Cholesterol Depletion of Enterocytes
2000; Elsevier BV; Volume: 275; Issue: 7 Linguagem: Inglês
10.1074/jbc.275.7.5136
ISSN1083-351X
AutoresGert H. Hansen, Lise-Lotte Niels-Christiansen, Evy Thorsen, Lissi Immerdal, E. Michael Danielsen,
Tópico(s)Lipid Membrane Structure and Behavior
ResumoIntestinal brush border enzymes, including aminopeptidase N and sucrase-isomaltase, are associated with "rafts" (membrane microdomains rich in cholesterol and sphingoglycolipids). To assess the functional role of rafts in the present work, we studied the effect of cholesterol depletion on apical membrane trafficking in enterocytes. Cultured mucosal explants of pig small intestine were treated for 2 h with the cholesterol sequestering agent methyl-β-cyclodextrin and lovastatin, an inhibitor of hydroxymethylglutaryl-coenzyme A reductase. The treatment reduced the cholesterol content >50%. Morphologically, the Golgi complex/trans-Golgi network was partially transformed into numerous 100–200 nm vesicles. By immunogold electron microscopy, aminopeptidase N was localized in these Golgi-derived vesicles as well as at the basolateral cell surface, indicating a partial missorting. Biochemically, the rates of the Golgi-associated complex glycosylation and association with rafts of newly synthesized aminopeptidase N were reduced, and less of the enzyme had reached the brush border membrane after 2 h of labeling. In contrast, the basolateral Na+/K+-ATPase was neither missorted nor raft-associated. Our results implicate the Golgi complex/trans-Golgi network in raft formation and suggest a close relationship between this event and apical membrane trafficking. Intestinal brush border enzymes, including aminopeptidase N and sucrase-isomaltase, are associated with "rafts" (membrane microdomains rich in cholesterol and sphingoglycolipids). To assess the functional role of rafts in the present work, we studied the effect of cholesterol depletion on apical membrane trafficking in enterocytes. Cultured mucosal explants of pig small intestine were treated for 2 h with the cholesterol sequestering agent methyl-β-cyclodextrin and lovastatin, an inhibitor of hydroxymethylglutaryl-coenzyme A reductase. The treatment reduced the cholesterol content >50%. Morphologically, the Golgi complex/trans-Golgi network was partially transformed into numerous 100–200 nm vesicles. By immunogold electron microscopy, aminopeptidase N was localized in these Golgi-derived vesicles as well as at the basolateral cell surface, indicating a partial missorting. Biochemically, the rates of the Golgi-associated complex glycosylation and association with rafts of newly synthesized aminopeptidase N were reduced, and less of the enzyme had reached the brush border membrane after 2 h of labeling. In contrast, the basolateral Na+/K+-ATPase was neither missorted nor raft-associated. Our results implicate the Golgi complex/trans-Golgi network in raft formation and suggest a close relationship between this event and apical membrane trafficking. glycosylphosphatidylinositol Madin-Darby canine kidney cell polyacrylamide gel electrophoresis According to the "membrane cluster" hypothesis (1.Simons K. van Meer G. Biochemistry. 1988; 27: 6197-6202Crossref PubMed Scopus (1091) Google Scholar, 2.Simons K. Ikonen E. Nature. 1997; 387: 569-572Crossref PubMed Scopus (8157) Google Scholar, 3.Brown D.A. London E. Annu. Rev. Cell Biol. 1998; 14: 111-136Crossref Scopus (2557) Google Scholar), cholesterol together with sphingoglycolipids spontaneously form microdomains or "rafts" in the exoplasmic leaflet of the bilayer of biological membranes. In physical terms, rafts are characterized as membrane domains that are in a liquid-ordered or gel phase, probably by virtue of the ability of cholesterol and sphingolipids to form hydrogen bonds and, in case of the latter, by having higher acyl chain melting temperatures than glycerolipids (4.van Meer G. Burger K.N.J. Trends Cell Biol. 1992; 2: 332-337Abstract Full Text PDF PubMed Scopus (27) Google Scholar, 5.Schroeder R. London E. Brown D. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 12130-12134Crossref PubMed Scopus (638) Google Scholar, 6.Schroeder R.J. Ahmed S.N. Zhu Y. London E. Brown D.A. J. Biol. Chem. 1998; 273: 1150-1157Abstract Full Text Full Text PDF PubMed Scopus (375) Google Scholar, 7.Ahmed S.N. Brown D.A. London E. Biochemistry. 1997; 36: 10944-10953Crossref PubMed Scopus (615) Google Scholar). The biological significance of rafts lies in their ability to act as lateral sorting platforms for various subsets of membrane proteins that associate with them (2.Simons K. Ikonen E. Nature. 1997; 387: 569-572Crossref PubMed Scopus (8157) Google Scholar, 3.Brown D.A. London E. Annu. Rev. Cell Biol. 1998; 14: 111-136Crossref Scopus (2557) Google Scholar). At the cell surface, raft association may be involved in formation of caveolae and other types of endocytotic trafficking as well as in membrane signaling events. In the biosynthetic exocytotic pathway, raft association in the trans-Golgi network may be involved in formation of transport vesicles specifically targeted to the apical cell surface in epithelial cell types and to the axonal membrane in neuronal cells, but evidence indicates that the raft pathway may well be operating also in non-polarized cells (8.Müsch A. Xu H. Shields D. Rodriguez-Boulan E. J. Cell Biol. 1996; 133: 543-558Crossref PubMed Scopus (148) Google Scholar, 9.Yoshimori T. Keller P. Roth M.-G. Simons K. J. Cell Biol. 1996; 133: 247-256Crossref PubMed Scopus (202) Google Scholar). Rafts can conveniently be isolated from cells by virtue of their resistance to detergent solubilization at low temperature (1.Simons K. van Meer G. Biochemistry. 1988; 27: 6197-6202Crossref PubMed Scopus (1091) Google Scholar, 10.Brown D.A. Rose J.K. Cell. 1992; 68: 533-544Abstract Full Text PDF PubMed Scopus (2618) Google Scholar). Rafts, thus prepared from small intestine as "detergent-insoluble complexes" (also commonly referred to as DIGs or DRMs), were previously found to be highly enriched in several of the major brush border enzymes, including the transmembrane aminopeptidase N and sucrase-isomaltase (11.Danielsen E.M. Biochemistry. 1995; 34: 1596-1605Crossref PubMed Scopus (129) Google Scholar) as well as the GPI1-anchored melanotransferrin (12.Danielsen E.M. van Deurs B. J. Cell Biol. 1995; 131: 939-950Crossref PubMed Scopus (71) Google Scholar). Likewise, in the enterocyte-like Caco-2 cell line, GPI-anchored proteins and sucrase-isomaltase were enriched in the raft fraction (13.Garcia M. Mirre C. Quaroni A. Reggio H. Le Bivic A. J. Cell Sci. 1993; 104: 1281-1290Crossref PubMed Google Scholar, 14.Mirre C. Monlauzeur L. Garcia M. Delgrossi M.-H. Le Bivic A. Am. J. Physiol. 1996; 271: C887-C894Crossref PubMed Google Scholar). In addition, we have identified a member of the galectin family, galectin 4, to be associated with rafts, by forming clusters with the brush border enzymes (15.Danielsen E.M. van Deurs B. Mol. Biol. Cell. 1997; 8: 2241-2251Crossref PubMed Scopus (90) Google Scholar). Recently we observed that the transcytotic route taken by IgA across the enterocyte passes through a raft containing compartment, most likely the apical endosome (16.Hansen G.H. Niels-Christiansen L.-L. Immerdal L. Hunziker W. Kenny A.J. Danielsen E.M. Gastroenterology. 1999; 116: 610-622Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar). Together, these findings imply that raft formation may be of functional importance in the apical membrane trafficking occurring in intestinal epithelial cells. Cholesterol-sequestering agents, sometimes used in combination with cholesterol synthesis inhibitors such as lovastatin (17.Alberts A.W. Chen J. Kuron G. Hunt V. Huff J. Hoffman C. Rothrock J. Lopez M. Joshua H. Harris E. Patchett A. Monaghan R. Currie S. Stapley E. Albers-Schonberg G. Hensens O. Hirshfield J. Hoogsteen K. Liesch J. Springer J. Proc. Natl. Acad. Sci. U. S. A. 1980; 77: 3957-3961Crossref PubMed Scopus (1465) Google Scholar), have been employed in a number of studies to characterize functionally membrane microdomains. Thus, low density lipoprotein deprivation caused a reduction in surface expression of the GPI-anchored gD1-DAF in MDCK cells, but the apical polarity was unaffected (18.Hannan L.A. Edidin M. J. Cell Biol. 1996; 133: 1265-1276Crossref PubMed Scopus (68) Google Scholar). Other types of cholesterol-sequestering agents such as filipin and cyclodextrins have been shown in several studies to disrupt caveolae and to inhibit non-clathrin-coated internalization of various ligands (19.Schnitzer J.E. Oh P. Pinney E. Allard J. J. Cell Biol. 1994; 127: 1217-1232Crossref PubMed Scopus (778) Google Scholar, 20.Orlandi P.A. Fishman P.H. J. Cell Biol. 1998; 141: 905-915Crossref PubMed Scopus (640) Google Scholar, 21.Hailstones D. Sleer L.S. Parton R.G. Stanley K.K. J. Lipid Res. 1998; 39: 369-379Abstract Full Text Full Text PDF PubMed Google Scholar). By using a combination of a rapid membrane cholesterol sequestering by methyl-β-cyclodextrin and cholesterol synthesis inhibition by lovastatin, an apical to basolateral missorting of the transmembrane influenza virus hemagglutinin was observed together with a reduced raft association (22.Keller P. Simons K. J. Cell Biol. 1998; 140: 1357-1367Crossref PubMed Scopus (472) Google Scholar). By using the same experimental approach, an axonal to dendritic missorting was observed for the GPI-anchored Thy-1 in cultured neuronal cells (23.Ledesma M.D. Simons K. Dotti C.G. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 3966-3971Crossref PubMed Scopus (199) Google Scholar). In the present work, methyl-β-cyclodextrin and lovastatin were used to study the effect of cholesterol depletion on the membrane trafficking of aminopeptidase N in enterocytes from mucosal explants. Compared with controls, the newly synthesized brush border enzyme was processed to its mature, complex glycosylated form and associated with rafts at a slower rate. In addition, an extensive vesiculation of the Golgi complex occurred together with an increased basolateral missorting. In contrast, the localization of a basolateral membrane protein, Na+/K+-ATPase, was unaffected by cholesterol depletion. Taken together, the results indicate a role for raft formation in the apical exocytotic membrane trafficking in enterocytes and implicate the Golgi complex as a key organelle in this process. Lovastatin was a kind gift from Alice Friis Petersen of Merck Sharp and Dohme (Glostrup, Denmark). Methyl-β-cyclodextrin was obtained from Sigma. Equipment for performing organ culture, including Trowell's T-8 medium and culture dishes with grids, was obtained as described previously (24.Danielsen E.M. Sjöström H. Norén O. Bro B. Dabelsteen E. Biochem. J. 1982; 202: 647-654Crossref PubMed Scopus (65) Google Scholar). A mouse monoclonal antibody to the α-subunit of human Na+/K+-ATPase was from Affinity Bioreagents, Inc. (Golden, CO). The rabbit antibodies to pig small intestinal aminopeptidase N used for immunogold electron microscopy and immunopurification were as described previously (25.Hansen G.H. Sjöström H. Norén O. Dabelsteen E. Eur. J. Cell Biol. 1987; 43: 253-259PubMed Google Scholar, 26.Sjöström H. Norén O. Christiansen L.A. Wacker H. Semenza G. J. Biol. Chem. 1980; 255: 11332-11338Abstract Full Text PDF PubMed Google Scholar). Segments of pig small intestine were kindly provided by Letty Klarskov and Mette Olesen from the Department of Experimental Medicine, the Panum Institute, Copenhagen, Denmark. After a rinse of the small intestinal segments in ice-cold Hanks' buffered salt solution, mucosal explants were excised and cultured in Trowell's T-8 medium for periods up to 4 h (24.Danielsen E.M. Sjöström H. Norén O. Bro B. Dabelsteen E. Biochem. J. 1982; 202: 647-654Crossref PubMed Scopus (65) Google Scholar). For cholesterol depletion, methyl-β-cyclodextrin (2%) was added either alone or together with 5 mm lovastatin. In some experiments, methyl-β-cyclodextrin-containing medium was preloaded with cholesterol prior to culture by incubation with 1% (w/v) free cholesterol for 24 h at 37 °C under constant shaking. After culture, the explants were either frozen quickly at −20 °C or immersed in fixative for electron microscopy. After culture, mucosal explants were fixed for ultrastructural analysis in 2.5% glutaraldehyde in 0.1 msodium phosphate buffer, pH 7.2, for 2 h. For immunogold labeling the mucosal explants were fixed in either 2% paraformaldehyde and 0.1% glutaraldehyde for 2 h or in 2% paraformaldehyde and 3% glutaraldehyde in the above buffer for 15 min. After a rinse in the above buffer, the explants were prepared for electron microscopy according to standard procedure or infused with 2.3 msucrose at room temperature before storage in liquid nitrogen. Ultracryosectioning and immunogold single and double labeling was performed as described previously (16.Hansen G.H. Niels-Christiansen L.-L. Immerdal L. Hunziker W. Kenny A.J. Danielsen E.M. Gastroenterology. 1999; 116: 610-622Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar, 27.Danielsen E.M. Hansen G.H. Niels-Christiansen L.-L. Gastroenterology. 1995; 109: 1039-1050Abstract Full Text PDF PubMed Scopus (16) Google Scholar). For preparation of rafts from total membranes, explants of about 100 mg of wet weight were thawed in 1 ml of 25 mm HEPES-HCl, 150 mm NaCl, pH 7.0, containing 10 μg of each of aprotinin and leupeptin, and homogenized in a manually operated Potter-Elvehjem homogenizer. The homogenate was cleared by centrifugation at 500 × g, 3 min, and then centrifuged at 48,000 × g, 30 min to obtain a pellet of total membranes. The pellet was resuspended in 1 ml of the above buffer, extracted with 1% Triton X-100 for 10 min on ice, and mixed with an equal volume of 80% sucrose in the same buffer. Low temperature detergent-insoluble complexes (rafts) were prepared according to Brown and Rose (10.Brown D.A. Rose J.K. Cell. 1992; 68: 533-544Abstract Full Text PDF PubMed Scopus (2618) Google Scholar), by layering a 5–30% sucrose gradient on top of the extract and centrifugation in an SW40 Ti rotor (Beckman Instruments, Palo Alto, CA) for 20–22 h at 35,000 rpm (g max = 217,000), as described previously (11.Danielsen E.M. Biochemistry. 1995; 34: 1596-1605Crossref PubMed Scopus (129) Google Scholar). In some experiments, mucosal explants were fractionated by the divalent cation precipitation method into Mg2+-precipitated (intracellular and basolateral membranes) and microvillar membranes (28.Booth A.G. Kenny A.J. Biochem. J. 1974; 142: 575-581Crossref PubMed Scopus (434) Google Scholar), as described previously (29.Danielsen E.M. Biochem. J. 1982; 204: 639-645Crossref PubMed Scopus (50) Google Scholar). SDS-PAGE in 10% gels under reducing conditions was performed according to Laemmli (30.Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207537) Google Scholar). Radioactively labeled proteins were visualized by autoradiography and quantitated by PhosphorImager analysis (Molecular Dynamics Inc., Sunnyvale, CA). The free cholesterol concentration in total mucosal explant membranes and sucrose gradient fractions was determined spectrophotometrically by a cholesterol oxidase/peroxidase assay (31.Gamble W. Vaughan M. Kruth H.S. Avigan J. J. Lipid Res. 1978; 19: 1068-1070Abstract Full Text PDF PubMed Google Scholar), and protein concentration was determined by the method of Bradford (32.Bradford M.M. Anal. Biochem. 1976; 72: 248-254Crossref PubMed Scopus (217544) Google Scholar). Aminopeptidase N activity was determined spectrophotometrically using alanyl-p-nitroanilide as substrate (33.Sjöström H. Norén O. Jeppesen L. Staun M. Svensson B. Christiansen L. Eur. J. Biochem. 1978; 88: 503-511Crossref PubMed Scopus (101) Google Scholar), and the brush border enzyme was immunopurified from solubilized membrane fractions as described previously (27.Danielsen E.M. Hansen G.H. Niels-Christiansen L.-L. Gastroenterology. 1995; 109: 1039-1050Abstract Full Text PDF PubMed Scopus (16) Google Scholar). As shown in TableI, culture of mucosal explants for 2 h in the presence of methyl-β-cyclodextrin caused a reduction in the cellular free cholesterol/protein ratio to less than 60% of the level of controls. When lovastatin was included as well, this ratio was further reduced to less than half the control value. TableII compares the cholesterol content of Mg2+-precipitated (intracellular and basolateral) membranes and microvillar membranes of cholesterol-depleted and control mucosal explants. As should be expected, microvillar membranes had the highest cholesterol content and, being directly exposed to methyl-β-cyclodextrin in the culture medium, were depleted to 36% of the controls. However, the cholesterol reduction was also substantial for the Mg2+-precipitated membranes (61% of the controls), indicating that also intracellular membranes of the enterocyte were affected by the treatment. As shown in Fig.1, about half the amount of cholesterol from a low temperature detergent extract of membranes of control explants floated in the upper fractions (fractions 5–9) after a sucrose gradient centrifugation. This indicates that a major fraction of the cellular cholesterol resides in detergent-resistant glycolipid/cholesterol microdomains or rafts (10.Brown D.A. Rose J.K. Cell. 1992; 68: 533-544Abstract Full Text PDF PubMed Scopus (2618) Google Scholar). Culture for 2 h in the presence of methyl-β-cyclodextrin and lovastatin markedly reduced the amount of cholesterol in the upper fractions of the gradient, indicating that cholesterol depletion affects the composition of glycolipid rafts. However, the fact that some cholesterol remained in the floating fractions following the depletion treatment also shows that rafts per se were not completely abolished by this treatment.Table ICholesterol depletion of mucosal explants by methyl-β-cyclodextrin and lovastatinExplantProteinCholesterolCholesterol/proteinmg/mlTop controlControl3.13 ± 0.300.334 ± 0.067100CD3.51 ± 0.210.213 ± 0.03258CD/lovastatin2.99 ± 0.490.145 ± 0.02546Explants were cultured for 2 h in the absence (control) or presence of 2% methyl-β-cyclodextrin alone (CD) or 2% methyl-β-cyclodextrin and 5 mm lovastatin (CD/lovastatin). After culture, they were homogenized in 1 ml of 25 mm HEPES-HCl, 150 mm NaCl, pH 7.0, and the homogenates were cleared by centrifugation at 500 × g for 3 min. Homogenate concentrations of protein and cholesterol were determined as described under "Experimental Procedures," and each of the data represent the mean ± S.E. of three experiments. Open table in a new tab Table IICholesterol depletion affects both Mg2+-precipitated and microvillar membranesExplantCholesterol/protein (w/w) in Mg2+- precipitated membranesCholesterol/protein (w/w) in microvillar membranesControl0.098 ± 0.011 (100%)0.211 ± 0.030 (100%)CD/lovastatin0.060 ± 0.003 (61%)0.075 ± 0.017 (36%)Explants were cultured as described in the legend to Table I. After culture, they were homogenized in 3 ml of 2 mm Tris-HCl, 50 mm mannitol, pH 7.1, and Mg2+-precipitated (intracellular and basolateral) membranes and microvillar membranes were prepared as described under "Experimental Procedures." The membrane fractions were resuspended in 0.5 ml of 25 mmHEPES-HCl, 150 mm NaCl, pH 7.0, and the concentrations of cholesterol and protein were determined. Each of the data represents the mean ± S.E. of three experiments. (Numbers in parentheses indicate the cholesterol/protein ratio as a percentage of the corresponding control fraction.) Open table in a new tab Explants were cultured for 2 h in the absence (control) or presence of 2% methyl-β-cyclodextrin alone (CD) or 2% methyl-β-cyclodextrin and 5 mm lovastatin (CD/lovastatin). After culture, they were homogenized in 1 ml of 25 mm HEPES-HCl, 150 mm NaCl, pH 7.0, and the homogenates were cleared by centrifugation at 500 × g for 3 min. Homogenate concentrations of protein and cholesterol were determined as described under "Experimental Procedures," and each of the data represent the mean ± S.E. of three experiments. Explants were cultured as described in the legend to Table I. After culture, they were homogenized in 3 ml of 2 mm Tris-HCl, 50 mm mannitol, pH 7.1, and Mg2+-precipitated (intracellular and basolateral) membranes and microvillar membranes were prepared as described under "Experimental Procedures." The membrane fractions were resuspended in 0.5 ml of 25 mmHEPES-HCl, 150 mm NaCl, pH 7.0, and the concentrations of cholesterol and protein were determined. Each of the data represents the mean ± S.E. of three experiments. (Numbers in parentheses indicate the cholesterol/protein ratio as a percentage of the corresponding control fraction.) Culture of mucosal explants for 2 h in the presence of methyl-β-cyclodextrin and lovastatin did not cause any major morphological changes in enterocytes (Fig.2). However, the microvillar length of the brush border was modestly reduced; measurements of 100 microvilli, taken from 20 views of control and cholesterol-depleted enterocytes, respectively, showed a reduction in microvillar length from 1.07 ± 0.16 to 0.84 ± 0.25 μm (Fig.3). The density of microvilli also appeared somewhat reduced and microvesiculation was frequently observed, but the microvilli otherwise appeared normal in shape including the actin cytoskeleton extending well into the underlying terminal web. In parallel cultures using methyl-β-cyclodextrin preloaded with free cholesterol, the microvillar length (0.98 ± 0.17 μm) and density resembled that of control enterocytes. The only intracellular changes in morphology caused by cholesterol depletion concerned the Golgi complex. Compared with those of control explants, the Golgi cisternae were often reduced in size and were surrounded by numerous vesicles, ranging in size from 100 to 200 nm (Fig. 3). In parallel cultures, methyl-β-cyclodextrin preloaded with cholesterol had no visible effect on the morphology of the Golgi complex, indicating that methyl-β-cyclodextrin by itself did not perturb the enterocytes (Fig. 3).Figure 3Microvilli and the Golgi complex are affected by cholesterol depletion. Electron micrographs of the brush border and Golgi complex areas of enterocytes from explants cultured in the absence (A and D) or presence of methyl-β-cyclodextrin (B and E) or methyl-β-cyclodextrin preloaded with free cholesterol prior to culture (C and F). Note that cholesterol depletion caused a moderate decrease in the length and density of microvilli and frequently induced microvillar microvesiculation. Intracellularly, the treatment caused an extensive vesiculation of the Golgi complex. None of these effects were observed with cholesterol-preloaded methyl-β-cyclodextrin. Bars, 0.5 μm.View Large Image Figure ViewerDownload Hi-res image Download (PPT) By immunogold electron microscopy, aminopeptidase N was seen in control enterocytes in the Golgi complex/trans-Golgi network, whereas labeling of the basolateral plasma membrane of mature enterocytes and of endosomes lying in proximity to this cell surface was very sparse, indicating a low steady-state concentration of aminopeptidase N in this region of the cell (Fig. 4). In enterocytes treated for 2 h with methyl-β-cyclodextrin and lovastatin, the vesiculated Golgi complex was also labeled by the antibody to aminopeptidase N, and labeling along the enterocyte basolateral plasma membrane was markedly increased (Fig. 4), suggesting a missorting of the enzyme. To substantiate this, a morphometric analysis based on 100 views of controls and cholesterol-depleted explants, respectively, indicated that the labeling density over the basolateral plasma membrane increased from 1.4 gold particles/view in controls to 3.4 gold particles/view. Both aminopeptidase N as well as the basolateral marker Na+/K+-ATPase were observed in the Golgi areas of both control and cholesterol-depleted enterocytes (Fig.5). The aminopeptidase N labeling density over the Golgi complex and surrounding vesicles was not markedly affected by cholesterol depletion. However, a morphometrical analysis of 20 Golgi areas of control and cholesterol-depleted enterocytes, respectively, revealed a 2-fold increase in the overall size of the organelle (from a supranuclear cytoplasmic density of approximately 5–10%). This implies that an increased amount of newly synthesized brush border enzyme is present in the Golgi area. The numerous vesicles surrounding the Golgi complex in cholesterol-depleted cells were also labeled by antibodies to the Golgi markers β-COP (not shown) and K58 (Fig. 6), indicating that they are indeed derived from the Golgi membranes. In contrast to aminopeptidase N, Na+/K+-ATPase labeling over the brush border membrane remained very weak also in cholesterol-depleted enterocytes (Fig. 7), indicating that the treatment caused no missorting of this basolateral enzyme.Figure 5Colocalization of aminopeptidase N and Na+ /K+-ATPase in the Golgi area. Double immunogold labeling of aminopeptidase N (13-nm gold particles) and Na+/K+-ATPase (7-nm gold particles) in Golgi complexes (GO) of control (A) and cholesterol-depleted (B) enterocytes. Both the apical and basolateral cell surface markers label the Golgi complex in A as well as the vesicles (VE) surrounding the Golgi complex in B. Bars, 100 nm.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 6Colocalization of aminopeptidase N and K58 in the Golgi area of cholesterol-depleted enterocytes. Double immunogold labeling of aminopeptidase N (7-nm particles) and K58 (13-nm particles) in the vesicular structures of the Golgi area of cholesterol-depleted enterocytes. The colocalization with K58 shows that these aminopeptidase N-positive structures are derived from Golgi membranes. Bar, 200 nm.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 7Cell surface colocalization of aminopeptidase N and Na+ /K+-ATPase in cholesterol-depleted enterocytes. Double immunogold labeling of aminopeptidase N (13-nm gold particles) and Na+/K+-ATPase (7-nm gold particles) at the apical (A) and basolateral (B) areas of cholesterol-depleted enterocytes. A, aminopeptidase N is abundantly present in the microvillar membrane (MM) and in an apical endosome (EN), whereas Na+/K+-ATPase labeling in this region of the cell is virtually absent. B, both Na+/K+-ATPase and aminopeptidase N labeling is seen along the basolateral membrane (BM). The latter is also localized in basolateral endosomes (EN). Bars, 100 nm.View Large Image Figure ViewerDownload Hi-res image Download (PPT) In contrast to the Na+/K+-ATPase that was found to be fully solubilized by Triton X-100 at low temperature (Fig.8), brush border enzymes are the most abundant proteins in rafts prepared from intestinal mucosa, and their raft association has previously been shown to take place intracellularly (11.Danielsen E.M. Biochemistry. 1995; 34: 1596-1605Crossref PubMed Scopus (129) Google Scholar). To study the effect of cholesterol depletion on Golgi-associated complex glycosylation and raft association, aminopeptidase N was immunopurified from the low temperature detergent-soluble and -insoluble fractions of explants pulse-labeled for 1 h at 20 °C and chased for 0–90 min at 37 °C (Fig.9). At 20 °C membrane trafficking in enterocytes was almost completely arrested (34.Danielsen E.M. Hansen G.H. Cowell G.M. Eur. J. Cell Biol. 1989; 49: 123-127PubMed Google Scholar), and most of the newly synthesized aminopeptidase N appeared as the transient, high mannose-glycosylated form of 140 kDa in both control and cholesterol-depleted explants. During the chase at 37 °C, conversion to the mature, complex-glycosylated form of 166 kDa of the enzyme rapidly began and reached 60–70% by 90 min in control enterocytes. In comparison, complex glycosylation occurred at a slower rate in enterocytes treated with methyl-β-cyclodextrin and lovastatin and was only about 40% by the end of the chase. Raft association (low temperature detergent insolubility) of the 166-kDa form only started to increase after 30 min of chase and was 40% for the control enterocytes by 90 min of chase, a value close to that previously reported for the enzyme at steady state (11.Danielsen E.M. Biochemistry. 1995; 34: 1596-1605Crossref PubMed Scopus (129) Google Scholar). As with complex glycosylation, raft association occurred more slowly in cholesterol-depleted enterocytes, reaching only 20% by the end of the chase. Fig.10 shows aminopeptidase N from Mg2+-precipitated and microvillar fractions of explants, labeled for 2 h. Compared with the control, a reduced amount of mature 166-kDa form was present in the microvillar fraction of cholesterol-depleted enterocytes, indicating an impaired transport to the apical cell surface. In conclusion, cholesterol depletion affects the rates of both Golgi-associated complex glycosylation and raft association of aminopeptidase N, and less of the newly synthesized enzyme is seen at the brush border by 2 h. Another raft-associated brush border enzyme, sucrase-isomaltase, was studied in parallel with aminopeptidase N, and the rates of complex glycosylation as well as raft association were similarly reduced for this enzyme (data not shown).Figure 9Cholesterol depletion reduces the rate of complex glycosylation and raft association of aminopeptidase N. A, mucosal explants were preincubated for 1 h at 20 °C in the absence (Control) or presence of methyl-β-cyclodextrin and lovastatin (CD/lova). 0.5 mCi/ml [35S]methionine was then added to the media, and the culture at 20 °C continued for 1 h. After labeling, the explants were chased in non-radioactive medium at 37 °C for the indicated periods (still in the absence or presence of methyl-β-cyclodextrin and lovastatin, respectively). After culture, low temperature detergent-soluble (S) and -insoluble (I) fractions were prepared as described in the legend to Fig. 8, and aminopeptidase N was immunopurified from the two fractions and subjected to SDS-PAGE. Molecular weights (in kilodaltons) of transient, high mannose-glycosylated and mature, complex glycosylated forms (140 and 166 kDa, respectively) of aminopeptidase N are indicated. B, a quantification of the above gel tracks by PhosphorImager analysis. The percentage of mature, complex glycosylated form (M-form) and insoluble mature, complex glycosylated form (Ins) was determined for each time point.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 10Cholesterol depletion reduces apical surface expression. Cholesterol-depleted (CD/lova) and control (Control) mucosal explants of comparable size were labeled with 0.5 mCi/ml [35S]methionine for 2 h and fractionated into Mg2+-precipitated (Mg) and microvillar (Mic) membranes. The membrane fractions were resuspended in 0.5 ml of 25 mm HEPES-HCl, 150 mm NaCl, pH 7.0, and solubilized by 1% Triton X-100 for 10 min at 37 °C. Aminopeptidase N was immunopurified from the solubilized membrane fractions and analyzed by SDS-PAGE. Molecular masses (in kilodaltons) are indicated.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Methyl-β-cyclodextrin treatment has previously been reported to be a rapid and efficient method to remove selectively cholesterol from the plasma membrane of living cells (35.Kilsdonk E.P.C. Yancey P.G. Stoudt G.W. Bangerter F.W. Johnson W.J. Phillips M.C. Rothblat G.H. J. Biol. Chem. 1995; 270: 17250-17256Abstract Full Text Full Text PDF PubMed Scopus (706) Google Scholar, 36.Yancey P.G. Rodrigueza W.V. Kilsdonk E.P.C. Stoudt G.W. Johnson W.J. Phillips M.C. Rothblat G.H. J. Biol. Chem. 1996; 271: 16026-16034Abstract Full Text Full Text PDF PubMed Scopus (389) Google Scholar, 37.Neufeld 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). Used either on its own or in combination with inhibitors of hydroxymethylglutaryl-coenzyme A reductase, a removal of free cellular cholesterol by 50% or more is usually achieved within a few hours of treatment but is reportedly increased by shaking of cells (in monolayer) or by placing cells in suspension (36.Yancey P.G. Rodrigueza W.V. Kilsdonk E.P.C. Stoudt G.W. Johnson W.J. Phillips M.C. Rothblat G.H. J. Biol. Chem. 1996; 271: 16026-16034Abstract Full Text Full Text PDF PubMed Scopus (389) Google Scholar). Barred from these options, we consider the 50% reduction achieved by treatment for 2 h in the present work acceptable and probably near-optimal without compromising explant viability, as assessed by morphology as well as by protein biosynthesis. The enterocyte brush border membrane is rich in cholesterol, reportedly containing almost equimolar amounts of cholesterol and phospholipids (38.Wang H. Dudley Jr, A.W. Dupont J. Reeds P.J. Hachey D.L. Dudley M.A. J. Nutr. 1996; 126: 1455-1462Crossref PubMed Scopus (15) Google Scholar), and is the cell surface from where the main cholesterol efflux was observed to take place. This may account for the reduction in microvillar length and density caused by cholesterol depletion, but despite this the overall brush border architecture remained intact. In fact, the only intracellular membrane compartment visibly affected was the Golgi complex/trans-Golgi network, the central organelle in secretory membrane traffic and sorting (39.Farquhar M.G. Palade G.E. Trends Cell Biol. 1998; 8: 2-10Abstract Full Text PDF PubMed Scopus (219) Google Scholar, 40.Mazzarello P. Bentaivoglio M. Nature. 1998; 392: 543-544Crossref PubMed Scopus (19) Google Scholar). Its complex functioning depends on small membrane-bound transport intermediates that shuttle proteins into and out of the stack of cisternae and associated tubules and vesicles. This membrane flux thereby determines the size and location of the Golgi complex, the dynamics of which has recently been magnificently visualized in a video essay compilation, making use of the green fluorescent protein technology (41.Presley J.F. Smith C. Hirschberg K. Miller C. Cole N.B. Zaal K.J.M. Lippincott-Schwartz J. Mol. Biol. Cell. 1998; 9: 1617-1626Crossref PubMed Scopus (74) Google Scholar). Our observation that cholesterol depletion caused the appearance of a large number of Golgi-derived, aminopeptidase N-positive vesicles thus implies a disturbance of the normal steady-state membrane flux through the organelle. Likewise, the markedly increased localization of the brush border enzyme in the basolateral plasma membrane points to an erratic sorting in the trans-Golgi network. Since the Na+/K+-ATPase was not observed to be similarly mistargeted to the brush border, cholesterol depletion seems to affect primarily the apical membrane traffic. The incomplete removal of cholesterol by the treatment with methyl-β-cyclodextrin and lovastatin may explain why a less than complete block of carbohydrate processing, raft formation, and apical transport was achieved. Furthermore, by studies on model membranes, it has recently been demonstrated that raft formation is not prevented by removal of cholesterol in membranes having high concentrations (33%) of sphingolipids (6.Schroeder R.J. Ahmed S.N. Zhu Y. London E. Brown D.A. J. Biol. Chem. 1998; 273: 1150-1157Abstract Full Text Full Text PDF PubMed Scopus (375) Google Scholar). Our results may reflect this fact since the sphingolipid content of pig intestinal brush border membranes is of similar magnitude (42.Christiansen K. Carlsen J. Biochim. Biophys. Acta. 1981; 647: 188-195Crossref PubMed Scopus (87) Google Scholar). At any rate, the results obtained argue in favor of a close relationship in the enterocyte between raft formation in the Golgi complex/trans-golgi network and apical exocytotic membrane trafficking. Since a similar relationship has recently been shown to exist in MDCK cells (22.Keller P. Simons K. J. Cell Biol. 1998; 140: 1357-1367Crossref PubMed Scopus (472) Google Scholar) and cultured neurons (23.Ledesma M.D. Simons K. Dotti C.G. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 3966-3971Crossref PubMed Scopus (199) Google Scholar), evidence is accumulating that rafts play a general role as lateral sorting platforms for apically destined membrane traffic. The long standing question concerning the actual size of rafts in vivo has been addressed by two different approaches (43.Jacobson K. Dietrich C. Trends Cell Biol. 1999; 9: 87-91Abstract Full Text Full Text PDF PubMed Scopus (378) Google Scholar). By fluorescence resonance energy transfer and chemical cross-linking studies on GPI-anchored proteins, raft domains at the cell surface were estimated to be less than 70 nm in diameter (44.Varma R. Mayor S. Nature. 1998; 394: 798-801Crossref PubMed Scopus (1031) Google Scholar) and to harbor only about 15 protein molecules per raft (45.Friedrichson T. Kurchalia T.V. Nature. 1998; 394: 802-805Crossref PubMed Scopus (480) Google Scholar). This is a much smaller size than rafts prepared by low temperature detergent extraction and well below the resolution power of fluorescence microscopy, explaining why clustering of GPI-anchored proteins at the cell surface in raft domains so far has been difficult to visualize. Still, much about rafts remains to be sorted out, including the structural clues (other than GPI anchors) that signal raft association. Some evidence indicates that carbohydrate-lectin interactions (46.Scheiffele P. Peranen J. Simons K. Nature. 1995; 378: 96-98Crossref PubMed Scopus (417) Google Scholar), as well as the transmembrane domain (47.Scheiffele P. Roth M.G. Simons K. EMBO J. 1997; 16: 5501-5508Crossref PubMed Scopus (571) Google Scholar), can serve this function, suggesting that raft localization may occur by several mechanisms. Another unsolved question concerns the molecular machinery involved in the budding of raft-containing vesicles from the trans-Golgi network and their subsequent docking and fusion with the apical cell surface (48.Weimbs T. Low S.H. Chapin S.J. Mostov K.E. Trends Cell Biol. 1997; 7: 393-399Abstract Full Text PDF PubMed Scopus (98) Google Scholar). Ikonen et al.(49.Ikonen E. Tagaya M. Ullrich O. Montecucco C. Simons K. Cell. 1995; 81: 571-580Abstract Full Text PDF PubMed Scopus (222) Google Scholar) initially reported the apical trafficking in MDCK cells to occur independently of the SNARE mechanism and instead suggested a member of the annexin family, annexin XIIIb, to be involved. Other investigators have since described the transcytotic route from the basolateral to the apical surface of MDCK cells to rely on the SNARE machinery (50.Apodaca G. Cardone M.H. Whiteheart S.W. DasGupta B.R. Mostov K.E. EMBO J. 1996; 15: 1471-1481Crossref PubMed Scopus (71) Google Scholar) and have localized a member of the t-SNARE family, SNAP-23, to the apical (as well as the basolateral) surface of this cell type (51.Leung S.-M. Chen D. DasGupta B.R. Whiteheart S.W. Apodaca G. J. Biol. Chem. 1998; 273: 17732-17741Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). More recently, other components of the SNARE machinery (syntaxin 3 and TI-VAMP) have been identified in apical carrier vesicles (52.Lafont F. Verkade P. Galli T. Wimmer C. Louvard D. Simons K. Proc. Natl. Acad. Sci. U. S. A. 1999; 30: 3734-3738Crossref Scopus (208) Google Scholar), indicating that the raft-associated apical route from thetrans-Golgi network after all relies on the "classical" molecular machinery for docking and fusion. Obviously, however, more work is needed before we have a satisfactory understanding of apical membrane trafficking. We thank Professors Ove Norén and Hans Sjöström for valuable discussions on this manuscript.
Referência(s)