A GDP/GTP Exchange-stimulatory Activity for the Rab5-RabGDI Complex on Clathrin-coated Vesicles from Bovine Brain
1995; Elsevier BV; Volume: 270; Issue: 19 Linguagem: Inglês
10.1074/jbc.270.19.11257
ISSN1083-351X
AutoresHisanori Horiuchi, Angelika Giner, Bernard Hoflack, Marino Zerial,
Tópico(s)Calcium signaling and nucleotide metabolism
ResumoSmall GTPases of the Rab family are key regulators of intracellular transport. They are associated with the cytoplasmic surface of distinct exocytic and endocytic organelles and with transport vesicles connecting these compartments. Rab proteins are also present in the cytosol in the GDP-bound conformation complexed to Rab GDP dissociation inhibitor (RabGDI). Upon membrane association, RabGDI is released, and the Rab protein is converted into the GTP-bound form. In this paper we have investigated whether Rab5, which regulates the clathrin-coated vesicle-mediated pathway of endocytosis, can directly associate with the membrane of clathrin-coated vesicles (CCV) purified from bovine brain in vitro. We found that RabGDI can specifically deliver Rab5 but not Rab7, which is localized to late endosomes, to CCV. Furthermore, CCV contain a heat- and trypsin-sensitive activity that stimulates the dissociation of GDP from Rab5, but not from Rab7, and the subsequent binding of GTP. The activity was found to be associated with the CCV membrane but not with the coat components. CCV weakly stimulated GDP release from either post-translationally modified or unmodified Rab5 alone. However, maximal GDP dissociation stimulation required the presence of RabGDI, suggesting that the factor(s) responsible for the membrane association and GDP/GTP exchange of Rab5 recognize the protein complexed to RabGDI. These data demonstrate that CCV are competent for acquiring Rab5 and for converting the molecule into the GTP-bound active form. Small GTPases of the Rab family are key regulators of intracellular transport. They are associated with the cytoplasmic surface of distinct exocytic and endocytic organelles and with transport vesicles connecting these compartments. Rab proteins are also present in the cytosol in the GDP-bound conformation complexed to Rab GDP dissociation inhibitor (RabGDI). Upon membrane association, RabGDI is released, and the Rab protein is converted into the GTP-bound form. In this paper we have investigated whether Rab5, which regulates the clathrin-coated vesicle-mediated pathway of endocytosis, can directly associate with the membrane of clathrin-coated vesicles (CCV) purified from bovine brain in vitro. We found that RabGDI can specifically deliver Rab5 but not Rab7, which is localized to late endosomes, to CCV. Furthermore, CCV contain a heat- and trypsin-sensitive activity that stimulates the dissociation of GDP from Rab5, but not from Rab7, and the subsequent binding of GTP. The activity was found to be associated with the CCV membrane but not with the coat components. CCV weakly stimulated GDP release from either post-translationally modified or unmodified Rab5 alone. However, maximal GDP dissociation stimulation required the presence of RabGDI, suggesting that the factor(s) responsible for the membrane association and GDP/GTP exchange of Rab5 recognize the protein complexed to RabGDI. These data demonstrate that CCV are competent for acquiring Rab5 and for converting the molecule into the GTP-bound active form. Members of the Rab subfamily of Ras-related GTPases are distributed among exocytic and endocytic organelles and function as regulators of intracellular vesicle transport (1Goud B. McCaffrey M. Curr. Opin. Cell Biol. 1991; 3: 626-633Crossref PubMed Scopus (103) Google Scholar, 2Takai Y. Kaibuchi K. Kikuchi A. Kawata M. Int. Rev. Cytol. 1992; 133: 187-222Crossref PubMed Scopus (254) Google Scholar, 3Zerial M. Stenmark H. Cur. Opin. Cell Biol. 1993; 5: 613-620Crossref PubMed Scopus (343) Google Scholar, 4Novick P. Brennwald P. Cell. 1993; 75: 597-601Abstract Full Text PDF PubMed Scopus (314) Google Scholar, 5Simons K. Zerial M. Neuron. 1993; 11: 789-799Abstract Full Text PDF PubMed Scopus (256) Google Scholar, 6Pfeffer S.R. Curr. Opin. Cell Biol. 1994; 6: 522-526Crossref PubMed Scopus (295) Google Scholar). Rab proteins are involved in the process of docking/fusion of transport vesicles with their target compartments, although a role in the budding process cannot be excluded (7Nuoffer C. Davidson H.W. Matteson J. Meinkoth J. Balch W.E. J. Cell Biol. 1994; 125: 225-237Crossref PubMed Scopus (191) Google Scholar, 8Peter F. Nuoffer C. Pind S.N. Balch W.E. J. Cell Biol. 1994; 126: 1393-1406Crossref PubMed Scopus (71) Google Scholar). In the yeast secretory pathway, Ypt1p and Sec4p are sequentially required for the fusion of vesicles derived from the endoplasmic reticulum with the Golgi complex and of post-Golgi vesicles with the plasma membrane, respectively (9Salminen A. Novick P.J. Cell. 1987; 49: 527-538Abstract Full Text PDF PubMed Scopus (589) Google Scholar, 10Segev N. Mulholland J. Botstein D. Cell. 1988; 52: 915-924Abstract Full Text PDF PubMed Scopus (452) Google Scholar). Furthermore, Rab proteins may also regulate both heterotypic and homotypic fusion events. For example, Rab5 regulates the fusion of plasma membrane-derived clathrin-coated vesicles with early endosomes and the fusion between early endosomes (11Gorvel J.-P. Chavrier P. Zerial M. Gruenberg J. Cell. 1991; 64: 915-925Abstract Full Text PDF PubMed Scopus (854) Google Scholar, 12Bucci C. Parton R.G. Mather I.H. Stunnenberg H. Simons K. Hoflack B. Zerial M. Cell. 1992; 70: 715-728Abstract Full Text PDF PubMed Scopus (1110) Google Scholar, 13Li G. Stahl P.D. J. Biol. Chem. 1993; 268: 24475-24480Abstract Full Text PDF PubMed Google Scholar, 14Stenmark H. Parton R.G. Steele-Mortimer O. Lütcke A. Gruenberg J. Zerial M. EMBO J. 1994; 13: 1287-1296Crossref PubMed Scopus (762) Google Scholar). Rab proteins are synthesized as soluble proteins, but they are modified by addition of the 20-carbon isoprenoid geranylgeranyl to C-terminal cysteines (15Farnsworth C.C. Kawata M. Yoshida Y. Takai Y. Gelb M.H. Glomset J.A. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 6196-6200Crossref PubMed Scopus (160) Google Scholar, 16Ullrich O. Stenmark H. Alexandrov K. Huber L.A. Kaibuchi K. Sasaki T. Takai Y. Zerial M. J. Biol. Chem. 1993; 268: 18143-18150Abstract Full Text PDF PubMed Google Scholar). This reaction is catalyzed by a multisubunit enzyme, Rab geranylgeranyltransferase, composed of a catalytic heterodimer of α and β subunits, and a Rab escort protein (REP-1 and REP-2) 1The abbreviations used are: REPRab escort proteinGDIGDP dissociation inhibitorCCVclathrin-coated vesiclesCHAPS3-[(3cholamidopropyl) dimethylammonio]-1-propanesulfonic acidBSAbovine serum albuminGEFguanine nucleotide exchange factorGTPγSguanosine 5′-3-0-(thio)triphosphate1The abbreviations used are: REPRab escort proteinGDIGDP dissociation inhibitorCCVclathrin-coated vesiclesCHAPS3-[(3cholamidopropyl) dimethylammonio]-1-propanesulfonic acidBSAbovine serum albuminGEFguanine nucleotide exchange factorGTPγSguanosine 5′-3-0-(thio)triphosphate (17Seabra M.C. Brown M.S. Slaughter C.A. Südhof T.C. Goldstein J.L. Cell. 1992; 70: 1049-1057Abstract Full Text PDF PubMed Scopus (238) Google Scholar, 18Seabra M.C. Goldstein J.L. Südhof T.C. Brown M.S. J. Biol. Chem. 1992; 267: 14497-14503Abstract Full Text PDF PubMed Google Scholar, 19Andres D.A. Seabra M.C. Brown M.S. Armstrong S.A. Smeland T.E. Cremers F.P.M. Goldstein J.L. Cell. 1993; 73: 1091-1099Abstract Full Text PDF PubMed Scopus (283) Google Scholar, 20Cremers F.P.M. Armstrong S.A. Seabra M.C. Brown M.S. Goldstein J.L. J. Biol. Chem. 1994; 269: 2111-2117Abstract Full Text PDF PubMed Google Scholar). In vitro studies have suggested that REP-1 presents newly synthesized Rab proteins to the catalytic component of the enzyme and, after the transfer reaction, mediates the association of the prenylated proteins with their target membranes (19Andres D.A. Seabra M.C. Brown M.S. Armstrong S.A. Smeland T.E. Cremers F.P.M. Goldstein J.L. Cell. 1993; 73: 1091-1099Abstract Full Text PDF PubMed Scopus (283) Google Scholar, 21Alexandrov K. Horiuchi H. Steele-Mortimer O. Seabra M. Zerial M. EMBO J. 1994; 13: 5262-5273Crossref PubMed Scopus (196) Google Scholar). Rab proteins are also detected in the cytosol complexed to Rab GDP dissociation inhibitor (RabGDI) (22Araki S. Kikuchi A. Hata Y. Isomura M. Takai Y. J. Biol. Chem. 1990; 265: 13007-13015Abstract Full Text PDF PubMed Google Scholar). Recently, different GDI isoforms have been identified (23Shisheva A. Südhof T.C. Czech M.P. Mol. Cell. Biol. 1994; 14: 3459-3468Crossref PubMed Scopus (72) Google Scholar, 24Nishimura N. Nakamura H. Takai Y. Sano K. J. Biol. Chem. 1994; 269: 14191-14198Abstract Full Text PDF PubMed Google Scholar). RabGDI was first identified for its ability to inhibit GDP dissociation from Rab proteins (25Sasaki T. Kikuchi A. Araki S. Hata Y. Isomura M. Kuroda S. Takai Y. J. Biol. Chem. 1990; 265: 2333-2337Abstract Full Text PDF PubMed Google Scholar) and to solubilize GDP-bound Rab proteins from the membrane (16Ullrich O. Stenmark H. Alexandrov K. Huber L.A. Kaibuchi K. Sasaki T. Takai Y. Zerial M. J. Biol. Chem. 1993; 268: 18143-18150Abstract Full Text PDF PubMed Google Scholar, 22Araki S. Kikuchi A. Hata Y. Isomura M. Takai Y. J. Biol. Chem. 1990; 265: 13007-13015Abstract Full Text PDF PubMed Google Scholar, 26Regazzi R. Kikuchi A. Takai Y. Wollheim C.B. J. Biol. Chem. 1992; 267: 17512-17519Abstract Full Text PDF PubMed Google Scholar, 27Soldati T. Riederer M.A. Pfeffer S.R. Mol. Biol. Cell. 1993; 4: 425-434Crossref PubMed Scopus (123) Google Scholar). RabGDI was also found to mediate the reassociation of Rab proteins with the membrane, a process that is followed by GDP/GTP exchange (28Soldati T. Shapiro A.D. Dirac Svejstrup A.B. Pfeffer S.R. Nature. 1994; 369: 76-78Crossref PubMed Scopus (148) Google Scholar, 29Ullrich O. Horiuchi H. Bucci C. Zerial M. Nature. 1994; 368: 157-160Crossref PubMed Scopus (247) Google Scholar). RabGDI can interact only with prenylated Rab proteins (30Araki S. Kaibuchi K. Sasaki T. Hata Y. Takai Y. Mol. Celt. Biol. 1991; 11: 1438-1447Crossref PubMed Scopus (61) Google Scholar). In contrast, REP-1, which also acts as a GDP dissociation inhibitor, interacts with both prenylated and nonprenylated proteins (21Alexandrov K. Horiuchi H. Steele-Mortimer O. Seabra M. Zerial M. EMBO J. 1994; 13: 5262-5273Crossref PubMed Scopus (196) Google Scholar). Thus, REP-1 appears to accompany newly synthesized Rab proteins to their correct compartments, whereas RabGDI is responsible for the constitutive shuttling of Rab proteins between the cytosol and the membranes. Rab escort protein GDP dissociation inhibitor clathrin-coated vesicles 3-[(3cholamidopropyl) dimethylammonio]-1-propanesulfonic acid bovine serum albumin guanine nucleotide exchange factor guanosine 5′-3-0-(thio)triphosphate Rab escort protein GDP dissociation inhibitor clathrin-coated vesicles 3-[(3cholamidopropyl) dimethylammonio]-1-propanesulfonic acid bovine serum albumin guanine nucleotide exchange factor guanosine 5′-3-0-(thio)triphosphate According to current models (1Goud B. McCaffrey M. Curr. Opin. Cell Biol. 1991; 3: 626-633Crossref PubMed Scopus (103) Google Scholar, 2Takai Y. Kaibuchi K. Kikuchi A. Kawata M. Int. Rev. Cytol. 1992; 133: 187-222Crossref PubMed Scopus (254) Google Scholar, 3Zerial M. Stenmark H. Cur. Opin. Cell Biol. 1993; 5: 613-620Crossref PubMed Scopus (343) Google Scholar, 4Novick P. Brennwald P. Cell. 1993; 75: 597-601Abstract Full Text PDF PubMed Scopus (314) Google Scholar, 5Simons K. Zerial M. Neuron. 1993; 11: 789-799Abstract Full Text PDF PubMed Scopus (256) Google Scholar, 6Pfeffer S.R. Curr. Opin. Cell Biol. 1994; 6: 522-526Crossref PubMed Scopus (295) Google Scholar), Rab proteins bind to the membrane of the donor compartment and, having adopted the GTP-bound active conformation, are incorporated into nascent transport vesicles. Following vesicle fusion with the acceptor compartment, GTP is hydrolyzed, in a reaction catalyzed by a GTPase activating protein (GAP) (31Burstein E.S. Macara I.G. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 1154-1158Crossref PubMed Scopus (72) Google Scholar). At this point, the GDP-bound Rab proteins are solubilized by RabGDI and recycled via cytosol to the donor compartment for a new round of vesicular transport. While in most cases Rab proteins are detected on both donor and acceptor compartments (3Zerial M. Stenmark H. Cur. Opin. Cell Biol. 1993; 5: 613-620Crossref PubMed Scopus (343) Google Scholar), it is not clear whether they can directly associate with transport vesicles. The finding that mutated Rab proteins interfere with the budding process has suggested that the presence of a functional Rab protein is a prerequisite for vesicle formation (7Nuoffer C. Davidson H.W. Matteson J. Meinkoth J. Balch W.E. J. Cell Biol. 1994; 125: 225-237Crossref PubMed Scopus (191) Google Scholar, 8Peter F. Nuoffer C. Pind S.N. Balch W.E. J. Cell Biol. 1994; 126: 1393-1406Crossref PubMed Scopus (71) Google Scholar). However, the function of Ypt1p does not appear to be required for budding in yeast, and studies in vitro have suggested that endoplasmic reticulum-derived transport vesicles may acquire this GTPase en route to the Golgi apparatus (10Segev N. Mulholland J. Botstein D. Cell. 1988; 52: 915-924Abstract Full Text PDF PubMed Scopus (452) Google Scholar, 32Lian J.P. Ferro-Novick S. Cell. 1993; 73: 735-745Abstract Full Text PDF PubMed Scopus (121) Google Scholar, 33Barlowe C. Orci L. Yeung T. Hosobuchi M. Hamamoto S. Salama N. Rexach M.F. Ravazzola M. Amherdt M. Schekman R. Cell. 1994; 77: 895-907Abstract Full Text PDF PubMed Scopus (1025) Google Scholar). Further evidence for the association of these GTPases with transport intermediates was suggested by studies on Sec4p, which has been shown to be present on post-Golgi-derived transport vesicles in both wild-type and sec6–4 mutant cells (34Goud B. Salminen A. Walworth N.C. Novick P.J. Cell. 1988; 53: 753-768Abstract Full Text PDF PubMed Scopus (390) Google Scholar). Direct evidence for the binding of Rab proteins to transport vesicles in mammalian cells, however, has not been obtained to date. In this paper, we have investigated whether Rab5, which regulates the clathrin-coated pathway of internalization (12Bucci C. Parton R.G. Mather I.H. Stunnenberg H. Simons K. Hoflack B. Zerial M. Cell. 1992; 70: 715-728Abstract Full Text PDF PubMed Scopus (1110) Google Scholar), can associate with purified clathrin-coated vesicles (CCV) in vitro. First, we found that RabGDI can specifically mediate the association of Rab5 but not of Rab7, which is localized to late endosomes, with CCV. Second, we detected a Rab5-specific GDP/GTP exchange activity on CCV, which required the presence of the Rab5-RabGDI complex for maximal stimulation. These results demonstrate that Rab5 can bind to transport vesicles in vitro and that CCV contain one or more factors that promote release of RabGDI from Rab5 and conversion of the protein into the GTP-bound form. Materials and Chemicals—RabGDI was purified as a histidine-tagged protein from overexpressing E. coli by using Ni2+-agarose gel (Qiagen) as described (35Ullrich O. Horiuchi H. Alexandrov K. Zerial M. Methods Enzymol. 1995; (in press)Google Scholar). Post-translationally modified and unmodified Rab5 were purified to homogeneity from Sf9 cells infected with a Rab5-recombinant baculovirus as described (36Horiuchi H. Ullrich O. Bucci C. Zerial M. Methods Enzymol. 1995; (in press)Google Scholar). Briefly, modified Rab5 was extracted with 0.6% CHAPS (Sigma) from the membrane fraction of the Sf9 cells and purified by Mono Q (Pharmacia Biotech Inc.) column chromatography. Unmodified Rab5 was purified from the cytosolic fraction by hydroxyapatite (Seikagaku-kogyo, Japan) column chromatography. Consistent with previous results (30Araki S. Kaibuchi K. Sasaki T. Hata Y. Takai Y. Mol. Celt. Biol. 1991; 11: 1438-1447Crossref PubMed Scopus (61) Google Scholar), purified RabGDI efficiently inhibited GDP dissociation from modified Rab5 but did not affect that from unmodified Rab5 (21Alexandrov K. Horiuchi H. Steele-Mortimer O. Seabra M. Zerial M. EMBO J. 1994; 13: 5262-5273Crossref PubMed Scopus (196) Google Scholar). Rab7-overexpressing Sf9 cells were prepared and post-translationally modified Rab7 was purified essentially in the same way as for modified Rab5 except that Rab7 was eluted from the Mono Q column at about 100 mm NaCl. CCV were prepared from bovine brain by a series of high and low speed centrifugation steps followed by a single step gradient in 8% sucrose in D2O and Sephacryl S-1000 gel filtration column chromatography as described (37Sahagian G. Steer C.J. J. Biol. Chem. 1985; 260: 9838-9842Abstract Full Text PDF PubMed Google Scholar, 38Méresse S. Ludwig T. Frank R. Hoflack B. J. Biol. Chem. 1990; 265: 18833-18842Abstract Full Text PDF PubMed Google Scholar). Typically, this procedure yielded 15 mg of CCV from four bovine brains. The quality and purity of the CCV preparation was routinely inspected by negative staining and electron microscopy analysis. Trypsin treatment of CCV was performed by incubating CCV in the presence of 20% (weight) trypsin (Sigma) at 37 °C for various periods of time and then by adding 60% (weight) soybean trypsin inhibitor (Sigma) to stop the digestion. Control CCV were simultaneously incubated with 20% trypsin and 60% soybean trypsin inhibitor. Uncoating of the CCV was performed by incubation in the presence of 1 M Tris/HCl at pH 7.5 and 1 mm EDTA as described (38Méresse S. Ludwig T. Frank R. Hoflack B. J. Biol. Chem. 1990; 265: 18833-18842Abstract Full Text PDF PubMed Google Scholar) except that samples were incubated at 4 °C for 4 h. Soluble coat proteins were separated from uncoated vesicles by centrifugation at 280,000 × g at 4 °C for 30 min. Bovine brain membrane pellet and cytosol fractions were obtained by centrifugation of a bovine brain homogenate at 100,000 × g for 1 h. Dog pancreas microsomes were a gift of Dr. Bernhard Dobberstein (DKFZ, Heidelberg). Assay for [3H]GDP Dissociation from Rab Protein—The experiments of Fig. 1, Fig. 2 were carried out using purified [3H]GDP-Rab protein-RabGDI complex. [3H]GDP-Rab5-RabGDI complex was prepared as described (29Ullrich O. Horiuchi H. Bucci C. Zerial M. Nature. 1994; 368: 157-160Crossref PubMed Scopus (247) Google Scholar, 35Ullrich O. Horiuchi H. Alexandrov K. Zerial M. Methods Enzymol. 1995; (in press)Google Scholar). Briefly, [3H]GDP-bound post-translationally modified Rab5 was prepared as described (25Sasaki T. Kikuchi A. Araki S. Hata Y. Isomura M. Kuroda S. Takai Y. J. Biol. Chem. 1990; 265: 2333-2337Abstract Full Text PDF PubMed Google Scholar) and incubated with RabGDI to form a complex. The complex was then purified by Superose-12 (Pharmacia) gel filtration column chromatography in order to eliminate the detergent used during protein purification. [3H]GDP-Rab7-RabGDI complex was purified in the same way. [3H]GDP-Rab protein-RabGDI complex (50 nM, 6,000–8,000 cpm/assay) was incubated with various concentrations of CCV at 30 °C in the presence of 5 mm GTP and 0.0037% Triton X-100 unless specified. After the indicated period, [3H]GDP-Rab protein was measured by a filtration method (25Sasaki T. Kikuchi A. Araki S. Hata Y. Isomura M. Kuroda S. Takai Y. J. Biol. Chem. 1990; 265: 2333-2337Abstract Full Text PDF PubMed Google Scholar). Alternatively, for the experiments of Fig. 5, Fig. 6, [3H)GDP-bound posttranslationally modified and unmodified Rab5 were incubated at 30 °C with various concentrations of RabGDI in the presence of 2 mg/ml BSA, 10 mm GTP, and 20 mm MgCl2 for 10 min to allow the formation of a Rab protein-RabGDI complex. The mixture containing 0.1 μΜ [3H)GDP-Rab protein (8,000–10,000 cpm/assay) was then incubated with different amounts of CCV in the presence of 0.03% CHAPS. After the indicated period of time, [3H]GDP-Rab5 was measured as described above. Data from time course experiments are expressed as a ratio of the counts of [3H]GDP-Rab protein after incubation to counts before incubation, in percent. For CCV concentration-dependent activity, values of stimulated [3H]GDP dissociation are expressed as a ratio of [3H]GDP-Rab5 incubated without CCV-[3H]GDP-Rab5 with CCV to [3H]GDP-Rab5 before incubation, in percent.Fig. 2Effect of heat (A) and trypsin (B) treatment of CCV on the GDP dissociation activity on Rab5. A, CCV were first incubated at 95 °C for 3 min (lane 2) or 56 °C for 10 min (lane 3) or 20 min (lane 4) before assaying for stimulation of dissociation of GDP from Rab5. CCV without heat treatment were used as a control (lane 1). B, CCV were treated with trypsin (closed bars) or with trypsin together with trypsin inhibitor (open bars) at 37 °C for 0 min (lanes 1 and 2), 2 min (lanes 3 and 4), 5 min (lanes 5 and 6), and 10 min (lanes 7 and 8) as described under "Experimental Procedures." Pretreated CCV (0.5 mg/ml) were then incubated with [3H]GDP-Rab5-RabGDI for 20 min at 30 °C. Dissociation of [3H]GDP from Rab5 was measured as described under "Experimental Procedures." Results shown are representative of three independent experiments with similar results.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Fig. 5Effect of RabGDI on time-dependent CCV-stimulated GDP dissociation from Rab5. [3H]GDP-bound post-translationally modified Rab5 (A and B) and unmodified Rab5 (C and D) were incubated in the absence (A and C) or in the presence of 0.4 μΜ RabGDI (B and D). The samples were subsequently incubated in the absence (closed squares) or in the presence (open squares) of 0.6 mg/ml CCV for various periods of time as described under "Experimental Procedures." Data are shown as mean ± S.E. of the ratio of [3H]GDP-Rab5 after incubation to [3H]GDP-Rab5 before incubation in percent, from three independent experiments.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Fig. 6Low concentration of RabGDI enhances whereas high concentration inhibits CCV-stimulated GDP dissociation from Rab5. After incubation of 0.1 μΜ [3H]GDP-bound post-translationally modified Rab5 without and with low (A) and high (B) concentrations of RabGDI, the samples were incubated with various concentrations of CCV for 20 min as described under "Experimental Procedures." Rab-GDI concentrations were 0 μΜ (closed squares), 0.04 μΜ (open circles), 0.4 μΜ (open squares), 4 μΜ (closed circles), and 40 μΜ (closed triangles). Data are shown as mean ± S.E. of percentages of stimulated [3H]GDP dissociation (see "Experimental Procedures") from three independent experiments.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Assay for [35S]GTPγS Binding to Rab5 on CCV—GDP-Rab5-RabGDI was prepared in the same way as [3H)GDP-Rab5-RabGDI. GDP-Rab5-RabGDI (3.6 pmol, 72 nM) or buffer alone was incubated in the presence of 10 μΜ [35S]GTPγS (90,000 cpm/pmol) and 2.8 mm MgCl2 with 1.0 mg/ml CCV at 30 °C for the indicated periods of time. The exchange reactions were quenched by the addition of 10 mm GTP and 20 mm MgCl2. The samples were then centrifuged at 280,000 × g at 4 °C for 30 min. The CCV pellets were extracted in 1% Nonidet P-40, 20 mm HEPES/KOH, pH 7.2, 10 mm MgCl2, 40 mm NaCl, and 1 mm GTP at 4 ºC for 30 min. The samples were again centrifuged at 280,000 × g at 4 ºC for 30 min. Most of the Rab5 was recovered in the supernatant, and no Rab5 was detected in the pellet by Western blot analysis as described below. Rab5 was immunoprecipitated in the supernatant using monoclonal anti-Rab5 antibody (39Bucci C. Wandinger-Ness A. Lütcke A. Chiariello M. Bruni C.B. Zerial M. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 5061-5065Crossref PubMed Scopus (110) Google Scholar), and the Rab5-associated radioactivity was measured as described (29Ullrich O. Horiuchi H. Bucci C. Zerial M. Nature. 1994; 368: 157-160Crossref PubMed Scopus (247) Google Scholar, 35Ullrich O. Horiuchi H. Alexandrov K. Zerial M. Methods Enzymol. 1995; (in press)Google Scholar) except that beads were washed with the same buffer used for the extraction. As a control, the GDP-Rab5-RabGDI complex was incubated in parallel in the presence of [35S]GTPγS and in the absence of CCV under the same conditions, and the protein-associated radioactivity was measured directly without centrifugation by the filter method described above. Assay for Binding of Rab Protein to CCV—[3H]GDP-Rab5-RabGDI and [3H]GDP-Rab7-RabGDI complexes (each 50 nM) were incubated with or without CCV or dog pancreas microsomes at the indicated concentrations in the presence of 0.5 mm GTPγS in 126 μl (Fig. 4A) or 42 μl (Fig. 4B) at 30 °C for various periods of time. Aliquots were used to measure the release of [3H]GDP from Rab proteins, and the rest of the CCV samples were pelleted by centrifugation at 280,000 × g at 4 °C for 30 min. Pellets were electrophoretically separated and analyzed by Western blotting using monoclonal anti-Rab5 (39Bucci C. Wandinger-Ness A. Lütcke A. Chiariello M. Bruni C.B. Zerial M. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 5061-5065Crossref PubMed Scopus (110) Google Scholar) and polyclonal anti-Rab7 antibodies (40Chavrier P. Parton R.G. Hauri H.P. Simons K. Zerial M. Cell. 1990; 62: 317-329Abstract Full Text PDF PubMed Scopus (875) Google Scholar). The blot was visualized by ECL (Amersham Corp.) and quantified by densitometry. Protein concentrations were determined by densitometric scanning of Coomassie Blue-stained sodium dodecyl sulfate-polyacrylamide gel or by Bradford's method (Bio-Rad) using BSA as a standard. Previous studies have demonstrated that the association of Rab5 complexed to RabGDI with Streptolysin O-permeabilized cell membranes is accompanied by release of RabGDI into the cytosol and nucleotide exchange (29Ullrich O. Horiuchi H. Bucci C. Zerial M. Nature. 1994; 368: 157-160Crossref PubMed Scopus (247) Google Scholar). Given the regulatory role of Rab5 in clathrin-coated vesicle-mediated endocytosis (12Bucci C. Parton R.G. Mather I.H. Stunnenberg H. Simons K. Hoflack B. Zerial M. Cell. 1992; 70: 715-728Abstract Full Text PDF PubMed Scopus (1110) Google Scholar, 14Stenmark H. Parton R.G. Steele-Mortimer O. Lütcke A. Gruenberg J. Zerial M. EMBO J. 1994; 13: 1287-1296Crossref PubMed Scopus (762) Google Scholar), we investigated whether RabGDI could deliver the protein directly to CCV purified from bovine brain as described (37Sahagian G. Steer C.J. J. Biol. Chem. 1985; 260: 9838-9842Abstract Full Text PDF PubMed Google Scholar, 38Méresse S. Ludwig T. Frank R. Hoflack B. J. Biol. Chem. 1990; 265: 18833-18842Abstract Full Text PDF PubMed Google Scholar). The structural preservation and purity of these vesicles were confirmed morphologically by negative staining and electron microscopy analysis (data not shown). We first used the [3H]GDP-Rab5-RabGDI complex, i.e. the cytosolic form of Rab5, to measure the effect of CCV on the dissociation of GDP from Rab5 (Fig. 1A). Whereas the dissociation of [3H]GDP from Rab5 was low in the absence of CCV, the rate of [3H]GDP dissociation was stimulated in the presence of CCV. Identical results were obtained when experiments were performed in the presence of either 5 mm GTP or GDP (data not shown). This activity was specific for Rab5 since no stimulation of [3H]GDP dissociation was detected for Rab7 (Fig. 1B), a GTPase localized to late endosomes (40Chavrier P. Parton R.G. Hauri H.P. Simons K. Zerial M. Cell. 1990; 62: 317-329Abstract Full Text PDF PubMed Scopus (875) Google Scholar). CCV efficiently stimulated [3H]GDP dissociation from Rab5 in a concentration-dependent manner (Fig. 1C). Weak stimulation was detected in a bovine brain crude membrane fraction, but no activity was detected in either bovine brain cytosol or an endoplasmic reticulum enriched fraction (dog pancreas microsomes) or in the presence of BSA. These data suggest that the Rab5 GDP dissociation activity was associated with membranes and enriched in the CCV preparation. This activity was temperature-dependent since it was undetectable at 4 °C (data not shown). It was also sensitive to heat since incubation of CCV at 95 °C for 3 min or at 56 °C for 20 min abrogated the stimulation of GDP dissociation (Fig. 2A). Furthermore, it was trypsin-sensitive, thus indicating its proteinaceous nature (Fig. 2B). To determine whether stimulation of GDP release was accompanied by GTP binding, CCV were incubated with or without GDP-Rab5 complexed with RabGDI in the presence of [35S]GTPγS for different periods of time. The Rab5-associated radioactivity was measured after immunoprecipitation, and the results are shown in Fig. 3. When the complex was incubated in the absence of CCV very little [35S]GTPγS was detected bound to Rab5. In contrast, the presence of CCV strongly stimulated the binding of [35S]GTPγS to Rab5 in a time-dependent manner. Twenty-five and 34% of exogenously added Rab5 bound the labeled nucleotide after 15- and 30-min incubations, respectively. The observation that [35S]GTPγS-Rab5 was detected associated with the CCV (see "Experimental Procedures") suggests that GDP/GTP exchange takes place upon binding of Rab5 to the membrane. We confirmed this point by inspecting the binding of Rab5 to CCV by Western blot analysis (Fig. 4). Endogenous Rab5 but not Rab7 was detected in the CCV preparation (Fig. 4A, lanes 4 and 9). Consistent with the lack of GDP dissociation activity for Rab7, very low binding of this GTPase to CCV was detected (Fig. 4A, lane 10). In contrast, Rab5 efficiently (30% of added Rab5) associated with the membrane of CCV (Fig. 4A, lane 5). This value corresponds to an 8-fold enrichment over the endogenous level. Concomitantly, GDP dissociation was not stimulated for Rab7, whereas there was a 31% increase in GDP release from Rab5 in the presence of CCV. These values are in good agreement with the observed efficiency of GDP/GTP exchange. In addition, membrane binding was time-dependent, with 19 and 30% of exogenous Rab5 binding CCV after 7- and 15-min incubations, respectively (Fig. 4B, lanes 1–4). In contrast, only low amounts of Rab5 bound microsomal membranes, and (most importantly) this fraction did not increase with time (Fig. 4B, lanes 5–8). These data demonstrate that GDP/GTP exchange is coupled to attachment of Rab5 to the membrane of CCV. To investigate whether the activity was due to the constituents of the coat (clathrin and underlying adaptors), these were stripped from CCV by incubation in the presence of 1 m Tris/HCl, pH 7.5, and 1 mM EDTA for 4 h at 4 °C (38Méresse S. Ludwig T. Frank R. Hoflack B. J. Biol. Chem. 1990; 265: 18833-18842Abstract Full Text PDF PubMed Google Scholar). Soluble coat proteins were separated from vesicles by centrifugation at high speed, and the stimulation of [3H]GDP dissociation from Rab5 complexed to RabGDI was measured in the pellet and the supernatant fractions. Sixty percent of total [3H]GDP dissociation activity was recovered in the vesicle pellet, which accounted for 30% of the total protein content of CCV, whereas no activity was detected in the coat proteins fraction (Fig. 4C). These results demonstrate that the activity that elicits GDP dissociation from Rab5 is not part of the coat constituents but is associated with the CCV membrane. RabGDI is essential for the correct delivery of Rab proteins to their target membranes (41Dirac-Svejstrup A.B. Soldati T. Shapiro A.D. Pfeffer S.R. J. Biol. Chem. 1994; 269: 15427-15430Abstract Full Text PDF PubMed Google Scholar). We next investigated whether RabGDI is also required for the guanine nucleotide exchange reaction that takes place on the membrane (28Soldati T. Shapiro A.D. Dirac Svejstrup A.B. Pfeffer S.R. Nature. 1994; 369: 76-78Crossref PubMed Scopus (148) Google Scholar, 29Ullrich O. Horiuchi H. Bucci C. Zerial M. Nature. 1994; 368: 157-160Crossref PubMed Scopus (247) Google Scholar). Posttranslationally modified (geranylgeranylated) [3H]GDP-bound Rab5 was solubilized in 0.03% CHAPS (see "Experimental Procedures"). At this concentration the detergent does not affect the interaction with RabGDI (data not shown). The protein was incubated in the presence or absence of RabGDI, and the effect of CCV on the dissociation of the nucleotide was measured. In the absence of RabGDI addition of CCV weakly stimulated the dissociation of [3H]GDP from Rab5 in a time-dependent manner (Fig. 5A). In contrast, the presence of RabGDI markedly accelerated the rate of dissociation of the nucleotide bound to Rab5 (Fig. 5B). Evidence that the activity required the formation of a complex between Rab5-GDP and RabGDI was provided by the weak stimulation observed when post-translationally unmodified Rab5, which cannot complex with RabGDI (30Araki S. Kaibuchi K. Sasaki T. Hata Y. Takai Y. Mol. Celt. Biol. 1991; 11: 1438-1447Crossref PubMed Scopus (61) Google Scholar), was assayed in the presence of RabGDI (Fig. 5D). Interestingly, low but significant stimulation of GDP dissociation from Rab5 was detected when both the modified and unmodified protein were incubated with CCV in the absence of RabGDI (Fig. 5, A and C). These data indicate that CCV weakly promote GDP dissociation from Rab5 when this GTPase is not complexed to RabGDI but that maximal stimulation of GDP dissociation from Rab5 requires the presence of a Rab5-RabGDI complex. We then examined the concentration-dependent effect of RabGDI on the dissociation of GDP from Rab5 in the presence of various amounts of CCV (Fig. 6). Under all conditions, posttranslationally modified Rab5 was present at a concentration of 0.1 μΜ, and the samples were incubated at 30 °C for 20 min. The stimulation of dissociation of [3H]GDP from Rab5 by CCV was both concentration-dependent and saturable (Fig. 6A). Addition of increasing amounts of RabGDI increased the release of [3H]GDP at all CCV concentrations, and maximal stimulation was observed in the presence of 0.4 μΜ RabGDI (Fig. 6A). However, higher concentrations of RabGDI (4 μΜ) inhibited the GDP dissociation activity of CCV (Fig. 6B). The inhibitoryeffect depended upon the concentration of CCV present in the mixture and was neutralized in the presence of 1.2 mg/ml CCV. Addition of a large excess of RabGDI (40 μΜ) completely inhibited the stimulation of [3H]GDP release from Rab5 by CCV (Fig. 6B). These data further demonstrate that CCV contain a saturable activity that stimulates only weakly the dissociation of GDP from Rab5 alone but efficiently from the Rab5-RabGDI complex. Furthermore, high levels of RabGDI inhibit this activity. The results presented in this study indicate that RabGDI can mediate the association of Rab5 with CCV in vitro. This process was selective since Rab7, which is not present in the early endocytic pathway, remained in a soluble complex with Rab-GDI. A specific GDP dissociation activity for Rab5 was detected in the CCV membrane but not in the coat components. Release of GDP was found to be accompanied by binding of GTP. The heat and trypsin sensitivity of the activity indicated the involvement of a protein(s). The activity was enriched in CCV compared with total membranes and was undetectable in cytosol or endoplasmic reticulum microsomes that exhibited low and nonspecific binding of Rab5. As for membrane association, very little activity was detected for the Rab7-RabGDI complex. CCV weakly stimulated GDP release from either post-translationally modified or unmodified Rab5. A potent stimulation of GDP release was only observed when Rab5 was complexed to RabGDI. Several lines of evidence support the idea that Rab5 regulates the clathrin-coated pathway of internalization. First, Rab5 is localized to the plasma membrane, clathrin-coated vesicles, and early endosomes (11Gorvel J.-P. Chavrier P. Zerial M. Gruenberg J. Cell. 1991; 64: 915-925Abstract Full Text PDF PubMed Scopus (854) Google Scholar, 12Bucci C. Parton R.G. Mather I.H. Stunnenberg H. Simons K. Hoflack B. Zerial M. Cell. 1992; 70: 715-728Abstract Full Text PDF PubMed Scopus (1110) Google Scholar, 14Stenmark H. Parton R.G. Steele-Mortimer O. Lütcke A. Gruenberg J. Zerial M. EMBO J. 1994; 13: 1287-1296Crossref PubMed Scopus (762) Google Scholar, 40Chavrier P. Parton R.G. Hauri H.P. Simons K. Zerial M. Cell. 1990; 62: 317-329Abstract Full Text PDF PubMed Scopus (875) Google Scholar). Although Jahn and colleagues (42Stahl B. Fisher von Mollard G. Walch-Solimena C. Jahn R. J. Biol. Chem. 1994; 269: 24770-24776Abstract Full Text PDF PubMed Google Scholar) have recently reported that clathrin-coated vesicles lack Rab5, our study (Fig. 4) confirms our previous findings that Rab5 is indeed present, although at low concentration, in the CCV from bovine brain (12Bucci C. Parton R.G. Mather I.H. Stunnenberg H. Simons K. Hoflack B. Zerial M. Cell. 1992; 70: 715-728Abstract Full Text PDF PubMed Scopus (1110) Google Scholar). Second, an additional argument that validates the localization of Rab5 to CCV is the ability of this GTPase to stimulate the pathway of clathrin-coated vesicle-mediated endocytosis when overexpressed in vivo (12Bucci C. Parton R.G. Mather I.H. Stunnenberg H. Simons K. Hoflack B. Zerial M. Cell. 1992; 70: 715-728Abstract Full Text PDF PubMed Scopus (1110) Google Scholar, 13Li G. Stahl P.D. J. Biol. Chem. 1993; 268: 24475-24480Abstract Full Text PDF PubMed Google Scholar, 14Stenmark H. Parton R.G. Steele-Mortimer O. Lütcke A. Gruenberg J. Zerial M. EMBO J. 1994; 13: 1287-1296Crossref PubMed Scopus (762) Google Scholar). Thus, the association of Rab5 with CCV is consistent with the function of the protein. Third, Rab5 can specifically associate with purified CCV in vitro. The regulatory role of Rab5 in the endocytic pathway suggests that this protein binds to the fraction of vesicles that originated from the plasma membrane rather than those derived from the Golgi complex. Thus, these vesicles are competent for acquiring Rab5 and for converting the molecule into the GTP-bound active form. This is, to our knowledge, the first direct evidence for the binding of a Rab protein to transport intermediates after their formation. Our data further demonstrate that Rab5 is rate-limiting in CCV and suggest that overexpression of Rab5 may increase the fraction of this GTPase in the vesicles, thereby accelerating the kinetics of fusion with the early endosomes. Fourth, this hypothesis is corroborated by recent data that demonstrate a role of Rab5 in the fusion of plasma membrane-derived clathrin-coated vesicles with early endosomes in vitro. 2O. Ullrich et al., manuscript in preparation. Studies in vitro have shown that RabGDI regulates the membrane association of several different Rab proteins (16Ullrich O. Stenmark H. Alexandrov K. Huber L.A. Kaibuchi K. Sasaki T. Takai Y. Zerial M. J. Biol. Chem. 1993; 268: 18143-18150Abstract Full Text PDF PubMed Google Scholar, 22Araki S. Kikuchi A. Hata Y. Isomura M. Takai Y. J. Biol. Chem. 1990; 265: 13007-13015Abstract Full Text PDF PubMed Google Scholar, 26Regazzi R. Kikuchi A. Takai Y. Wollheim C.B. J. Biol. Chem. 1992; 267: 17512-17519Abstract Full Text PDF PubMed Google Scholar, 27Soldati T. Riederer M.A. Pfeffer S.R. Mol. Biol. Cell. 1993; 4: 425-434Crossref PubMed Scopus (123) Google Scholar, 43Sasaki T. Kaibuchi K. Kabcenell A.K. Novick P.J. Takai Y. Mol. Cell. Biol. 1991; 11: 2909-2912Crossref PubMed Scopus (53) Google Scholar). Targeting of Rab proteins to their specific compartments is determined by sequences in the hypervariable C-terminal region (44Chavrier P. Gorvel J.-P. Steltzer E. Simons K. Gruenberg J. Zerial M. Nature. 1991; 353: 769-772Crossref PubMed Scopus (315) Google Scholar, 45Brennwald P. Novick P. Nature. 1993; 362: 560-563Crossref PubMed Scopus (154) Google Scholar, 46Stenmark H. Valencia A. Martinez O. Ullrich O. Goud B. Zerial M. EMBO J. 1994; 13: 575-583Crossref PubMed Scopus (146) Google Scholar). However, despite RabGDI being a general regulator of Rab proteins, it also contributes to the selectivity of this process, since substitution of RabGDI with BSA led to unspecific membrane binding and spontaneous GDP/GTP exchange of Rab9 (41Dirac-Svejstrup A.B. Soldati T. Shapiro A.D. Pfeffer S.R. J. Biol. Chem. 1994; 269: 15427-15430Abstract Full Text PDF PubMed Google Scholar). We here demonstrate that RabGDI is also required for the stimulation of GDP dissociation from Rab5, which follows the attachment of the protein to the CCV membrane. Three observations suggest that the factors that mediate membrane association and GDP/GTP exchange of Rab5 preferentially recognize the protein complexed to RabGDI. First, post-translationally modified GDP-Rab5 alone was a poor substrate for the GDP dissociation-stimulatory activity of CCV, suggesting that the GDP-bound conformation is not sufficient for interaction with these factors. Second, regardless of the presence of RabGDI, only low stimulation of GDP release was detected for post-translationally unmodified Rab5, which cannot complex with RabGDI. Third, in agreement with previous findings (16Ullrich O. Stenmark H. Alexandrov K. Huber L.A. Kaibuchi K. Sasaki T. Takai Y. Zerial M. J. Biol. Chem. 1993; 268: 18143-18150Abstract Full Text PDF PubMed Google Scholar, 27Soldati T. Riederer M.A. Pfeffer S.R. Mol. Biol. Cell. 1993; 4: 425-434Crossref PubMed Scopus (123) Google Scholar, 28Soldati T. Shapiro A.D. Dirac Svejstrup A.B. Pfeffer S.R. Nature. 1994; 369: 76-78Crossref PubMed Scopus (148) Google Scholar), an excess of RabGDI inhibited the GDP dissociation activity. One possible explanation for this effect is that RabGDI interacts with the membrane factors in a competitive manner. We cannot, however, rule out the possibility that this inhibition is due to a rapid retrieval of Rab proteins prior to nucleotide exchange. Membrane association of Rab proteins is accompanied by GDP/GTP exchange (21Alexandrov K. Horiuchi H. Steele-Mortimer O. Seabra M. Zerial M. EMBO J. 1994; 13: 5262-5273Crossref PubMed Scopus (196) Google Scholar, 28Soldati T. Shapiro A.D. Dirac Svejstrup A.B. Pfeffer S.R. Nature. 1994; 369: 76-78Crossref PubMed Scopus (148) Google Scholar, 29Ullrich O. Horiuchi H. Bucci C. Zerial M. Nature. 1994; 368: 157-160Crossref PubMed Scopus (247) Google Scholar, this study). It is not clear whether one or more factors mediate these distinct events. The finding that targeting of Rab proteins to their compartments involves the formation of a transient GDP-bound intermediate on the membrane, after the release of RabGDI and prior to GDP/GTP exchange, would support the second possibility. It is interesting in this respect that previously identified GDP/GTP exchange factors (GEFs) for Rab proteins do not require the presence of RabGDI. Dss4p and Mss4 stimulate GDP/GTP exchange on unmodified Sec4p and Rab3A (47Moya M. Roberts D. Novick P. Nature. 1993; 361: 460-463Crossref PubMed Scopus (90) Google Scholar, 48Burton J. Roberts D. Montaldi M. Novick P. De Camilli P. Nature. 1993; 361: 464-467Crossref PubMed Scopus (98) Google Scholar). Furthermore, RabGDI was shown to antagonize the guanine nucleotide releasing factor of Rab3A (49Burstein E.S. Brondyk W.H. Macara I.G. Kaibuchi K. Takai Y. J. Biol. Chem. 1993; 268: 22247-22250Abstract Full Text PDF PubMed Google Scholar) at a concentration similar to that eliciting GDP dissociation from Rab5 by CCV. Given the requirement for RabGDI, it thus appears that the GDP dissociation activity associated with CCV is distinct from the exchange factors for Rab proteins identified so far. Alternatively, the low but significant GDP dissociation stimulation of CCV on Rab5 in the absence of RabGDI may reflect the presence of a GEF distinct from the molecule that interacts with the Rab5-RabGDI complex and promotes its disassembly. RabGDI may position Rab5 on the membrane and make it competent to efficiently interact with the GEF, Upon nucleotide exchange, Rab5 would then become unavailable for removal by RabGDI. The identification of the factors mediating the membrane association of Rab5 will be essential to test this hypothesis. In this respect, CCV appear to be a valuable source of material from which these activities can be purified. We thank Dr. B. Dobberstein for the gift of dog pancreas microsomes and Drs. H. Stenmark and C. Murphy for critical reading of the manuscript.
Referência(s)