Portal de Periódicos da CAPES

Overexpression of an ADP-ribosylation Factor-Guanine Nucleotide Exchange Factor, BIG2, Uncouples Brefeldin A-induced Adaptor Protein-1 Coat Dissociation and Membrane Tubulation

2002; Elsevier BV; Volume: 277; Issue: 11 Linguagem: Inglês

10.1074/jbc.m112427200

1083-351X

Chisa Shinotsuka, Yusaku Yoshida, Kazumasa Kawamoto, Hiroyuki Takatsu, Kazuhisa Nakayama,

Calcium signaling and nucleotide metabolism

BIG2 is a guanine nucleotide exchange factor (GEF) for the ADP-ribosylation factor (ARF) family of small GTPases, which regulate membrane association of COPI and adaptor protein (AP)-1 coat protein complexes. A fungal metabolite, brefeldin A (BFA), inhibits ARF-GEFs and leads to redistribution of coat proteins from membranes to the cytoplasm and membrane tubulation of the Golgi complex and the trans-Golgi network (TGN). To investigate the function of BIG2, we examined the effects of BIG2-overexpression on the BFA-induced redistribution of ARF, coat proteins, and organelle markers. The BIG2 overexpression blocked BFA-induced redistribution from membranes of ARF1 and the AP-1 complex but not that of the COPI complex. These observations indicate that BIG2 is implicated in membrane association of AP-1, but not that of COPI, through activating ARF. Furthermore, not only BIG2 but also ARF1 and AP-1 were found as queues of spherical swellings along the BFA-induced membrane tubules emanating from the TGN. These observations indicate that BFA-induced AP-1 dissociation from TGN membranes and tubulation of TGN membranes are not coupled events and suggest that a BFA target other than ARF-GEFs exists in the cell. BIG2 is a guanine nucleotide exchange factor (GEF) for the ADP-ribosylation factor (ARF) family of small GTPases, which regulate membrane association of COPI and adaptor protein (AP)-1 coat protein complexes. A fungal metabolite, brefeldin A (BFA), inhibits ARF-GEFs and leads to redistribution of coat proteins from membranes to the cytoplasm and membrane tubulation of the Golgi complex and the trans-Golgi network (TGN). To investigate the function of BIG2, we examined the effects of BIG2-overexpression on the BFA-induced redistribution of ARF, coat proteins, and organelle markers. The BIG2 overexpression blocked BFA-induced redistribution from membranes of ARF1 and the AP-1 complex but not that of the COPI complex. These observations indicate that BIG2 is implicated in membrane association of AP-1, but not that of COPI, through activating ARF. Furthermore, not only BIG2 but also ARF1 and AP-1 were found as queues of spherical swellings along the BFA-induced membrane tubules emanating from the TGN. These observations indicate that BFA-induced AP-1 dissociation from TGN membranes and tubulation of TGN membranes are not coupled events and suggest that a BFA target other than ARF-GEFs exists in the cell. ADP-ribosylation factor brefeldin A brefeldin A-ADP-ribosylated substrate expressed sequence tag guanine nucleotide exchange factor Golgi-localizing, γ-adaptin ear homology domain ARF-binding protein GGA homology glutathione S-transferase hemagglutinin mannosidase II mannose 6-phosphate receptor trans-Golgi network Membrane traffic between intracellular organelles in eukaryotic cells is mediated primarily by small vesicles that bud from a donor compartment and fuse with a target compartment to deliver cargo molecules. The budding of carrier vesicles is triggered by membrane binding of the ADP-ribosylation factor (ARF)1 and Sar1 families of small GTPases (reviewed in Refs. 1Moss J. Vaughan M. J. Biol. Chem. 1995; 270: 12327-12330Abstract Full Text Full Text PDF PubMed Scopus (272) Google Scholar, 2Schmid S.L. Damke H. FASEB J. 1995; 9: 1445-1453Crossref PubMed Scopus (63) Google Scholar, 3Springer S. Spang A. Schekman R. Cell. 1999; 97: 145-148Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar). ARFs are found in all eukaryotes and have been shown to function in a variety of membrane-trafficking events. In mammals, the ARF family members are divided into three classes based on sequence similarity, Class I (ARF1-ARF3), Class II (ARF4 and ARF5), and Class III (ARF6) (reviewed in Refs. 1Moss J. Vaughan M. J. Biol. Chem. 1995; 270: 12327-12330Abstract Full Text Full Text PDF PubMed Scopus (272) Google Scholar and 4Chavrier P. Goud B. Curr. Opin. Cell Biol. 1999; 11: 466-475Crossref PubMed Scopus (422) Google Scholar). The Class I ARFs, especially ARF1, have been most extensively studied and shown to regulate the assembly of coat protein complexes onto vesicle-budding sites including the COPI complex on the Golgi complex, the AP-1 clathrin adaptor complex on thetrans-Golgi network (TGN), and the AP-3 complex on endosomes (5Hirst J. Robinson M.S. Biochim. Biophys. Acta. 1998; 1404: 173-193Crossref PubMed Scopus (327) Google Scholar, 6Lippincott-Schwartz J. Cole N.B. Donaldson J.G. Histochem. Cell Biol. 1998; 109: 449-462Crossref PubMed Scopus (70) Google Scholar). The only Class III member, ARF6, functions in endosome-plasma membrane recycling system and in the remodeling of actin cytoskeleton (7Al-Awar O. Radhakrishna H. Powell N.N. Donaldson J.G. Mol. Cell. Biol. 2000; 20: 5998-6007Crossref PubMed Scopus (70) Google Scholar). Little is known concerning the roles of the Class II ARFs. Similar to other GTP-binding proteins, ARF cycles between a GDP-bound inactive state and a GTP-bound active state. GDP-bound ARF is primarily cytosolic although weakly associated with membranes, whereas the active GTP-bound form binds tightly to membranes in which it encounters effectors. For instance, ARF1-GTP promotes the formation of COPI-coated and clathrin/AP-1-coated vesicles at the pre-Golgi intermediates and at the TGN, respectively, by inducing an assembly of the coat complexes onto membranes and possibly by activating lipid mediators (8Stamnes M.A. Rothman J.E. Cell. 1993; 73: 999-1005Abstract Full Text PDF PubMed Scopus (340) Google Scholar, 9Traub L.M. Ostrom J.A. Kornfeld S. J. Cell Biol. 1993; 123: 561-573Crossref PubMed Scopus (252) Google Scholar, reviewed in Refs. 1Moss J. Vaughan M. J. Biol. Chem. 1995; 270: 12327-12330Abstract Full Text Full Text PDF PubMed Scopus (272) Google Scholar, 4Chavrier P. Goud B. Curr. Opin. Cell Biol. 1999; 11: 466-475Crossref PubMed Scopus (422) Google Scholar, and 10Roth M.G. Trends Cell Biol. 1999; 9: 174-179Abstract Full Text Full Text PDF PubMed Scopus (99) Google Scholar). Furthermore, brefeldin A (BFA), which blocks guanine nucleotide exchange on ARF, inhibits the assembly of the COPI and AP-1 complexes onto membranes and induce the formation of membrane tubules from the Golgi complex and the TGN that subsequently fuse with the endoplasmic reticulum and with endosomes, respectively (11–14, reviewed in Refs. 6Lippincott-Schwartz J. Cole N.B. Donaldson J.G. Histochem. Cell Biol. 1998; 109: 449-462Crossref PubMed Scopus (70) Google Scholar, 15Klausner R.D. Donaldson J.G. Lippincott-Schwartz J. J. Cell Biol. 1992; 116: 1071-1080Crossref PubMed Scopus (1542) Google Scholar, and 16Jackson C.L. Subcell. Biochem. 2000; 34: 233-272Crossref PubMed Google Scholar). Therefore, it has been suggested that the BFA-induced dissociation of coat proteins from membranes is a prerequisite for membrane tubulation (15Klausner R.D. Donaldson J.G. Lippincott-Schwartz J. J. Cell Biol. 1992; 116: 1071-1080Crossref PubMed Scopus (1542) Google Scholar, 17Scheel J. Pepperkok R. Lowe M. Griffiths G. Kreis T.E. J. Cell Biol. 1997; 137: 319-333Crossref PubMed Scopus (72) Google Scholar). However, accumulating evidence shows that membrane tubules are also formed under physiological conditions, indicating that membrane-bound compartments communicate with each other by tubules as well as vesicles (for example, see Refs. 18Cooper M.S. Cornell-Bell A.H. Chernjavsky A. Dani J.W. Smith S.J. Cell. 1990; 61: 135-145Abstract Full Text PDF PubMed Scopus (167) Google Scholar, 19Cluett E.B. Wood S.A. Banta M. Brown W.J. J. Cell Biol. 1993; 120: 15-24Crossref PubMed Scopus (72) Google Scholar, 20Sciaky N. Presley J. Smith C. Zaal K.J.M. Cole N. Moreira J.E. Terasaki M. Siggia E. Lippincott-Schwartz J. J. Cell Biol. 1997; 139: 1137-1155Crossref PubMed Scopus (410) Google Scholar, 21Klumperman J. Schweizer A. Clausen H. Tang B.L. Hong W. Oorschot V.V. Hauri H.-P. J. Cell Sci. 1998; 111: 3411-3425PubMed Google Scholar, 22Nakata T. Terada S. Hirokawa N. J. Cell Biol. 1998; 140: 659-674Crossref PubMed Scopus (266) Google Scholar, 23Shima D.T. Scales S.J. Kreis T.E. Pepperkok R. Curr. Biol. 1999; 9: 821-824Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar, 24Blum R. Stephens D.J. Schulz I. J. Cell Sci. 2000; 113: 3151-3159PubMed Google Scholar, 25Polishchuk R.S. Polishchuk E.V. Marra P. Alberti S. Buccione R. Luini A. Mironov A.A. J. Cell Biol. 2000; 148: 45-58Crossref PubMed Scopus (270) Google Scholar). The nucleotide exchange on ARF is catalyzed by guanine nucleotide exchange factors (GEFs). All ARF-GEFs identified so far possess a Sec7 domain of ∼200 amino acids that is a minimum unit for the catalysis (reviewed in Refs. 26Donaldson J.G. Jackson C.L. Curr. Opin. Cell Biol. 2000; 12: 475-482Crossref PubMed Scopus (319) Google Scholar and 27Jackson C.L. Casanova J.E. Trends Cell Biol. 2000; 10: 60-67Abstract Full Text Full Text PDF PubMed Scopus (391) Google Scholar). Members of the Sec7 family of ARF-GEFs can be subdivided into two major classes on the basis of sequence similarity and functional differences (26Donaldson J.G. Jackson C.L. Curr. Opin. Cell Biol. 2000; 12: 475-482Crossref PubMed Scopus (319) Google Scholar, 27Jackson C.L. Casanova J.E. Trends Cell Biol. 2000; 10: 60-67Abstract Full Text Full Text PDF PubMed Scopus (391) Google Scholar). The ARF-GEFs of high molecular mass (>100 kDa) include Saccharomyces cerevisiae Sec7p, Gea1p, and Gea2p, ArabidopsisGNOM/EMB30, and mammalian GBF1, BIG1, and BIG2. All but one, GBF1, have been reported to be sensitive to BFA (28Peyroche A. Paris S. Jackson C.L. Nature. 1996; 384: 479-481Crossref PubMed Scopus (236) Google Scholar, 29Morinaga N. Moss J. Vaughan M. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12926-12931Crossref PubMed Scopus (83) Google Scholar, 30Claude A. Zhao B.-P. Kuziemsky C.E. Dahan S. Berger S.J. Yan J.-P. Armold A.D. Sullivan E.M. Melançon P. J. Cell Biol. 1999; 146: 71-84Crossref PubMed Google Scholar, 31Mansour S.J. Skaug J. Zhao X.-H. Giordano J. Scherer S.W. Melançon P. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 7968-7973Crossref PubMed Scopus (101) Google Scholar, 32Togawa A. Morinaga N. Ogasawara M. Moss J. Vaughan M. J. Biol. Chem. 1999; 274: 12308-12315Abstract Full Text Full Text PDF PubMed Scopus (121) Google Scholar). All mammalian ARF-GEFs of high molecular weight are localized in the Golgi region and are believed to be involved in membrane trafficking (29Morinaga N. Moss J. Vaughan M. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12926-12931Crossref PubMed Scopus (83) Google Scholar, 30Claude A. Zhao B.-P. Kuziemsky C.E. Dahan S. Berger S.J. Yan J.-P. Armold A.D. Sullivan E.M. Melançon P. J. Cell Biol. 1999; 146: 71-84Crossref PubMed Google Scholar, 33Yamaji R. Adamik R. Takeda K. Togawa A. Pacheco-Rodriguez G. Ferrans V.J. Moss J. Vaughan M. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 2567-2572Crossref PubMed Scopus (99) Google Scholar). The second subfamily of smaller ARF-GEFs (45–50 kDa) include mammalian ARNO, cytohesin-1, GRP1/ARNO3, cytohesin-4, and EFA6. In contrast to the high molecular weight group, all low molecular weight ARF-GEFs are insensitive to BFA and are believed to be involved in endosomal recycling and cytoskeletal reorganization through activating ARF6 primarily (reviewed in Refs. 26Donaldson J.G. Jackson C.L. Curr. Opin. Cell Biol. 2000; 12: 475-482Crossref PubMed Scopus (319) Google Scholar and 27Jackson C.L. Casanova J.E. Trends Cell Biol. 2000; 10: 60-67Abstract Full Text Full Text PDF PubMed Scopus (391) Google Scholar). In this study, we examined the role of one of the high molecular weight ARF-GEFs, BIG2. When cells overexpressing BIG2 were treated with BFA, the dissociation from membranes of the AP-1 complex but not that of the COPI complex was blocked, suggesting that BIG2 is implicated in the formation of clathrin/AP-1-coated vesicles through activating ARFs but not that of COPI-coated vesicles. However, in the BIG2-overexpressing cells, the BFA treatment still induced the formation of membrane tubules with which not only BIG2 itself but also ARF, AP-1, and mannose 6-phosphate receptor (MPR) were associated. These observations indicate that BFA-induced coat dissociation is not necessarily a prerequisite for membrane tubulation and that another target of BFA, which inherently blocks membrane tubulation, is present in the cell. A BLAST search was performed against a human expressed sequence tag (EST) data base using the yeast Sec7p sequence as a query and identified an EST (GenBankTMaccession number AA121376) that coded for a part of a protein showing significant homology to the Sec7 domain sequence of Sec7p. A part of the EST sequence was amplified from a human liver cDNA library (Invitrogen) by polymerase chain reaction and used to screen a human hepatoma HepG2 cDNA library (34Hosaka M. Murakami K. Nakayama K. Biomed. Res. 1994; 15: 383-390Crossref Scopus (24) Google Scholar). Ten positive clones were identified by screening of ∼1.8 × 106clones, and the longest clone (clone 7) was subjected to further analysis. The cDNA insert of clone 7 was ∼3.0-kb long and covered a part of human BIG2 polypeptide (amino acid residues 103–1362). Subsequent screening of the same HepG2 library with the clone 7 insert identified several positive clones, and two of them (clones 22 and 67) covered the missing BIG2 region (amino acids 1–208 and 881–1719, respectively). However, the most C-terminal region of BIG2 did not appear to be covered by the cloned cDNAs. Another BLAST search against the human EST data base using the most 3′ part of the clone 67 insert identified an EST (GenBankTM accession numberAA594443) that appeared to cover the missing region, and the corresponding cDNA fragment (7C-AA) covering amino acids 1710–1785 was amplified from the human liver cDNA library. The amino acid sequence deduced from the composite cDNA sequence was confirmed to be identical to the reported human BIG2 sequence (32Togawa A. Morinaga N. Ogasawara M. Moss J. Vaughan M. J. Biol. Chem. 1999; 274: 12308-12315Abstract Full Text Full Text PDF PubMed Scopus (121) Google Scholar). To construct expression vectors for epitope-tagged BIG2, an oligonucleotide for an epitope sequence for hemagglutinin (HA) or for Myc was first introduced between theHindIII and BamHI sites of the pcDNA4/HisMax or pcDNA3 vector (Invitrogen), respectively. The entire coding sequence for BIG2 was then ligated downstream of the epitope sequence of the resulting pcDNA4-HAN or pcDNA3-MycN vector. The construction of expression vectors for C-terminally HA-tagged ARF1 was described previously (35Hosaka M. Toda K. Takatsu H. Torii S. Murakami K. Nakayama K. J. Biochem. 1996; 120: 813-819Crossref PubMed Scopus (63) Google Scholar). An expression vector for the GGA homology (GGAH) domain of Golgi-localizing γ-adaptin ear homology domain ARF-binding protein 1 (GGA1) fused to glutathioneS-transferase (GST) was constructed by subcloning of a cDNA fragment covering the GGAH domain (amino acids 141–326) of human GGA1 (36Takatsu H. Yoshino K. Nakayama K. Biochem. Biophys. Res. Commun. 2000; 271: 719-725Crossref PubMed Scopus (94) Google Scholar) into the pGEX4T-2 vector (Amersham Biosciences, Inc.). Monoclonal mouse antibody (2G11) to cation-independent MPR (37Dintzis S.M. Pfeffer S.R. EMBO J. 1990; 9: 77-84Crossref PubMed Scopus (14) Google Scholar) was kindly provided by Dr. Suzanne Pfeffer (Stanford University, Palo Alto, CA) or purchased from Affinity Bioreagents. Monoclonal mouse antibodies to γ-adaptin (100.3), mannosidase II (Man II) (53FC3), Lamp-1 (H4A3), and EEA1 (clone 14) and polyclonal rabbit anti-β-COP antibodies were purchased from Sigma, BAbCo, PharMingen, Transduction Laboratories, and Affinity Bioreagents, respectively. Monoclonal rat anti-HA (3F10) and monoclonal mouse anti-Myc (9E10) antibodies were from Roche Diagnostics and Santa Cruz Biotechnology, Inc., respectively. Alexa488-conjugated antibodies and Cy3- and peroxidase-conjugated secondary antibodies were from Molecular Probes and Jackson ImmunoResearch Laboratories, respectively. For the expression of HA-tagged BIG2, HeLa cells or Clone 9 rat hepatocytes were transfected with its expression vector using a FuGENE 6 transfection reagent (Roche Molecular Biochemicals) and incubated for 6 h in the presence of 20 mm sodium butyrate. For the coexpression of HA-tagged ARF1 and Myc-tagged BIG2, their expression vectors were microinjected into the nuclei of HeLa cells using Eppendorf Micromanipulator 5171 and Transjector 5246, and the cells were incubated for 3 h. Staining of the cells was performed by essentially the same procedures as described previously (38Torii S. Banno T. Watanabe T. Ikehara Y. Murakami K. Nakayama K. J. Biol. Chem. 1995; 270: 11574-11580Abstract Full Text Full Text PDF PubMed Scopus (50) Google Scholar, 39Takatsu H. Sakurai M. Shin H.-W. Murakami K. Nakayama K. J. Biol. Chem. 1998; 273: 24693-24700Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar). The transfected or microinjected cells were fixed with 4% paraformaldehyde in phosphate-buffered saline and permeabilized with 0.1% Triton X-100 or fixed and permeabilized with cold methanol. The cells were then incubated with the indicated combinations of primary antibodies followed by a combination of Alexa488-conjugated and Cy3-conjugated secondary antibodies and observed using a confocal microscope (TCS-SP2, Leica). The activation levels of ARF1 in cells were estimated by a recently developed pull-down assay using GGA (40Santy L.C. Casanova J.E. J. Cell Biol. 2001; 154: 599-610Crossref PubMed Scopus (314) Google Scholar). Human embryonic kidney 293 cells grown on a 10-cm dish were transfected with an expression vector for either wild type or mutant ARF1-HA alone or in combination with that for HA-tagged BIG2 using the FuGENE 6 reagent and incubated for 18 h in normal medium and then for 6 h in medium containing 20 mm sodium butyrate. The cells were then lysed in 0.65 ml of cell lysis buffer (50 mm Tris, pH 7.5, 100 mm NaCl, 2 mmMgCl2, 0.1% SDS, 0.5% sodium deoxycholate, 1% Triton X-100, 10% glycerol) containing a CompleteTM protease inhibitor mixture (Roche Molecular Biochemicals). The lysates were precleared by incubation with GST prebound to glutathione-Sepharose 4B beads (Amersham Biosciences, Inc.) for 30 min at 4 °C and centrifugation at 3000 rpm for 5 min at 4 °C in a microcentrifuge. The supernatants (0.5 ml) were then incubated with the fusion protein of GST and the GGA1 GGAH domain prebound to glutathione-Sepharose 4B beads for 30 min at 4 °C and centrifuged at 3000 rpm for 5 min at 4 °C in a microcentrifuge. The pellets were washed three times with cell lysis buffer and boiled in SDS-PAGE sample buffer. The samples were electrophoresed on a 15% SDS-polyacrylamide gel for detection of ARF or 7.5% SDS-polyacrylamide gel for detection of BIG2 and electroblotted onto an Immobilon-P membrane (Millipore). The blot was incubated sequentially with monoclonal rat anti-HA antibody (3F10) and peroxidase-conjugated anti-rat IgG and detected using a Renaissance Chemiluminescence Reagent Plus (PerkinElmer Life Sciences). Human BIG2 has recently been cloned and found to be associated with the Golgi region (32Togawa A. Morinaga N. Ogasawara M. Moss J. Vaughan M. J. Biol. Chem. 1999; 274: 12308-12315Abstract Full Text Full Text PDF PubMed Scopus (121) Google Scholar, 33Yamaji R. Adamik R. Takeda K. Togawa A. Pacheco-Rodriguez G. Ferrans V.J. Moss J. Vaughan M. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 2567-2572Crossref PubMed Scopus (99) Google Scholar). In these studies, however, neither its precise intracellular localization nor its physiological role has been determined. Because we have independently cloned human BIG2 cDNAs (see “Materials and Methods”), we set out to address these issues. To this end, we first attempted to raise antibodies to BIG2, but they did not work well in immunofluorescence studies. Therefore, we transiently transfected epitope-tagged (HA- or Myc-tagged) BIG2 into cells and examined its subcellular localization. When expressed at moderate levels in HeLa cells and in Clone 9 rat hepatocytes, HA-BIG2 was localized in the perinuclear region (Fig. 1,A–F). Similar perinuclear localization was observed for Myc-tagged BIG2 (see Fig. 3 C′). The staining for BIG2 overlapped with that for β-COP (a COPI subunit, Fig. 1 A), Man II (a medial Golgi membrane protein, Fig. 1 B), γ-adaptin (a subunit of the AP-1 complex, Fig. 1 C), or MPR (a protein recycling between the TGN and late endosomes, Fig.1 D). In contrast, BIG2 was not significantly colocalized with Lamp-1 (a marker for late endosomes and lysosomes, Fig.1 E) or EEA1 (an early endosomal protein, Fig.1 F). These observations are compatible with the data presented by Yamaji et al. (33Yamaji R. Adamik R. Takeda K. Togawa A. Pacheco-Rodriguez G. Ferrans V.J. Moss J. Vaughan M. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 2567-2572Crossref PubMed Scopus (99) Google Scholar) that BIG2 is localized primarily in the Golgi region.Figure 3Effect of BIG2-coexpression on the BFA-induced redistribution of ARF1. HeLa cells microinjected with an expression vector for HA-tagged ARF1 alone (A andB) or in combination with that for Myc-tagged BIG2 (C and D) were left untreated (A andC) or treated with 5 μg/ml BFA for 15 min (Band D) and processed for staining with anti-HA antibody (A and B) or double-staining with anti-HA (C and D) and anti-myc (C′ andD′) antibodies as described under “Materials and Methods.” Merged images are shown in C" andD".View Large Image Figure ViewerDownload Hi-res image Download (PPT) A previous study showed that BIG2 catalyzes guanine nucleotide exchange on ARF in vitro (32Togawa A. Morinaga N. Ogasawara M. Moss J. Vaughan M. J. Biol. Chem. 1999; 274: 12308-12315Abstract Full Text Full Text PDF PubMed Scopus (121) Google Scholar, 33Yamaji R. Adamik R. Takeda K. Togawa A. Pacheco-Rodriguez G. Ferrans V.J. Moss J. Vaughan M. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 2567-2572Crossref PubMed Scopus (99) Google Scholar). However, the study did not address whether it is also functional in vivo. To address this issue, we made use of a recently developed pull-down assay (40Santy L.C. Casanova J.E. J. Cell Biol. 2001; 154: 599-610Crossref PubMed Scopus (314) Google Scholar) for active GTP-bound ARF using GST fused to the GGAH domain of GGA, which interacts directly with ARF in a GTP-dependent manner (41Boman A.L. Zhang C.-J. Zhu X. Kahn R.A. Mol. Biol. Cell. 2000; 11: 1241-1255Crossref PubMed Scopus (224) Google Scholar, 42Dell'Angelica E.C. Puertollano R. Mullins C. Aguilar R.C. Vargas J.D. Hartnell L.M. Bonifacino J.S. J. Cell Biol. 2000; 149: 81-93Crossref PubMed Scopus (324) Google Scholar). The specificity of this assay was tested using lysates from cells transfected with wild type or mutants of C-terminally HA-tagged ARF1. As shown in Fig. 2, top panel, the binding efficiency to GST-GGAH of GTP-bound mutant of ARF1, ARF1(Q71L) (lane 3), was extremely high as compared with those of wild type ARF1 (lane 4) and its nucleotide-free mutant, ARF1(N126I) (lane 2). Although two bands for ARF1-HA were detected in each lane, we speculate that the upper band may represent myristoylated full-length ARF1, and the lower band may be its degradation product or its non-myristoylated form. We then examined the activation of ARF by BIG2 in cells using this assay. As shown in Fig.2, top panel, the binding of ARF1 from cells cotransfected with ARF1-HA and HA-BIG2 (lane 5) to GST-GGAH was increased 2–3-fold as compared with that from cells transfected with ARF1-HA alone (lane 4). The data indicate that BIG2 does activate ARF1 in cells. Although the -fold activation of ARF1 by BIG2 appears to be low, this may reflect the difference between them in the subcellular localization; namely, BIG2 is associated mainly with the TGN (see below), whereas ARF1 is associated not only with the TGN but with other compartments as well. BFA inhibits ARF-GEFs and thereby inhibits the recruitment of ARF onto Golgi membranes. Therefore, it is possible that the overexpression of an ARF-GEF may block the BFA-induced redistribution of ARF from membranes to the cytoplasm. To address this possibility, we examined the effect of BIG2 overexpression on the localization and BFA-induced redistribution of ARF. To this end, HeLa cells were microinjected with the expression vector for C-terminally HA-tagged ARF1 alone or in combination with that for Myc-tagged BIG2. The microinjected cells were left untreated or treated with BFA for 15 min and processed for immunofluorescence analysis. As reported previously (35Hosaka M. Toda K. Takatsu H. Torii S. Murakami K. Nakayama K. J. Biochem. 1996; 120: 813-819Crossref PubMed Scopus (63) Google Scholar), when expressed alone, ARF1 was localized in the perinuclear Golgi region (Fig.3 A). The treatment of the cells with BFA resulted in the redistribution of ARF1 into the cytoplasm (Fig. 3 B) as reported previously (44Teal S.B. Hsu V.W. Peters P.J. Klausner R.D. Donaldson J.G. J. Biol. Chem. 1994; 269: 3135-3138Abstract Full Text PDF PubMed Google Scholar, 45Zhang C.-J. Rosenwald A.G. Willingham M.C. Skuntz S. Clark J. Kahn R.A. J. Cell Biol. 1994; 124: 289-300Crossref PubMed Scopus (151) Google Scholar). The perinuclear localization of ARF1 was superimposed on that of BIG2 and was not apparently affected by the coexpression of BIG2 (Fig. 3,C–C"). When the BIG2-coexpressing cells were treated with BFA, we met with an unexpected phenomenon. Namely, in the BFA-treated cells, BIG2 was present on membrane tubules emanating from the perinuclear region (Fig. 3 D′). Furthermore, to our surprise, some fraction of ARF1 was present along the BIG2-positive tubules, although most of the fraction appeared to dissociate into the cytoplasm (Fig. 3, D–D"). The labeling for BIG2 and ARF1 often looked like queues of spherical swellings along the same tubular processes. These observations make it possible for BIG2 to be involved in the activation of ARF in vivo. Furthermore, these findings suggest that the BFA-induced membrane tubulation is not necessarily coupled with inactivation of ARF. Because BFA inhibits membrane association of coat protein complexes such as COPI and AP-1 through inhibiting ARF-GEFs, thereby blocking membrane recruitment of ARFs, it was reasonably speculated that the overexpression of ARF-GEF may also block the BFA-induced redistribution of the coat proteins. To see for which coat recruitment BIG2 is responsible, the localization of γ-adaptin or β-COP was examined when BIG2-transfected cells were treated with BFA. As shown in Fig. 4,A–A", when BIG2-transfected cells were treated with BFA for 15 min, the labeling for γ-adaptin was also observed along the BIG2-positive tubules, whereas in non-transfected cells seen in the same field marked by asterisks, γ-adaptin was redistributed into the cytoplasm. Again, the labeling for BIG2 and γ-adaptin often looked similar to queues of spherical swellings along the same tubular processes. The membrane tubules were still observed by a much longer incubation period (up to 2 h) in the presence of BFA (data not shown), suggesting that the tubular processes did not fuse with any compartment other than the compartment from which the tubules originated. By contrast, β-COP was redistributed into the cytoplasm by treatment with BFA for only 1 min even in the BIG2-transfected cells and was not observed along the BIG2-positive tubules (Fig.5, A–A").Figure 5Effect of BIG2-overexpression on BFA-induced redistribution of β-COP and Man II. Clone 9 cells transfected with the expression vector for HA-tagged BIG2 were treated with 5 μg/ml BFA for 1 min (A and B) or 15 min (C and D) and processed for double-staining with antibodies to the HA epitope (A–D) and to either β-COP (A′ and C′) or Man II (B′ and D′) as described under “Materials and Methods.” Merged images are shown in A"–D". Asterisks indicate non-transfected cells.View Large Image Figure ViewerDownload Hi-res image Download (PPT) To more clearly show that γ-adaptin remained associated with membranes in cells overexpressing BIG2 when treated with BFA, we exploited the fact that BFA-induced membrane tubules extend along microtubules, and the treatment of cells with nocodazole, a microtubule-disrupting drug, results in complete inhibition of the tubule formation (12Lippincott-Schwartz J. Yuan L. Tipper C. Amherdt M. Orci L. Klausner R.D. Cell. 1991; 67: 601-616Abstract Full Text PDF PubMed Scopus (680) Google Scholar, 13Wood S.A. Park J.E. Brown W.J. Cell. 1991; 67: 591-600Abstract Full Text PDF PubMed Scopus (289) Google Scholar, 14Robinson M.S. Kreis T.E. Cell. 1992; 69: 129-138Abstract Full Text PDF PubMed Scopus (286) Google Scholar, 18Cooper M.S. Cornell-Bell A.H. Chernjavsky A. Dani J.W. Smith S.J. Cell. 1990; 61: 135-145Abstract Full Text PDF PubMed Scopus (167) Google Scholar). As shown in Fig.6, the treatment of the BIG2-transfected cells with nocodazole for 2 h caused not only the labeling for BIG2 (Fig. 6, A and C) but also the labeling for γ-adaptin (Fig. 6 A′) and β-COP (Fig. 6 C′) to fragment and disperse throughout the cytoplasm. When the nocodazole-treated cells were then treated with BFA for 15 min in the continued presence of nocodazole, γ-adaptin remained associated with the fragmented BIG2-positive structures while being redistributed into the cytoplasm in non-transfected cells seen in the same field (Fig. 6,B and B′). In contrast, β-COP was redistributed from membranes of the fragmented Golgi into the cytoplasm even in the BIG2-overexpressing cells (Fig. 6, D andD′). These observations indicate that BIG2 is involved in the membrane recruitment of the AP-1 complex but not that of the COPI complex. The above results suggest that BIG2 functions primarily on TGN membranes. To corroborate this finding, cells overexpressing HA-BIG2 were treated with BFA for 15 min and double stained for HA and MPR. MPR is present in the TGN and late endosomes and appears in TGN-derived tubules after short periods of BFA treatment (13Wood S.A. Park J.E. Brown W.J. Cell. 1991; 67: 591-600Abstract Full Text PDF PubMed Scopus (289) Google Scholar, 43Wood S.A. Brown W.J. J. Cell Biol. 1992; 119: 273-285Crossref PubMed Scopus (69) Google Scholar). As shown in Fig.4, B–B", the labeling for MPR was observed along the BIG2-positive membrane tubules induced by the BFA treatment. Again, the MPR labeling often looked like a queue of spherical swellings. In contrast, punctate labeling for Lamp-1, a late endosome/lysosome marker, in untreated cells (see Fig. 1 E) was not apparently affected by the BFA treatment nor observed along the BFA-induced tubules containing BIG2 (data not shown). Thus, the BIG2-positive tubules appear to be derived from the TGN. We then examined whether BIG2 overexpression also affects the disorganization of the Golgi complex induced by BFA. It has been known that by a relatively short term incubation period (<1 min) of cells with BFA, COPI is dissociated from Golgi membranes, but the Golgi complex itself is apparently intact, whereas a longer incubation period results in the formation of Golgi tubules, which in turn fuse with the endoplasmic reticulum (11Donaldson J.G. Lippincott-Schwartz J. Bloom G.S. Kreis T.E. Klausner R.D. J. Cell Biol. 1990; 111: 2295-2306Crossref PubMed Scopus (292) Google Scholar). As shown in Fig. 5, A–A" andB–B", irrespective of BIG2 transfection, the treatment of cells with BFA for 1 min led to the redistribution of β-COP but not Man II, a medial Golgi marker. This finding indicates that the β-COP redistribution observed in BIG2-overexpressing cells did not result from the Golgi disorganization. When cells were treated with BFA for 15 min, the Man II labeling was a cytoplasmic pattern reminiscent of the endoplasmic reticulum even in the BIG2-transfected cells (Fig. 5,D–D"), indicating that BIG2 is unable to affect the BFA-induced Golgi disorganization. These observations all together indicate that (i) BIG2 functions primarily on TGN membranes, (ii) through activating ARF, BIG2 is responsible for the recruitment of the AP-1 complex, but not that of the COPI complex, onto membranes, and (iii) the dissociation of coat proteins from membranes is not necessarily a prerequisite for membrane tubulation induced by BFA. Guanine nucleotide exchange on ARFs catalyzed by their GEFs, especially high molecular weight ARF-GEFs, results in their membrane association that triggers subsequent membrane recruitment of coat protein complexes such as the COPI complex with cis-Golgi membranes and that of the AP-1 adaptor complex with TGN membranes. Thus, ARF-GEFs are key regulators for the formation of coated carrier vesicles (26Donaldson J.G. Jackson C.L. Curr. Opin. Cell Biol. 2000; 12: 475-482Crossref PubMed Scopus (319) Google Scholar, 27Jackson C.L. Casanova J.E. Trends Cell Biol. 2000; 10: 60-67Abstract Full Text Full Text PDF PubMed Scopus (391) Google Scholar). BFA is known to inhibit high molecular weight ARF-GEFs and thereby inhibit membrane association of ARF itself and that of coat protein complexes (6Lippincott-Schwartz J. Cole N.B. Donaldson J.G. Histochem. Cell Biol. 1998; 109: 449-462Crossref PubMed Scopus (70) Google Scholar, 15Klausner R.D. Donaldson J.G. Lippincott-Schwartz J. J. Cell Biol. 1992; 116: 1071-1080Crossref PubMed Scopus (1542) Google Scholar, 16Jackson C.L. Subcell. Biochem. 2000; 34: 233-272Crossref PubMed Google Scholar). Furthermore, this drug induces tubulation of Golgi and TGN membranes, which has been proposed to follow the coat dissociation (15Klausner R.D. Donaldson J.G. Lippincott-Schwartz J. J. Cell Biol. 1992; 116: 1071-1080Crossref PubMed Scopus (1542) Google Scholar, 17Scheel J. Pepperkok R. Lowe M. Griffiths G. Kreis T.E. J. Cell Biol. 1997; 137: 319-333Crossref PubMed Scopus (72) Google Scholar). Previous studies (32Togawa A. Morinaga N. Ogasawara M. Moss J. Vaughan M. J. Biol. Chem. 1999; 274: 12308-12315Abstract Full Text Full Text PDF PubMed Scopus (121) Google Scholar, 33Yamaji R. Adamik R. Takeda K. Togawa A. Pacheco-Rodriguez G. Ferrans V.J. Moss J. Vaughan M. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 2567-2572Crossref PubMed Scopus (99) Google Scholar) have shown that BIG2 localizes mainly in the Golgi region but have not addressed which coat recruitment or which transport step BIG2 regulates through activating ARF. In this study, we addressed these questions by comparing changes in response to BFA between normal and BIG2-overexpressing cells. Because the guanine nucleotide exchange activity of BIG2 on ARFs is inhibited by BFA, it was anticipated that the BIG2 overexpression was able to antagonize some aspects of the effects of BFA. When overexpressed in cells, BIG2 blocked BFA-induced dissociation of ARF and the AP-1 complex from membranes. By contrast, the BIG2 overexpression was unable to antagonize BFA-induced COPI redistribution. The data indicate that BIG2 is involved in the recruitment of the AP-1 complex with the TGN membranes through activating ARF but not in the recruitment of COPI with cis-Golgi membranes. In the course of the experiments, we unexpectedly found that the BIG2 overexpression did not inhibit the BFA-induced membrane tubulation and that ARF and AP-1 were associated with the BIG2-containing membrane tubules. The tubules appear to be derived from the TGN but not from the Golgi complex, because MPR but not Man II was also found on the BIG2-positive tubules. The tubular processes lasted up to at least 2 h in the presence of BFA and did not appear to fuse with any compartment other than the TGN. In normal cells, BFA induces tubules from the TGN, which then fuse with endosomal compartments (12Lippincott-Schwartz J. Yuan L. Tipper C. Amherdt M. Orci L. Klausner R.D. Cell. 1991; 67: 601-616Abstract Full Text PDF PubMed Scopus (680) Google Scholar, 13Wood S.A. Park J.E. Brown W.J. Cell. 1991; 67: 591-600Abstract Full Text PDF PubMed Scopus (289) Google Scholar). Given that prior uncoating is generally required for membrane fusion, it is probable that the continuous presence of the AP-1 coat led to the stable BFA-induced tubules observed in the BIG2-overexpressing cells by inhibiting membrane fusion. These observations also indicate that the dissociation of coat proteins from membranes is not necessarily a prerequisite for membrane tubulation and that a BFA target other than ARF-GEFs may exist in the cell and regulate membrane tubulation/fission. Such a candidate target is BFA-ADP-ribosylated substrate (BARS), which is ADP-ribosylated in a BFA-dependent manner in vitro and in vivo (46Mironov A. Colanzi A. Silletta M.G. Fiucci G. Flati S. Fusella A. Polishchuk R. Di Mironov Jr., A. Tullio G. Weigert R. Malhotra V. De Corda D. Matteis M.A. Luini A. J. Cell Biol. 1997; 139: 1109-1118Crossref PubMed Scopus (43) Google Scholar, 47Spanfò S. Silletta M.G. Colanzi A. Alberti S. Fiucci G. Valente C. Fusella A. Salmona M. Mironov A. Luini A. Corda D. J. Biol. Chem. 1999; 274: 17705-17710Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar). When added to permeabilized cells, BARS strongly inhibits BFA-induced Golgi tubular disassembly, whereas pre-ADP-ribosylated BARS does not. By contrast, BARS has no effect on BFA-induced coat dissociation. Recently, Weigert et al. (48Weigert R. Silletta M.G. Spanfò S. Turacchio G. Cericola C. Colanzi A. Senatore S. Mancini R. Polishchuk E.V. Salmona M. Facchiano F. Burger K.N.J. Mironov A. Luini A. Corda D. Nature. 1999; 402: 429-433Crossref PubMed Scopus (279) Google Scholar) have reported that BARS is directly or indirectly implicated in the acylation of lysophosphatidic acid to generate phosphatidic acid and regulates fission of Golgi membrane tubules. Because lysophosphatidic acid acyltransferase activity has been reported to be also required for another membrane fission event (49Schmidt A. Wolde M. Thiele C. Fest W. Kratzin H. Podtelejnikov A.V. Witke W. Huttner W.B. Söling H.-D. Nature. 1999; 401: 133-141Crossref PubMed Scopus (449) Google Scholar, reviewed in Ref. 50Scales S.J. Scheller R.H. Nature. 1999; 401: 123-124Crossref PubMed Scopus (45) Google Scholar), it is likely that BFA induces membrane tubules by inhibiting membrane fission by inhibiting the function of BARS or a related protein. However, we cannot exclude a possibility that inhibition of another BFA-sensitive ARF-GEF such as BIG1 may lead to the tubule formation. Future studies will be required to solve this problem. Another striking observation in this study is that the labeling for BIG2, AP-1, ARF, and MPR often looked like queues of spherical swellings along the BFA-induced tubular processes. Supporting our observation is the most recent report of Huang et al. (51Huang F. Nesterov A. Carter R.E. Sorkin A. Traffic. 2001; 2: 345-357Crossref PubMed Scopus (38) Google Scholar). They have observed a dynamic movement of the μ1 AP-1 subunit tagged with yellow fluorescent protein in living cells by time-lapse imaging and found that the μ1-positive structures move as queues of spherical swellings along the tubules, and vesicles are frequently formed at the tips of the tubular extensions. Thus, the images of the AP-1-positive swellings along the BFA-induced tubules we observed may correspond to frozen images at some time point of the dynamically moving swellings along the tubules formed under physiological conditions. In this context, an observation in the BARS study (48Weigert R. Silletta M.G. Spanfò S. Turacchio G. Cericola C. Colanzi A. Senatore S. Mancini R. Polishchuk E.V. Salmona M. Facchiano F. Burger K.N.J. Mironov A. Luini A. Corda D. Nature. 1999; 402: 429-433Crossref PubMed Scopus (279) Google Scholar) described earlier is also noteworthy. A feature of intermediates of Golgi tubule fission induced by BARS is a number of sites in which the tubule diameter is greatly reduced, in other words, queues of boluses or swellings are found in the tubule fission intermediates. Although the study (48Weigert R. Silletta M.G. Spanfò S. Turacchio G. Cericola C. Colanzi A. Senatore S. Mancini R. Polishchuk E.V. Salmona M. Facchiano F. Burger K.N.J. Mironov A. Luini A. Corda D. Nature. 1999; 402: 429-433Crossref PubMed Scopus (279) Google Scholar) did not determine whether the swollen regions are coated or whether the constricted sites along the tubules represent bona fidefission sites, it is tempting to speculate that the swollen regions in the intermediates correspond to those observed when BIG2-overexpressing cells were treated with BFA. We thank Dr. Suzanne Pfeffer for providing anti-MPR antibody.