Portal de Periódicos da CAPES

Cholesterol Depletion of Caveolae Causes Hyperactivation of Extracellular Signal-related Kinase (ERK)

1998; Elsevier BV; Volume: 273; Issue: 33 Linguagem: Inglês

10.1074/jbc.273.33.21099

1083-351X

Takemitsu Furuchi, Richard G.W. Anderson,

Cholesterol and Lipid Metabolism

Previously we showed that activation of Erk in quiescent cells occurs in the caveolae fraction isolated from fibroblasts. Since the structure and function of caveolae is sensitive to the amount of cholesterol in the membrane, it might be that a direct link exists between the concentration of membrane cholesterol and mitogen-activated protein (MAP) kinase activation. We acutely lowered the cholesterol level of the caveolae fraction by incubating Rat-1 cells in the presence of either cyclodextrin or progesterone. Cholesterol-depleted caveolae had a reduced amount of several key protein components of the MAP kinase complex, including Ras, Grb2, Erk2, and Src. Incubation of these cells in the presence of epidermal growth factor (EGF) caused a rapid loss of EGF receptor from the caveolae fraction, but the usual recruitment of c-Raf was markedly inhibited. Despite the reduced amount of c-Raf and Erk2 in the cholesterol-depleted caveolae fraction, EGF caused a hyperactivation of the remaining caveolae Erk isoenzymes. This was followed by an increase in the amount of active Erk in the cytoplasm. The increased amount of activated Erk produced under these conditions was linked to a 2-fold higher level of EGF-stimulated DNA synthesis. Even cholesterol depletion by itself stimulated Erk activation and DNA synthesis. These results suggest that the MAP kinase pathway can connect the cholesterol level of caveolae membrane to the control of cell division. Previously we showed that activation of Erk in quiescent cells occurs in the caveolae fraction isolated from fibroblasts. Since the structure and function of caveolae is sensitive to the amount of cholesterol in the membrane, it might be that a direct link exists between the concentration of membrane cholesterol and mitogen-activated protein (MAP) kinase activation. We acutely lowered the cholesterol level of the caveolae fraction by incubating Rat-1 cells in the presence of either cyclodextrin or progesterone. Cholesterol-depleted caveolae had a reduced amount of several key protein components of the MAP kinase complex, including Ras, Grb2, Erk2, and Src. Incubation of these cells in the presence of epidermal growth factor (EGF) caused a rapid loss of EGF receptor from the caveolae fraction, but the usual recruitment of c-Raf was markedly inhibited. Despite the reduced amount of c-Raf and Erk2 in the cholesterol-depleted caveolae fraction, EGF caused a hyperactivation of the remaining caveolae Erk isoenzymes. This was followed by an increase in the amount of active Erk in the cytoplasm. The increased amount of activated Erk produced under these conditions was linked to a 2-fold higher level of EGF-stimulated DNA synthesis. Even cholesterol depletion by itself stimulated Erk activation and DNA synthesis. These results suggest that the MAP kinase pathway can connect the cholesterol level of caveolae membrane to the control of cell division. Cholesterol is an important structural lipid that modulates the fluidity of biological membranes. Several new studies also indicate that membrane cholesterol has a direct role in cell signaling. The membrane-bound transcription factor SREBP 1The abbreviations used are: SREBPsterol regulatory element-binding proteinERendoplasmic reticulumMAPmitogen-activated proteinDMEMDulbecco's modified Eagle's mediumEGFepidermal growth factormAbmonoclonal antibodypAbpolyclonal antibodyCDcyclodextrin. controls cholesterol homeostasis (1Brown M.S. Goldstein J.L. Cell. 1997; 89: 331-340Abstract Full Text Full Text PDF PubMed Scopus (3028) Google Scholar). The release of the mature SREBP from the ER is regulated by SCAP, a protein containing up to 8 membrane-spanning regions. One of the transmembrane domains of SCAP appears to sense the level of cholesterol in the membrane and regulate the proteolytic cleavage of SREBP when cellular cholesterol gets low. Similar cholesterol “sensing” domains are present in the NPC1 (2Carstea E.D. Morris J.A. Coleman K.G. Loftus S.K. Zhang D. Cummings C. Gu J. Rosenfeld M.A. Pavan W.J. Krizman D.B. Nagle J. Polymeropoulos M.H. Sturley S.L. Ioannou Y.A. Higgins M.E. Comly M. Cooney A. Brown A. Kaneski C.R. Blanchette-Mackie E.J. Dwyer N.K. Neufeld E.B. Chang T.Y. Liscum L. Tagle D.A. et al.Science. 1997; 277: 228-231Crossref PubMed Scopus (1229) Google Scholar) and PATCHED gene products (3Perrimon N. Cell. 1996; 86: 513-516Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar). The former is a molecule involved in controlling cholesterol traffic from the lysosome whereas the later is a morphogen receptor. PATCHED is inhibited by the binding of Hh (hedgehog), a protein that is post-translationally modified in the ER by the covalent addition of cholesterol (4Porter J.A. Young K.E. Beachy P.A. Science. 1996; 274: 255-259Crossref PubMed Scopus (1113) Google Scholar, 5Ruiz i Altaba A. Cell. 1997; 90: 193-196Abstract Full Text Full Text PDF PubMed Scopus (168) Google Scholar). Cholesterol may anchor Hh to the membrane. These findings suggest that cholesterol-protein interactions can directly regulate the expression of genes important for cell behavior and development. sterol regulatory element-binding protein endoplasmic reticulum mitogen-activated protein Dulbecco's modified Eagle's medium epidermal growth factor monoclonal antibody polyclonal antibody cyclodextrin. Another protein that interacts with cholesterol is caveolin-1 (6Murata M. Peränen J. Schreiner R. Wieland F. Kurzchalia T.V. Simons K. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 10339-10343Crossref PubMed Scopus (774) Google Scholar). This integral membrane protein is found principally in surface caveolae and the Golgi apparatus (7Rothberg K.G. Heuser J.E. Donzell W.C. Ying Y.-S. Glenney J.R. Anderson R.G.W. Cell. 1992; 68: 673-682Abstract Full Text PDF PubMed Scopus (1873) Google Scholar, 8Kurzchalia T.V. Dupree P. Parton R.G. Kellner R. Virta H. Lehnert M. Simons K. J. Cell Biol. 1992; 118: 1003-1014Crossref PubMed Scopus (465) Google Scholar) of many cells. Caveolin-1 appears to have a major function in carrying cholesterol between ER and caveolae membranes (9Anderson R.G.W. Annu. Rev. Biochem. 1998; 67 (in press)Crossref PubMed Scopus (1727) Google Scholar). This is accomplished using a novel membrane traffic route that shuttles caveolin-1 back and forth between the two compartments (10Smart E.J. Ying Y. Donzell W.C. Anderson R.G.W. J. Biol. Chem. 1996; 271: 29427-29435Abstract Full Text Full Text PDF PubMed Scopus (463) Google Scholar). Ordinarily the cholesterol-to-protein ratio of the caveolae fraction is 4–5 times higher than the surrounding plasma membrane (10Smart E.J. Ying Y. Donzell W.C. Anderson R.G.W. J. Biol. Chem. 1996; 271: 29427-29435Abstract Full Text Full Text PDF PubMed Scopus (463) Google Scholar). The maintenance of this level of cholesterol in the caveolae fraction of fibroblasts and transformed lymphocytes (10Smart E.J. Ying Y. Donzell W.C. Anderson R.G.W. J. Biol. Chem. 1996; 271: 29427-29435Abstract Full Text Full Text PDF PubMed Scopus (463) Google Scholar) appears to depend on the caveolin-1 shuttle. Experimentally lowering the cholesterol level of the caveolae fraction disrupts the molecular organization of the domain (11Chang W.-J. Rothberg K.G. Kamen B.A. Anderson R.G.W. J. Cell Biol. 1992; 118: 63-69Crossref PubMed Scopus (211) Google Scholar) and inhibits internalization of both molecular (11Chang W.-J. Rothberg K.G. Kamen B.A. Anderson R.G.W. J. Cell Biol. 1992; 118: 63-69Crossref PubMed Scopus (211) Google Scholar) and particulate material by potocytosis (12Baorto D.M. Gao Z. Malaviya R. Dustin M.L. van der Merwe A. Lublin D.M. Abraham S.N. Nature. 1997; 389: 636-639Crossref PubMed Scopus (244) Google Scholar). Treatment of cells with cholesterol-binding drugs such as filipin appears to have a similar effect (13Schnitzer J.E. Oh P. Pinney E. Allard J. J. Cell Biol. 1994; 127: 1217-1232Crossref PubMed Scopus (778) Google Scholar). In addition to being a portal of entry into the cell, caveolae are the surface locations where multiple signaling pathways converge (9Anderson R.G.W. Annu. Rev. Biochem. 1998; 67 (in press)Crossref PubMed Scopus (1727) Google Scholar, 14Anderson R.G.W. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 10909-10913Crossref PubMed Scopus (570) Google Scholar). One of the pathways organized at this site links multiple receptor tyrosine kinases to MAP kinases (15Mineo C. James G.L. Smart E.J. Anderson R.G.W. J. Biol. Chem. 1996; 271: 11930-11935Abstract Full Text Full Text PDF PubMed Scopus (404) Google Scholar, 16Liu P. Ying Y. Ko Y.-G. Anderson R.G.W. J. Biol. Chem. 1996; 271: 10299-10303Abstract Full Text Full Text PDF PubMed Scopus (338) Google Scholar). Interactions between as many as 11 different molecules can take place during activation of a MAP kinase, and representatives of each are present in the caveolae fraction of quiescent cells (17Liu P. Ying Y.-S. Anderson R.G.W. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 13666-13670Crossref PubMed Scopus (193) Google Scholar). These molecules apparently are functionally linked to MAP kinase because incubation of fibroblasts in the presence of growth factors activates the Erk in the caveolae fraction, which is followed by the appearance of the active enzyme in the cytoplasm (17Liu P. Ying Y.-S. Anderson R.G.W. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 13666-13670Crossref PubMed Scopus (193) Google Scholar). Incubation of isolated caveolae fractions in the presence of platelet-derived growth factor also activates the resident Erk enzymes (17Liu P. Ying Y.-S. Anderson R.G.W. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 13666-13670Crossref PubMed Scopus (193) Google Scholar). The molecular machinery responsible for holding these molecules so they can optimally interact during signal transduction is not known. Based on what is known about caveolae membrane structure (9Anderson R.G.W. Annu. Rev. Biochem. 1998; 67 (in press)Crossref PubMed Scopus (1727) Google Scholar), however, cholesterol could be an important determinant in organizing the MAP kinase complex. We now report that depletion of cholesterol from the caveolae fraction results in a decline in several key proteins of the MAP kinase complex, an increase in the sensitivity of the resident Erk enzymes to stimulation, and an increase in mitogenesis. Fetal calf serum was from Hazleton Research Products, Inc. (Lenexa, KS). Dulbecco's modified Eagle's medium (DMEM), glutamine, trypsin-EDTA, penicillin/streptomycin, and OptiPrep were from Life Technologies, Inc. Percoll was from Amersham Pharmacia Biotech. EGF was from Calbiochem. Immobilon transfer nylon (polyvinylidene difluoride) was from Millipore. ECL Western blotting detection reagents were from Amersham. 2-Hydroxypropyl-β-cyclodextrin and cholesterol were from Sigma. [3H]Thymidine (86.2 Ci/mmol) was from NEN Life Sciences. Antibodies were obtained from the following sources. Anti-Raf-1 (mAb), anti-Ha-Ras (mAb), anti-Sos-1 IgG (mAb), anti-Erk2, anti-Grb2 IgG (mAb), and anti-caveolin-1 (pAb) were from Transduction Laboratories (Lexington, KY). Anti-pan Ras IgG (mAb) and anti-EGF receptor IgG (mAb) were from Calbiochem. Anti-Src (mAb) and anti-Shc (pAb) were from Upstate Biotechnology Inc. (Lake Placid, NY). Anti-active Erk1 and -2 (pAb) were from Promega (Madison, WI). Anti-Gαi (pAb) and anti-Gβ (pAb) were gifts from Dr. Susanne Mumby (The University of Texas Southwestern Medical Center at Dallas). Both goat anti-rabbit IgG and goat anti-mouse IgG conjugated to horseradish peroxidase were from Cappel (Durham, NC). Two culture conditions were used. For all of the immunoblotting experiments, Rat-1 cells (5 × 105cells) were seeded in 100-mm dishes and grown in 10 ml of DMEM supplemented with 10% (v/v) fetal calf serum for 2 days. Cells were then incubated for 24–40 h in serum-free DMEM before the indicated treatments. For all of the mitogenesis experiments, Rat-1 cells (105 cells) were seeded into 12-well plates and grown for 2 days in 1 ml of DMEM supplemented with 10% (v/v) fetal calf serum. Cells were washed once with 1 ml of DMEM and incubated further in DMEM without serum for 24 h at 37 °C. Cyclodextrin (CD) was dissolved in DMEM and used directly. Cholesterol depletion was carried out by incubating cells in the presence of cyclodextrin for 1 h at 37 °C. Cells were repleted with cholesterol by incubating them in the presence of a cholesterol/CD mixture for 1 h at 37 °C. A stock solution of 0.4 mg/ml cholesterol and 10% cyclodextrin was prepared by vortexing at ∼40 °C in 10 ml of 10% cyclodextrin with 200 μl of cholesterol (20 mg/ml in ethanol solution). Each solution was filtered through a 0.2-μm filter before use. Detergent-free caveolae fractions were prepared using two OptiPrep gradients according to the method of Smart et al. (18Smart E.J. Ying Y.-S. Mineo C. Anderson R.G.W. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 10104-10108Crossref PubMed Scopus (676) Google Scholar). Fourteen fractions were collected from the first OptiPrep gradient. Fractions 1–7 were collected, mixed with 50% OptiPrep, and overlaid with 1 ml of 15% OptiPrep plus 0.5 ml of 5% OptiPrep and centrifuged at 52,000 × g for 90 min. An opaque band located between 15% and 5% interface was collected and designated as the caveolae fraction. Fractions 8 and 9 from OptiPrep 1 were pooled and designated the intermediate fractions while pooled fractions 10–14 were designated the heavy fractions. Proteins were separated by SDS-polyacrylamide gel electrophoresis (19Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207523) Google Scholar), transferred to Immobilon-P membrane, and incubated in 5% nonfat milk dissolved in TBS (20 mm Tris-HCl, pH 7.5, 150 mm NaCl, 0.05% Tween 20). Primary antibodies were diluted in TBS plus 1% nonfat milk using the dilutions recommended by the suppliers. The membranes were incubated in the presence of the indicated primary antibody overnight at 4 °C and washed 4 times in TBS for 5 min at room temperature. The appropriate secondary antibody conjugated to horseradish peroxidase was diluted (from 1:5,000 to 1:30,000) in TBS plus 1% nonfat milk. The membranes were incubated in the presence of the second antibody at room temperature for 1–2 h. The membranes were then washed 5 times with TBS at room temperature. ECL was used to visualize the reactive proteins. Cells were incubated in the presence of the indicated concentrations of either CD or cholesterol-CD complex for 1 h at 37 °C. The cells were then incubated in the presence or absence of EGF (50 ng/ml) for 21 h at 37 °C. [3H]Thymidine (1 μCi, 0.4 Ci/mmol) was added to each well, and the cells were incubated for an additional 3 h at 37 °C. The medium was removed, and the cells were washed twice with 1 ml of phosphate-buffered saline. Ice-cold 10% trichloroacetic acid (1 ml) was then added to the well for 30 min to precipitate the DNA and remove all soluble isotope. Finally, the precipitate was dissolved in 0.5 ml of 0.1 n NaOH, mixed with 10 ml of scintillation fluid, and counted in a Beckman LS 6000SC. Protein concentrations were determined using the Bradford assay (20Bradford M.M. Anal. Biochem. 1976; 72: 248-254Crossref PubMed Scopus (217508) Google Scholar). Cholesterol was measured according to the method of Gamble et al. (21Gamble W. Vaughan M. Kruth H.S. Avigan J. J. Lipid Res. 1978; 19: 1068-1070Abstract Full Text PDF PubMed Google Scholar). We used Rat-1 cells and EGF to study the effects of cholesterol depletion on MAP kinase activation. These cells synthesize very little cholesterol after growing for 24–48 h in the absence of serum (data not shown). Cholesterol was specifically (22Neufeld E.B. Cooney A.M. Pitha J. Dawidowicz E.A. Dwyer N.K. Pentchev P.G. Blanchette-Mackie E.J. J. Biol. Chem. 1996; 271: 21604-21613Abstract Full Text Full Text PDF PubMed Scopus (323) Google Scholar) removed from the plasma membrane (∼65% decrease) by incubating the cells in the presence of 2% cyclodextrin (2-hydroxypropyl-β-cyclodextrin) for 1 h at 37 °C. The membrane cholesterol levels returned to near normal when the cells were subsequently incubated in the presence of a cholesterol (16 μg/ml)-cyclodextrin (0.4%) complex for 1 h. No change occurred, however, if the complex was omitted from the incubation media. We used detergent-free membrane fractionation (18Smart E.J. Ying Y.-S. Mineo C. Anderson R.G.W. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 10104-10108Crossref PubMed Scopus (676) Google Scholar) to determine how cyclodextrin affected the cholesterol level of caveolae fractions compared with other fractions of the plasma membrane (Fig.1). The first gradient used in this procedure separates low density plasma membranes rich in caveolae from bulk membrane protein. The majority of the protein (66%) was in fractions 8–14, and the protein profile did not change after cyclodextrin treatment (compare ■ with ○, C). The top seven fractions were pooled and used to enrich for caveolae on a second gradient (CV fraction, 2.6% of membrane protein) while fractions 8 and 9 (intermediate fractions) and 10–14 (heavy fractions) where pooled and analyzed directly. The cholesterol-to-protein ratio in the caveolae fraction declined by 68% in response to cyclodextrin treatment (A, compare lanes 1 and 2). The ratio remained low after the cells were incubated further for 1 h in the absence of cyclodextrin (A, lane 3) but returned to near normal after the addition of cholesterol-CD complex to the media (A, lane 4). By contrast, cyclodextrin had much less effect on the specific cholesterol content of intermediate (lanes 5–8) and heavy (lanes 9–12) fractions. Since we did not detect any change in the concentration of other major lipids in these fractions (including phosphatidic acid, phosphatidylethanolamine, phosphatidylcholine, and sphingomyelin (data not shown)), cyclodextrin can be used to both remove and restore membrane cholesterol without causing major changes to other lipids in the bilayer. Cyclodextrin also caused a reversible decrease in the amount of caveolin-1 in the caveolae fraction (Fig. 1 B,CV). Samples of caveolae (lanes 1–4, 5 μg/lane), intermediate (lanes 5–8, 10 μg/lane), and heavy (lanes 9–12, 10 μg/lane) fractions from cells subjected to various treatments were separated and blotted with anti-caveolin-1 IgG. As expected, caveolin-1 was highly enriched in the caveolae fraction from control cells (lane 1). Some caveolin-1 was also detected in the intermediate fractions (lane 5) but none was present in the heavy fractions (lane 9). The amount of caveolin-1 in the caveolae fractions significantly declined after cyclodextrin treatment (compare lanes 2 and 1). Cyclodextrin also caused a modest increase in the amount of caveolin-1 in the intermediate fractions (compare lanes 5and 6). Caveolin-1 levels in the caveolae fraction declined further (lane 3) when the cells were incubated an additional hour in the absence of cyclodextrin. Caveolin-1 in the intermediate fraction (lane 7) also decreased under these conditions but the total amount of caveolin-1 in the cell remained constant (data not shown). Once cholesterol was restored to the membranes by incubating cells in the presence of the cholesterol-CD complex (lanes 4, 8, and12), the concentration of caveolin-1 in the various fractions returned to nearly the same level as in control cells. The total protein recovered in the caveolae fraction was not altered by any of these treatments (Fig. 1 C), and the total caveolin-1 in the cell remained constant (data not shown). Therefore, lowering the cholesterol level in the caveolae fraction results in a redistribution of caveolin-1. Previously we showed that the major protein components of the MAP kinase complex are concentrated in the caveolae fractions isolated from quiescent cells (17Liu P. Ying Y.-S. Anderson R.G.W. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 13666-13670Crossref PubMed Scopus (193) Google Scholar). We used immunoblotting to determine whether cholesterol depletion had any effect on the concentration of these proteins (Fig. 2). Equal amounts of protein (5 μg/lane) from cholesterol-depleted (lane 2) and non-depleted (lane 1) caveolae fractions were separated by gel electrophoresis and blotted with specific antibodies. As we have shown in other studies (15Mineo C. James G.L. Smart E.J. Anderson R.G.W. J. Biol. Chem. 1996; 271: 11930-11935Abstract Full Text Full Text PDF PubMed Scopus (404) Google Scholar, 16Liu P. Ying Y. Ko Y.-G. Anderson R.G.W. J. Biol. Chem. 1996; 271: 10299-10303Abstract Full Text Full Text PDF PubMed Scopus (338) Google Scholar, 17Liu P. Ying Y.-S. Anderson R.G.W. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 13666-13670Crossref PubMed Scopus (193) Google Scholar, 23Chang W.-J. Ying Y.-S. Rothberg K.G. Hooper N.M. Turner A.J. Gambliel H.A. De Gunzburg J. Mumby S.M. Gilman A.G. Anderson R.G.W. J. Cell Biol. 1994; 126: 127-138Crossref PubMed Scopus (311) Google Scholar,24Wu C. Butz S. Ying Y. Anderson R.G.W. J. Biol. Chem. 1997; 272: 3554-3559Abstract Full Text Full Text PDF PubMed Scopus (239) Google Scholar), caveolae fractions from control cells were enriched in caveolin-1 (Cav), EGF receptors (A, EGF-R), all three isoforms of Shc (A, Shc), Grb2 (A, Grb2), both Gαi (A,Gα) and Gβ (B,Gβ) heterotrimeric proteins, both Ki- and Ha-Ras (A, K-Ras, H-Ras), and Src (B, Src). In cholesterol-depleted caveolae, by contrast, the amount of caveolin-1, Grb2, Ras, and Src was markedly reduced. Shc isoforms and Gβ were decreased to a lesser extent. EGF receptors and Gαi were unchanged, suggesting that cholesterol depletion does not alter the entire protein composition of caveolae. The concentration of each of the depleted proteins increased after cells were incubated in the presence of the cholesterol-CD complex (lane 4), but not if the cells were incubated in media alone (lane 3). Within minutes after Rat-1 cells are exposed to EGF, c-Raf is recruited to caveolae fractions (15Mineo C. James G.L. Smart E.J. Anderson R.G.W. J. Biol. Chem. 1996; 271: 11930-11935Abstract Full Text Full Text PDF PubMed Scopus (404) Google Scholar). We looked at the effect of cholesterol depletion on the EGF-stimulated movement of c-Raf and six other protein components of the MAP kinase complex including Erk2 (Fig.3). Untreated (CD−) and cyclodextrin-treated (CD+) cells were incubated in the presence of EGF for 0, 1, and 3 min before caveolae fractions were prepared and processed for immunoblotting. Cholesterol depletion markedly suppressed the appearance of c-Raf in the caveolae fraction (compare c-Raf lanes 1–3 with 4–6). It also suppressed the appearance of SOS and Shc (comparelanes 1–3 with 4–6). Proteins that ordinarily do not change in response to EGF (Erk2, Gαi, Grb2, and Ras) remained constant. The basal level of Erk2, however, was much lower in cholesterol-depleted caveolae fractions (lanes 4–6). The concentration of Ras, Grb2, and Shc was also lower. Therefore, cholesterol depletion had a profound effect on the molecular composition of the MAP kinase complex in caveolae fractions. We used an antibody that recognizes activated Erk1 and -2 to determine how the cholesterol depletion of the caveolae fraction affected the ability of EGF to activate these enzymes (Fig.4). Two different incubation sequences were used to look at early (0, 1, 3 min, lanes 1–6) and late (0, 5, 10 min, lanes 7–12) effects of EGF. Cells were preincubated in the presence (CD+) or absence (CD−) of cyclodextrin before the addition of EGF, and both the caveolae (Cav) and cytoplasmic (Cyt) fractions were analyzed. The EGF receptor (EGF-R) moved out of the cholesterol-depleted caveolae fraction with the same rapid kinetics as it does from untreated caveolae fractions (Cav, compare lanes 1–3 with4–6). Cholesterol depletion also suppressed c-Raf (c-Raf) recruitment and decreased the level of Erk2 (Erk2) without affecting the Gαi(Gα) concentration (Cav, compare lanes 1–3 with 4–6 and 9–11 with12–14). By contrast, cholesterol depletion caused a marked increase in the amount of activated Erk (phoMAPK) in the caveolae fraction Cav, compare lanes 3 with6 and 9–11 with 12–14). Even cyclodextrin alone caused a significant increase in activated Erk (Cav, compare lanes 1 with 4 and9 with 12). Usually the appearance of activated Erk in the cytoplasm lags a little behind the caveolae fraction (17Liu P. Ying Y.-S. Anderson R.G.W. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 13666-13670Crossref PubMed Scopus (193) Google Scholar), but after 5 or 10 min in the presence of EGF we saw a similar dramatic increase in the amount of active enzyme in the cytoplasmic fraction of cholesterol-depleted cells (Cyt, comparelanes 9–11 with 12–14. The cytoplasmic concentrations of c-Raf and Erk2 were unaffected by cholesterol depletion. The effect of cyclodextrin on Erk activation was blunted by cholesterol (Fig. 5). Cells were incubated in the presence (Chol/CD+) or absence (Chol/CD−) of cholesterol-CD complex for 1 h at 37 °C before the addition of EGF for 0, 1, and 3 min (lanes 1–6). Cyclodextrin plus cholesterol had little effect on c-Raf recruitment relative to control treatments (compare lanes 1–3 with 4–6). Moreover, EGF-stimulated Erk activation (phoMAPK) was only slightly elevated compared with untreated cells. These results suggest that it is the decrease in caveolae cholesterol, not the presence of cyclodextrin, that is responsible for the increase in activated Erk. The caveolae fraction can also be depleted of cholesterol by incubating the cells in the presence of progesterone. A 1-h incubation in the presence of 10 μg/ml progesterone reduced the specific cholesterol content of the caveolae fraction from 0.7 to 0.45 mg/mg of protein. Although this amount of loss was ∼50% less than what occurred with cyclodextrin, progesterone specifically affected the cholesterol level of caveolae. There was virtually no decrease in the specific cholesterol content of non-caveolae membrane fractions (data not shown). Progesterone-treated (+, lanes 4–6) and non-treated (−, lanes 1–3) cells were incubated in the presence of EGF for 0, 5, and 10 min (Fig. 6). Similar to cyclodextrin, progesterone suppressed EGF-stimulated c-Raf recruitment (Cav, c-Raf), reduced the level of Erk2 in the caveolae fraction (Cav, Erk2), but had little effect on the level of EGF receptors. In addition, the amount of both basal and EGF-stimulated activated Erk (phoMAPK) in the caveolae fraction was markedly increased in progesterone-treated cells (Cav, compare lanes 1–3 with4–6). The cytoplasmic level of activated kinase was also increased by progesterone (Cyt, compare lanes 1–3 with 4–6). Therefore, two different methods of lowering the cholesterol level of the caveolae fraction render the resident MAP kinase hyper-responsive to EGF. Since EGF acts through the MAP kinase pathway to stimulate mitogenesis, we used [3H]thymidine incorporation to determine whether cyclodextrin had any effect on DNA synthesis (Fig.7). Cells were incubated in the presence of various concentrations of cyclodextrin for 1 h. The cells were further incubated for 24 h in the presence (•) or absence (○) of EGF (A) before measuring [3H]thymidine incorporation. In the absence of EGF, increasing the concentration of cyclodextrin in the media caused an increase in mitogenesis, which is consistent with its ability to activate Erk. Stimulation plateaued at 1.5% cyclodextrin. EGF alone also stimulated [3H]thymidine incorporation, but cyclodextrin increased the level of DNA synthesis further. Cyclodextrin plus EGF caused a 3-fold higher level of DNA synthesis than no treatment. The synergistic effects of cyclodextrin and EGF were blocked by cholesterol (B). In the absence of cholesterol, EGF and cyclodextrin together caused a 2.5-fold stimulation in mitogenesis. As the concentration of cholesterol was increased, however, thymidine incorporation progressively declined (•). Two experimental conditions that acutely lower the cholesterol level of the caveolae fraction cause a resident population of MAP kinase to become hyper-responsive to EGF. While cyclodextrin caused a general cholesterol depletion of the plasma membrane, progesterone removed cholesterol specifically from the caveolae fraction. The effects of cholesterol depletion on caveolae MAP kinase activity were blunted by incubating the cells in the presence of cyclodextrin plus cholesterol. These results suggest that the caveolae MAP kinase signaling pathway is influenced by how much cholesterol is in the membrane. Cholesterol depletion caused a disruption in the molecular organization of the MAP kinase signaling complex. Lowering the cholesterol caused a decline in many of the key molecules in the complex, including Erk2. It also blocked EGF-stimulated c-Raf recruitment to caveolae, as well as the recruitment of other molecules in the complex. This loss of organization might have two potential effects. 1) It uncoupled a feedback loop between EGF receptors and Erk2, leading to the generation of higher than normal amounts of activated Erk2 per bound EGF. 2) The decreased concentration of Erk in caveolae caused a hyperphosphorylation of the enzyme by MEK (25Burack W.R. Sturgill T.W. Biochemistry. 1997; 36: 5929-5933Crossref PubMed Scopus (155) Google Scholar). The MAP kinase activated under these circumstances could stimulate DNA synthesis nearly 2-fold above EGF alone. Even cholesterol depletion by itself activated Erk and stimulated DNA synthesis. Therefore, a direct connection exists between the downstream signaling activities of Erk and the cholesterol content of the plasma membrane. Biophysical studies (26Schroeder R. London E. Brown D. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 12130-12134Crossref PubMed Scopus (638) Google Scholar, 27Ahmed S.N. Brown D.A. London E. Biochemistry. 1997; 36: 10944-10953Crossref PubMed Scopus (615) Google Scholar, 28Brown D.A. London E. Biochem. Biophys. Res. Commun. 1997; 240: 1-7Crossref PubMed Scopus (464) Google Scholar) suggest caveolae membrane lipids are in a liquid-order phase (referred to as the β phase) because of a high concentration of glycosphingolipids, sphingomyelin, and cholesterol. The β phase has characteristic physical properties, such as resistance to solubilization by Triton X-100 (28Brown D.A. London E. Biochem. Biophys. Res. Commun. 1997; 240: 1-7Crossref PubMed Scopus (464) Google Scholar) and a relatively low lateral mobility of resident lipids (26Schroeder R. London E. Brown D. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 12130-12134Crossref PubMed Scopus (638) Google Scholar). This special lipid environment appears to be important for attracting and organizing lipid-anchored proteins (glycosylphosphatidylinositol, acyl, and prenyl anchors) in caveolae. Since cholesterol is a critical ingredient for the β phase transition (28Brown D.A. London E. Biochem. Biophys. Res. Commun. 1997; 240: 1-7Crossref PubMed Scopus (464) Google Scholar), decreasing the cholesterol content of caveolae can directly affect the number and organization of lipid-anchored proteins as a result of changing the phase properties of the bilayer (11Chang W.-J. Rothberg K.G. Kamen B.A. Anderson R.G.W. J. Cell Biol. 1992; 118: 63-69Crossref PubMed Scopus (211) Google Scholar, 29Smart E.J. Mineo C. Anderson R.G.W. J. Cell Biol. 1996; 134: 1169-1177Crossref PubMed Scopus (79) Google Scholar). Any protein that interacts with lipid-anchored proteins in caveolae will be similarly affected. Therefore, the loss of components from the MAP kinase complex following cholesterol depletion as well as the hyper-responsive Erk can be traced to a stringent requirement for high cholesterol to maintain the proper organization of caveolae membranes. The maintenance of high cholesterol levels in caveolae is an active process. Newly synthesized cholesterol transported directly from the ER to the plasma membrane (30Liscum L. Underwood K.W. J. Biol. Chem. 1995; 270: 15443-15446Abstract Full Text Full Text PDF PubMed Scopus (198) Google Scholar) initially appears in the caveolae membrane (10Smart E.J. Ying Y. Donzell W.C. Anderson R.G.W. J. Biol. Chem. 1996; 271: 29427-29435Abstract Full Text Full Text PDF PubMed Scopus (463) Google Scholar). Caveolin-1 may be necessary for caveolae to have high cholesterol levels (10Smart E.J. Ying Y. Donzell W.C. Anderson R.G.W. J. Biol. Chem. 1996; 271: 29427-29435Abstract Full Text Full Text PDF PubMed Scopus (463) Google Scholar), which suggests it is a component of the ER-to-caveolae shuttle. Moreover, caveolin-1 moves from caveolae through the ER to the Golgi apparatus when caveolae cholesterol becomes oxidized (31Smart E.J. Ying Y.-U. Conrad P.A. Anderson R.G.W. J. Cell Biol. 1994; 127: 1185-1197Crossref PubMed Scopus (378) Google Scholar). Caveolae can also acquire cholesterol released from low density lipoproteins in lysosomes (32Fielding P.E. Fielding C.J. Biochemistry. 1996; 35: 14932-14938Crossref PubMed Scopus (89) Google Scholar). Furthermore, the high density lipoprotein-binding protein SR-B1, which binds both normal and damaged lipoproteins (33Acton S. Rigotti A. Landschulz K.T. Xu S. Hobbs H.H. Krieger M. Science. 1996; 271: 518-520Crossref PubMed Scopus (2011) Google Scholar), has been localized to caveolae (34Babitt J. Trigatti B. Rigotti A. Smart E.J. Anderson R.G. Xu S. Krieger M. J. Biol. Chem. 1997; 272: 13242-13249Abstract Full Text Full Text PDF PubMed Scopus (337) Google Scholar) where it may facilitate the bidirectional transfer of cholesterol between lipoproteins and caveolae. The convergence of both cholesterol import and export at caveolae provides the cell with redundant mechanisms for delivering cholesterol to this membrane domain. Several pathological conditions exist where the concentration of cholesterol in caveolae might influence the transduction of mitogenic signals. Caveolae cholesterol levels could either be too high or too low. Overaccumulation of cellular cholesterol is common in cells within atherosclerotic lesions. No one has studied the MAP kinase activity in caveolae of cholesterol-enriched cells, but our results predict the enzyme would be hyporesponsive to growth factors such as EGF. Recently, Pomerantz et al. (35Pomerantz K.B. Lander H.M. Summers B. Robishaw J.D. Balcueva E. Hajjar D.P. Biochemistry. 1997; 36: 9523-9531Crossref PubMed Scopus (22) Google Scholar) reported that cultured smooth muscle cells with elevated levels of cholesterol exhibit a reduced MAP kinase activity in response to platelet-derived growth factor. At the other extreme, transformed cells have below normal amounts of caveolae cholesterol, which is restored to normal after the cells are transfected with caveolin-1 (10Smart E.J. Ying Y. Donzell W.C. Anderson R.G.W. J. Biol. Chem. 1996; 271: 29427-29435Abstract Full Text Full Text PDF PubMed Scopus (463) Google Scholar). Loss of caveolin-1 often occurs in cells infected with oncogenic viruses (36Koleske A.J. Baltimore D. Lisanti M.P. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 1381-1385Crossref PubMed Scopus (472) Google Scholar), and expression of caveolin-1 in transformed cells can revert the cell to a normal phenotype (37Engelman J.A. Wykoff C.C. Yasuhara S. Song K.S. Okamoto T. Lisanti M.P. J. Biol. Chem. 1997; 272: 16374-16381Abstract Full Text Full Text PDF PubMed Scopus (335) Google Scholar). This suggests caveolin-1 acts as a tumor suppresser (38Couet J. Li S.W. Okamoto T. Scherer P.E. Lisanti M.P. Trends Cardiovasc. Med. 1997; 7: 103-110Crossref PubMed Scopus (111) Google Scholar). Since the hyper-responsive MAP kinase in cholesterol-depleted caveolae is mitogenic (Fig. 7), the tumor-suppressing effect of caveolin-1 may be due to its ability to return the concentration of cholesterol in caveolae to normal. The ability of cholesterol to directly influence an entire signal pathway suggests that the concentration of cholesterol in each membrane compartment within the cell must be tightly controlled. Otherwise signals originating from sources such as the EGF receptor would be unreliable. Therefore, in addition to balancing endogenous and exogenous sources of cholesterol (1Brown M.S. Goldstein J.L. Cell. 1997; 89: 331-340Abstract Full Text Full Text PDF PubMed Scopus (3028) Google Scholar), cells must regulate the supply of cholesterol to specific membrane compartments such as caveolae. This level of regulation should involve at least two components: a sensing device that detects how much cholesterol is present in the membrane of the compartment and a transporter that responds to the cholesterol sensor by delivering or removing cholesterol. Caveolin-1 is a candidate component of the transporter that supplies cholesterol for the caveolae compartment. So far no one has identified a protein in caveolae that contains a transmembrane domain homologous to those thought to be involved in detecting membrane cholesterol (1Brown M.S. Goldstein J.L. Cell. 1997; 89: 331-340Abstract Full Text Full Text PDF PubMed Scopus (3028) Google Scholar). We would like to thank William Donzell and Ann Horton for their valuable technical assistance and Stephanie Baldock for administrative assistance. Drs. Pingsheng Liu and Chieko Mineo provided valuable insights during the course of these studies.