent results indicate that epidermal growth factor-induced hydrolysis of phosphatidylinositol-bis-phosphate also occurs solely within CEMs from A431 cells (36). Moreover, caveolae may serve as sites for cross-talk between tyrosine kinase and lipid signaling pathways. Because signaling via the epidermal growth factor and platelet-derived growth factor receptors occurs within CEMs (21, 22), other receptor-linked tyrosine kinases may be resident in CEMs. These results raise the possibility that localization of p75NTR signaling to CEMs may have functional significance for cross-talk between Trk and p75NTR signaling pathways. We and others have reported that activation of Trk tyrosine kinases can inhibit p75NTR-dependent SM hydrolysis and ceramide generation (14, 16). Therefore, co-localization of p75NTR with Trk receptors within caveolae may enrich p75NTR in an area of the plasma membrane where it can effectively couple to its putative effector system (SM hydrolysis/ceramide generation) or interact with the Trk effector system.
p75NTR has been demonstrated to associate with the 42- and
44-kDa forms of mitogen activated protein kinase (37), an undefined 120/104-kDa kinase (38), and two unidentified phosphoproteins of 60 and
130 kDa (39). However, the role of these proteins in p75NTR
signaling remains unknown. The role of caveolin as a scaffolding protein involved in organizing and sequestering signaling molecules within caveolae (23, 24) suggests that the association of caveolin with
p75NTR may have functional consequences for p75NTR
signaling. For example, the association of p75NTR with caveolin
may directly or indirectly impart specificity to neurotrophin signaling
through p75NTR. Recent results indicate that NGF but not BDNF
nor neurotrophin-3 can activate NF-B in primary cultures of Schwaan
cells (40) and induce apoptosis in primary cultures of mature
oligodendrocytes (10). Presumably, both of these effects are mediated
through NGF-induced p75NTR-dependent ceramide
generation. Thus, although all the neurotrophins can couple to SM
hydrolysis through p75NTR in transfected fibroblasts
overexpressing the receptor (14), in primary cultures of glial cells,
other factors must affect neurotrophin signaling through
p75NTR.
In conclusion, the localization of p75NTR-dependent SM hydrolysis to CEMs extends our previous observations on p75NTR signaling through SM metabolism and in the broader perspective implicates caveolin as a potentially important component in affecting cellular responses to neurotrophins through p75NTR.
We thank Drs. Y. Hannun and Bob Palazzo for critical review of the manuscript and Dr. Moses Chao for various reagents and antibodies.