QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

Vol. 16, Issue 8, 3786-3799, August 2005

This Article Abstract Full Text (PDF) All Versions of this Article: E05-04-0289v1 16/8/3786    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Park, S.-K. Articles by Jackson, C. L. Search for Related Content PubMed PubMed Citation Articles by Park, S.-K. Articles by Jackson, C. L.

Mutations in a Highly Conserved Region of the Arf1p Activator GEA2 Block Anterograde Golgi Transport but Not COPI Recruitment to Membranes

Cell Biology and Metabolism Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD 20892

Submitted April 8, 2005; Revised May 23, 2005; Accepted May 25, 2005
Monitoring Editor: Akihiko Nakano

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
We have identified an important functional region of the yeast Arf1 activator Gea2p upstream of the catalytic Sec7 domain and characterized a set of temperature-sensitive (ts) mutants with amino acid substitutions in this region. These gea2-ts mutants block or slow transport of proteins traversing the secretory pathway at exit from the endoplasmic reticulum (ER) and the early Golgi, and accumulate both ER and early Golgi membranes. No defects in two types of retrograde trafficking/sorting assays were observed. We find that a substantial amount of COPI is associated with Golgi membranes in the gea2-ts mutants, even after prolonged incubation at the nonpermissive temperature. COPI in these mutants is released from Golgi membranes by brefeldin A, a drug that binds directly to Gea2p and blocks Arf1 activation. Our results demonstrate that COPI function in sorting of at least three retrograde cargo proteins within the Golgi is not perturbed in these mutants, but that forward transport is severely inhibited. Hence this region of Gea2p upstream of the Sec7 domain plays a role in anterograde transport that is independent of its role in recruiting COPI for retrograde transport, at least of a subset of Golgi-ER cargo.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
The current model for transport through the Golgi apparatus postulates that anterograde transport is an indirect consequence of retrograde transport of Golgi resident enzymes from earlier to later compartments (Bonfanti et al., 1998; Glick and Malhotra, 1998; de Graffenried and Bertozzi, 2004). Retrograde transport of many proteins is mediated by the COPI coat, a heptameric complex that cycles between membranes and cytoplasm (Bonifacino and Lippincott-Schwartz, 2003; Duden, 2003; Lee et al., 2004). COPI is recruited to Golgi membranes by Arf1-GTP, which is itself activated by one of a family of guanine-nucleotide exchange factors (GEFs) carrying a Sec7 domain (Jackson and Casanova, 2000). The Sec7 domain is sufficient to catalyze exchange of GTP for GDP on class I Arf proteins such as Arf1 and Arf3 in humans, and Arf1p and Arf2p in Saccharomyces cerevisiae. Two subfamilies of Arf GEFs localize to the Golgi complex in both yeast and mammalian cells: GBF1 and its yeast homologues Gea1p and Gea2p localize at steady state to the early Golgi, and Sec7p and its mammalian homologues BIG1 and BIG2 localize at steady state to the late Golgi (Zhao et al., 2002; Shin and Nakayama, 2004). In yeast, Gea1p and Gea2p are functionally redundant in that one or the other can be deleted with no detectable effect on growth rate or rates of transport to the cell surface and vacuole, whereas a gea1 gea2 double mutant is inviable (Peyroche et al., 1996, 2001). A similar situation is likely true of the BIG1–BIG2 pair in mammalian cells, given their extensive sequence homology (74% identity; Togawa et al., 1999; Yamaji et al., 2000) and the fact that a human disease (autosomal recessive periventricular heterotopia) is caused by almost complete deletion of the BIG2 gene (Sheen et al., 2004). The fact that these affected individuals with no BIG2 function are born and live several years (albeit with severe abnormalities) indicates that at the cellular level, there is a high level of redundancy between BIG1 and BIG2 for essential functions. Why multiple Arf GEFs are needed for Arf1 function at the Golgi in organisms from yeast to humans is not known but likely is involved in specificity of the multiple effectors of the same Arf protein acting at different membrane sites within the Golgi (Donaldson and Jackson, 2000).

In the cisternal maturation model of Golgi traffic, glycosylation enzymes and fusion machinery are transported from later Golgi compartments back to earlier ones and from the cis-Golgi and/or intermediate compartment to the endoplasmic reticulum (ER) (Bonfanti et al., 1998; Glick and Malhotra, 1998). In this way, an early Golgi compartment is transformed into a later one when it receives incoming glycosylation enzymes from this later compartment (de Graffenried and Bertozzi, 2004). There is evidence in both yeast and mammalian cells that COPI mediates the retrograde trafficking of glycosylation enzymes, cargo receptors, and fusion machinery (Nakano, 2004). The precise mechanism of this retrograde trafficking is a subject of ongoing research, and there is debate as to whether retrograde-directed glycosylation enzymes in mammalian cells are packaged into COPI vesicles or are sorted into tubules that connect Golgi cisternae (Kweon et al., 2004; Trucco et al., 2004). Trucco et al. (2004) have proposed that COPI vesicles contain not Golgi enzymes but primarily the fusion machinery, such as the SNARE protein membrin. Recently, it has been demonstrated biochemically that distinct populations of COPI-containing membranes can be isolated from cells, suggesting that there are different COPI-mediated retrograde trafficking pathways responsible for transporting different retrograde cargo (Malsam et al., 2005).

In this study, we have isolated a set of gea1 gea2-ts temperature-sensitive mutants (referred to as gea2-ts mutants) and characterized their secretion and morphological phenotypes. Surprisingly, in three mutants, anterograde trafficking is severely attenuated, whereas COPI-dependent retrograde trafficking of three marker proteins is not defective. In addition, COPI is recruited to membranes in these mutants in an Arf1-GTP–dependent manner, and both ER and early Golgi membranes accumulate. These results indicate either that Gea2p plays a direct role in anterograde transport or that there are multiple pathways of retrograde transport within the Golgi in yeast.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Mutagenesis and Construction of gea2-ts Strains
Strains and plasmids used in this study are listed in Tables 1 and 2, respectively. Four regions of GEA2 were mutagenized separately using five unique restriction sites spanning the gene, employing a mutagenic PCR approach as described previously (Muhlrad et al., 1992). For each region, a gapped plasmid made by digesting pSKP20 at two unique restriction sites was isolated by gel purification (Figure 1). Random mutagenesis of each region of the GEA2 open reading frame was performed by PCR amplification under conditions that decreased the fidelity of Taq polymerase (50 µM ATP, one fourth the concentration of the other three nucleotides). pSKP20 was used as template with primers annealing to sequences that flanked the corresponding pair of restriction sites used to gap pSKP20. Equimolar amounts of gapped plasmid and mutagenized DNA were cotransformed into CJY52–10-2 gea1 gea2/pCLJ92 (2µ-GEA1-URA3). LEU2+ transformants were replica plated onto 5-fluorootic acid (5-FOA) plates for counterselection of the URA3 plasmid carrying the GEA1 gene. Temperature-sensitive mutants of GEA2 were screened by selection of yeast cells capable of growing at 24°C but not at 37°C. pSKP20 plasmids bearing temperature-sensitive gea2-ts alleles were isolated and amplified in E. coli for DNA sequencing analysis to map mutations.

View larger version (23K): [in this window] [in a new window]   Figure 1. Schematic diagram showing strategy for isolating gea2-ts mutants. A centromeric plasmid carrying Gea2p-GFP (pSKP20) was cut with two unique restriction enzymes (XbaI and XhoI are shown here) to create a gap within the GEA2 gene. A region of GEA2 with sequences flanking the restriction sites was amplified by PCR, using mutagenic conditions that favor introduction of single nucleotide mismatches. The PCR product and gapped plasmid were cotransformed into strain CJY52-10-2 gea1 gea2/pCLJ92 in which the chromosomal copies of the GEA1 and GEA2 genes were deleted and the GEA1 gene was present on a URA3 plasmid. The yeast cells repaired the gapped plasmid, incorporating the mutagenized PCR product into the GEA2 gene. The wild-type GEA1 plasmid was then lost by counterselection against URA3 on 5-fluorootic acid (FOA) plates, and the resulting strains were then tested for temperature sensitivity.

Strain CJY101 gea1 gea2/pCLJ92 was constructed by replacing the GEA2 open reading frame with HIS3 in BY4742 gea1 carrying pCLJ92. SEC21-3xGFP strains CJY102/pSKP60, CJY102/pSKP61, CJY102/pSKP62, CJY102/pSKP63, and CJY102/pSKP64 were constructed from CJY101/pCLJ92 by plasmid shuffling on 5-FOA and then by pop-in–pop-out replacement of SEC21 with SEC21-3xGFP as described (Rossanese et al., 2001). Briefly, CJY101/pSKP60, CJY101/pSKP61, CJY101/pSKP62, CJY101/pSKP63, and CJY101/pSKP64 cells were transformed with YIplac211-SEC21-3xGFP selecting for "pop-in" integration in medium lacking uracil and then grown on 5-FOA to select for "pop-out" events that reconstitute SEC21-3xGFP at the chromosomal SEC21 locus. SEC7–3xGFP strains CJY103/pSKP60, CJY103/pSKP61, CJY103/pSKP62, CJY103/pSKP63, and CJY103/pSKP64 were constructed from CJY101/pCLJ92 in a similar manner, but using pSSEC7–3xE-GFP for pop-in–pop-out replacement of SEC7 with SEC7–3xGFP. CJY104 was constructed by pop-in–pop-out replacement of SEC21 with SEC21-3xGFP in strain BY4742 erg6.

To construct strain CJY105/pSKP60 + pAP45 and the gea2-ts and sec7-ts single and double mutant versions, strain APY045-18-3/pCLJ90 (CEN-TRP1-SUP11-GEA1) + pAP32 (2µ-URA3-SEC7; Peyroche et al., 1999) was first transformed with pSKP60, pSKP63, or pSKP64. APY045-18-3/pCLJ90 + pAP32 colonies are white because of SUP11 suppression of the nonsense ade2-101 mutation, whereas colonies of cells that have lost pCLJ90 are red. Hence clones of cells in which pCLJ90 was replaced by pSKP60, pSKP63, or pSKP64 were identified by red colonies or sectors among white colonies. Strains APY045-18-3/pSKP60 + pAP32, APY045-18-3/pSKP63 + pAP32, and APY045-18-3/pSKP64 + pAP32 were then transformed with pAP45, pSKP72, or pSKP78, and plasmid shuffling was carried out on 5-FOA plates to select for clones that had lost pAP32. The resulting strains were then transformed with plasmid YIplac211-SEC21-3xGFP to generate a fusion gene encoding Sec21p-3xGFP at the chromosomal SEC21 locus, as described above (Table 1).

Microscopy of Living Cells
We carried out live imaging of yeast cells as previously described (Rossanese et al., 2001) with minor modifications. To immobilize cells for imaging, a LabTek chamber slide (Nalge Nunc International, Rochester, NY) was pretreated with 2 mg/ml concanavalin A (Sigma, St. Louis, MO, L-7647), rinsed with distilled water, and dried. Yeast cultures were grown overnight in yeast synthetic complete (SC) medium (Qbiogene, Carlsbad, CA) at 24°C, diluted into fresh medium, and grown to early log phase (OD₆₀₀ = 0.3–0.5). For imaging at 37°C, an aliquot (1 ml) of the culture was added to a LabTek chamber and left undisturbed for 10 min at room temperature. The medium was removed and any unbound cells remaining were washed away by gently pipetting several aliquots of fresh medium into the chambers. A final aliquot of fresh medium prewarmed to 37°C was added to the bound cells in the chamber and kept at 37°C for 20 min before imaging. Cells were maintained at 37°C during imaging by use of a space heater (Air Stream Stage Incubator, Nevtek, Burnsville, VA) directed at the microscope stage. Cells were imaged using an UltraView LCL confocal imaging system (Perkin Elmer-Cetus, Boston, MA) mounted on a Nikon TE2000-S fluorescence microscope (Melville, NY). The GFP signal was visualized by excitation with a 488-nm laser and collection with a 505-nm filter. Brefeldin A (BFA) treatment was performed by adding BFA (Sigma, B-7651) dissolved in ethanol to a final concentration of 100 µg/ml to the medium covering the cells bound to a LabTek chamber. Fluorescence images were processed using Adobe Photoshop (San Jose, CA).

Immunofluorescence Analysis
Immunofluorescence was performed as described previously (Peyroche et al., 2001). Cells were grown overnight in selective SC medium lacking appropriate amino acids and/or uracil, diluted into fresh medium, and grown at 24°C to early log phase. An aliquot of cells were shifted to 37°C for 1 h. Cells were fixed by adding formaldehyde (Sigma, F-1635) to a final concentration of 3.7% and incubated for 20 min at the growth temperature. Cells were washed with 1 ml KP sorbitol buffer (50 mM potassium phosphate, 1.2 M sorbitol, 0.5 M MgCl₂) and refixed with 3.7% formaldehyde overnight at 4°C. To convert cells into spheroplasts, 50 µl of fixed cells were digested with Zymolyase 100T (2 µg/µl) and -mercaptoethanol (10%) for 20 min at 30°C. Fixed spheroplasts were incubated first with mouse monoclonal anti-HA.11 (MMS-101R, Babco-Covance, Princeton, NJ; 10 µg/ml final concentration) for 1 h at room temperature and then with a 1–100 dilution of anti-mouse IgG Alexa Fluor 568 (Molecular Probes, Eugene, OR) for 1 h in the dark at room temperature. For microscopy, cells were transferred to a microscope slide in Prolong Antifade solution (Molecular Probes).

Electron Microscopy
gea2-ts mutants and the GEA2 control were grown overnight at permissive temperature (24°C) and then shifted to 37°C for 30 min. Cells were fixed at 37°C in suspension with 2x 4% formaldehyde with 0.2% glutaraldehyde in 100 mM phosphate buffer at pH 6.7. Cells were pelleted, rinsed, and resuspended in 5% gelatin. The gelatin/yeast was cooled and cut into cubes then placed in 2.3 M sucrose. Cubes were frozen onto stubs and placed in a cryoultramicrotome for sectioning. Sections were produced and gathered and immunolabeled as previously described (Liou et al., 1996; Tokuyasu, 1986). The Och1-HA epitope tag was recognized with a rat anti-HA antibody (Roche, Indianapolis, IN). Ultrathin frozen sections were incubated with anti-HA antibody and then incubated with rabbit anti-rat antibody (Jackson ImmunoResearch, West Grove, PA). The Och1p-HA was visualized by the addition of 10-nm protein A-gold (University of Utrecht, Netherlands). Labeled sections were viewed and photographed with a Philips CM-10 transmission electron microscope (FEI, Beaverton, OR).

Cell Labeling and Immunoprecipitation
Cells grown overnight at 24°C in SC medium lacking methionine (Qbiogene, Carlsbad, CA) were resuspended in the same medium and incubated at 24°C until the cells reached early log phase. One-half of the culture was then shifted to 37°C for 60 min. Cells were concentrated to 5 OD/ml and pulsed-labeled for 10 min by addition of 125 µCi/ml ³⁵S-methionine (Perkin Elmer-Cetus, specific activity 1175 Ci/mmol) at the growth temperature. One-tenth volume of prewarmed 10x chase solution (50 mM methionine and 10 mM cysteine, 4% yeast extract, and 20% glucose) was added and incubation for 0, 5, 10, or 20 min was continued at the growth temperature. One milliliter of labeled cells was then immediately transferred to a microcentrifuge tube containing 100 µl of 10 mM sodium azide, washed with 1 ml of cold 10 mM sodium azide, and resuspended in 200 µl of lysis buffer (1% SDS, 50 mM Tris-HCl, pH 7.4, 1 mM EDTA) containing 1 mM phenylmethylsulfonyl fluoride (PMSF; Sigma, P-7626) and 25x EDTA-free protease inhibitor cocktail (Complete Mini, Roche). Labeled cells were lysed by vortexing for 30 s (4x) at 4°C in the presence of 80–90% (vol/vol) glass beads (Sigma, G-8772). IP buffer, 600 µl (15 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% SDS, 1% Triton X-100, 2 mM NaN₃), was added to each tube and centrifuged for 10 min at 16,000 x g at 4°C to clear the sample.

For immunoprecipitation, 1–2 µl of specific antiserum and 50 µl of 20% Protein G Sepharose (Amersham Biosciences, Piscataway, NJ) in IP buffer were added to 700 µl of labeled sample and incubated at 4°C overnight with constant gentle agitation. Pellets were washed, and the samples were then resolved by electrophoresis on a 4–20% NuPAGE Novex Bis-Tris gradient gel (Invitrogen, Carlsbad, CA). Secretion of proteins into the medium was assayed as described previously (Peyroche and Jackson, 2001). Briefly, labeled cells were spun down and the medium was transferred to a microcentrifuge tube containing one-fourth volume 100% cold trichloroacetic acid (TCA) and placed on ice for 15 min. TCA precipitates were collected by centrifugation, washed with acetone, and resolved by electrophoresis on a 4–20% NuPAGE Novex Bis-Tris gradient gel (Invitrogen). Autoradiograms were visualized by exposure onto Biomax MR film (Eastman Kodak, Rochester, NY).

Western Blotting
Strains carrying plasmid pWB-Am were grown to early log phase at 24 and 37°C. Sodium azide, 2 mM, was added to the cultures, and cells were harvested and then washed with nondenaturing lysis buffer (50 mM Tris-HCl, pH 7.4, 300 mM NaCl, 5 mM EDTA, 1% Triton X-100, 0.02% NaN₃). Cells were resuspended in 100 µl of nondenaturing lysis buffer with 1 mM PMSF and 25x EDTA-free protease inhibitor cocktail. Cells were lysed by vortexing four times for 30 s with glass beads. Lysed samples were cleared by centrifugation for 2 min at 16,000 x g and then resolved by electrophoresis on a 4–20% NuPAGE Novex Bis-Tris gradient gel (Invitrogen). Sec22-myc- derivatives were detected with polyclonal rabbit anti-Sec22p antibody (Lupashin and Waters, 1997) or with anti-myc antibody (9E10, Babco-Covance, Princeton, NJ).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
We introduced random mutations into the GEA2 gene using a PCR and plasmid shuffling approach (Figure 1, and see Materials and Methods). The strain used carried the wild-type GEA1 gene on a URA3 plasmid (with both gea1 and gea2 deletions of the chromosomal copies of the genes). The mutant gea2 alleles were introduced into this strain on a low-copy LEU2 plasmid, and after counterselection on FOA medium, the mutant gea2 allele became the sole copy of the redundant GEA1 and GEA2 genes in the cell (Figure 1). We mutagenized four regions of the GEA2 gene separately. Region 1 spanned amino acids 1–540, Region 2 amino acids 370–962, Region 3 amino acids 767-1197, and Region 4 amino acids 1019–1460. Transformants were tested for growth at 24 and 37°C, and clones failing to grow at 37°C were recovered and tested further. Plasmids from 127 mutants that reproducibly exhibited a ts growth defect were isolated from yeast cells and sequenced. Several mutants had a large number of substitutions and were not studied further. The localization of the temperature-sensitive gea2-ts-GFP was determined for 67 of these mutants. In most cases the signal was weaker than for the wild-type at 37°C, with some spots remaining and some cytoplasmic signal (our unpublished data).

A subset of 50 mutants had only one or two substitutions in the regions mutagenized. These mutations were not randomly distributed across the Gea2p protein, but were clustered (Figure 2A). The Sec7 domain, the catalytic domain of the protein, had the highest frequency of mutations, as expected given the functional importance of this region. A region of 200 amino acids just upstream of the Sec7 domain had the next-highest frequency of mutations. This region contains residues evolutionarily conserved in Arf GEFs of both the Gea/GBF and Sec7/BIG subfamilies, and we refer to it as B/G (Cox et al., 2004; Figure 2B). Of the 50 mutants with one or two amino acid substitutions, we tested 44 by Western blot analysis to determine the stability of the mutant protein. Interestingly, the majority of the mutant proteins were unstable at the nonpermissive temperature, with only 9 of the 44 mutants having a level of gea2-ts protein equivalent to that of the wild type. We chose these nine mutants to study further, because the mutant phenotype is not simply due to total loss of the Gea2p protein by degradation, but may have specific effects due to mutations in highly conserved domains. An 25 amino acid region contains one of the most highly conserved motifs in the Gea/GBF and Sec7/BIG Arf GEFs: N Y/F D C D/N (aa 483–487 in Gea2p; Figure 2B). We found the D485G or C486R mutations in combination with other substitutions in four mutants of the collection, two for each substitution. To determine the effects of these mutations alone, we introduced them as single mutations by site directed mutagenesis, creating gea2-610 and gea2-611, respectively. We tested the growth rates of these mutants along with several of the original single or double mutants (Figure 2C). All failed to grow at 37°C. Mutants gea2-28, gea2-74, and gea2-610 had the most severe growth defects, with no growth at 35.5, 36, or 37°C (Figure 2C). The mutant gea2-60 has the next strongest growth defect, with no growth at 36 or 37°C (Figure 2C). We have focused further studies on these four mutants.

View larger version (41K): [in this window] [in a new window]   Figure 2. Results of GEA2 mutagenesis. (A) Distribution of single and double amino acid substitutions along the Gea2p polypeptide. Some amino acids substitutions were found in multiple mutants. The region of Gea2p corresponding to the Sec7 domain is shaded red, and the region we focused on (shown in sequence alignment in B, is shaded in blue. (B) Domain structure of Gea2p showing homology regions (Cox et al., 2004). The portion of the B/G region (dark blue) that was indicated in A is shown in light blue, and its sequence, aligned with the homologous region from other Arf GEFs of the same subfamily, is shown. Amino acid substitutions in the four gea2-ts mutants analyzed further are indicated. (C) Growth of the indicated gea2-ts mutants at various temperatures.

The localization of the temperature-sensitive gea2p-GFP protein was examined in the mutant strains. In the Gea2p-GFP control, the pattern observed was typical of the Golgi apparatus in yeast: multiple spots distributed throughout the cells (Figure 3). In the gea2-28, gea2-60, gea2-74, and gea2-610 mutants, some spots were still visible after 20 min of shift to the nonpermissive temperature (Figure 3). Hence each of these four gea2-ts proteins is partially released from Golgi membranes, but a portion remains membrane-associated.

View larger version (53K): [in this window] [in a new window]   Figure 3. Localization of Gea2p-GFP, gea2-28p-GFP, gea2-60p-GFP, gea2-74-GFP, and gea2-610-GFP proteins. CJY101 cells carrying either SKP20 (pCEN-LEU2-GEA2-GFP), SKP21 gea2-28, SKP22 gea2-60, SKP23 gea2-74, or SKP24 gea2-610 were grown at 24°C to exponential phase and then either left at 24°C or shifted to 37°C for 20 min before imaging the GFP signal in living cells as described in Materials and Methods. Scale bar, 5 µm.

View larger version (88K): [in this window] [in a new window]   Figure 4. Localization of Sec63p-GFP, Och1p-HA, Sec21p-GFP, Sec7p-GFP, and FAPPI(PH)-GFP in the gea2-ts mutants and in the control strain carrying wild-type GEA2. The strains used were CJY101/pSKP60 + pMS329 (SEC63-GFP), CJY101/pSKP60 + pOH (OCH1-HA), CJY101/pSKP60 + pCS164 (FAPPI-PH), CJY102 SEC21-GFP + pSKP60, and CJY103 SEC7-GFP/pSKP60 (for the GEA2 control), and the same strains but carrying pSKP61 gea2-28, pSKP62 gea2-60, pSKP63 gea2-74, or pSKP64 gea2-610 in place of pSKP60. Cells were grown at 24°C to exponential phase and then shifted to 37°C for 20 min and either viewed directly (Sec63p-GFP, Sec21p-GFP, Sec7p-GFP, and FAPPI(PH)-GFP) or prepared for immunofluorescence analysis (Och1p-HA) and imaged as described in Materials and Methods. Scale bar, 5 µm.

To examine the ultrastructure of the membranes accumulated in the gea2-ts mutants, we carried out immuno-EM analysis on the mutants expressing Och1p-HA from a low-copy centromeric plasmid. A typical Golgi element in a wild-type strain has the appearance of a curved tubular network structure, sometimes forming a complete circle (Preuss et al., 1992; Rambourg et al., 2001). In the control strain labeled with anti-HA antibodies, typical Golgi structures were observed (Figure 5A). The gea2-28 mutant showed accumulation of abnormal, multilayered Golgi structures, an example of which is shown in Figure 5B. Golgi structures labeled with Och1p-HA in the gea2-60, gea2-74, and gea2-610 mutants were also complex, multilayered structures that differed significantly from the control structures (Figure 5, C–E). The Golgi elements in these mutants appeared to be spherical with accumulation of tubular networks in the interior of the structure. Och1p-HA was present in the outer layer of these structures as well as in the interior membranes (Figure 5, B–E). In gea2-28, the majority of structures appear as concentric circles of tubular network (that probably corresponds to a spiral structure in three dimensions), whereas gea2-60 and gea2-610 have many structures like those shown in Figure 5, C and D, with a more random orientation of the internal membrane structures. In all of the mutants, we observed Och1p-HA labeling in the ER, which was accumulated to a greater extent than in wild-type for all four mutants, consistent with the results of the immunofluorescence experiments. An example of this phenotype for gea2-74 is shown (Figure 5F). Hence the early Golgi structures accumulated in the gea2-ts mutants have an abnormal structure compared with the control strain, and accumulation of Och1p-HA in the ER is confirmed in all of the mutants by immuno-EM analysis.

View larger version (155K): [in this window] [in a new window]   Figure 5. Electron microscopy analysis of structural organization of Golgi membranes in the gea2-ts mutants and the GEA2 control strain. CJY52-10-2/pSKP20 (CEN-LEU2-GEA2-GFP), CJY52–10-2/pSKP21, CJY52–10-2/pSKP22, CJY52-10-2/pSKP23, and CJY52-10-2/pSKP24 cells were prepared as described in Materials and Methods. Immunocytochemistry was performed on thin frozen sections of the cells using a rat anti-HA antibody, rabbit anti-rat bridging antibody, and 10-nm protein A-gold to visualize Och1p-HA. N, nucleus; G, Golgi; m, mitochondrion; V, vacuole; ER, endoplasmic reticulum. Scale bars, 0.2 µm.

View larger version (60K): [in this window] [in a new window]   Figure 6. Secretion defects in the gea2-ts mutants. Strains CJY101/pSKP20, CJY101/pSKP21, CJY101/pSKP22, CJY101/pSKP23, and CJY101/pSKP24 were grown at 37°C for 60 min and then subjected to pulse-chase analysis as described in Materials and Methods. Chase times are shown in minutes in each panel. (A) Cell lysates were immunoprecipitated with anti-alkaline phosphatase antibodies and the immunoprecipitates visualized by SDS-PAGE and autoradiography. The precursor form (pALP) and mature vacuolar cleaved form (mALP) are indicated. (B) Cell lysates were immunoprecipitated with anti-CPY antibodies, and the p1 ER form, p2 Golgi form, and mature vacuolar form (mCPY) are indicated. (C) After labeling and chase, cells were removed, the medium was TCA precipitated, and pellets containing proteins secreted into the medium were visualized by SDS-PAGE and autoradiography.

The current model of trafficking through the Golgi apparatus, the cisternal maturation model, postulates that anterograde transport is accomplished through retrograde transport of Golgi enzymes from later back to earlier compartments. Hence the prediction of this model is that any mutant exhibiting a strong anterograde secretion defect will also have a strong retrograde defect. This has been demonstrated for mutants in several COPI subunits, in the Arf1p GTPase, in subunits of the COG complex and in the Neo1p aminophospholipid translocase (Gaynor and Emr, 1997; Sato et al., 2001; Yahara et al., 2001; Suvorova et al., 2002; Hua and Graham, 2003). We first used an assay developed by Schmitt and colleagues to test retrograde transport in the gea2-ts mutants (Ballensiefen et al., 1998). The assay utilizes a Sec22 fusion protein with a luminal myc epitope, Kex2p cleavage site and -factor repeat (Sec22-myc-). In this assay, the extent to which the Sec22-myc- chimera comes in contact with the trans-Golgi localized Kex2p protease is monitored. This Sec22-myc- chimeric protein, like the original Sec22p SNARE protein, cycles between the ER and early Golgi (Ballensiefen et al., 1998). Normally very little Sec22-myc- arrives at the late Golgi compartment containing Kex2p, so in wild-type cells, the majority of the protein is found in its full-length form and only a small amount of Kex2p cleavage product is observed (Figure 7). In contrast, mutants defective in retrograde trafficking such as ret1-1 (defective in -COP) had a marked accumulation of the cleavage product, with little if any full-length protein present (Figure 7). Surprisingly, none of the gea2-28, gea2-60, gea2-74, and gea2-610 mutants had a defect in this assay, each showing a phenotype indistinguishable from wild-type control strains (Figure 7, A and B). The gea1-6 gea2 mutant that we have previously shown to have a retrograde trafficking defect (Chantalat et al., 2003) accumulated a large amount of cleaved product in this assay (Figure 7C). Even a gea2 mutant alone had a defect in the assay, accumulating significantly more cleavage product than the wild-type control (Figure 7C), whereas a gea1 mutant had a phenotype indistinguishable from wild type (Figure 7C). This is the first demonstration of a difference in phenotype in a trafficking assay between the gea1 and gea2 mutants and suggests that Gea1p and Gea2p may act preferentially in different sorting pathways within the Golgi. In conclusion, we find that the gea2-ts mutants have no defect in retrograde trafficking using the Sec22-myc- assay, unlike COPI, arf1-ts and even gea1-ts mutants, which have severe defects in this assay.

View larger version (54K): [in this window] [in a new window]   Figure 7. The gea2-ts mutants are not defective in retrograde trafficking of Sec22p-myc-. Strains BY4742, CJY101/pSKP21 gea2-28, CJY101/pSKP23 gea2-74, and EGY101 ret1-1 (A); strains CJY101/pSKP20 GEA2, CJY101/pSKP22 gea2-60, CJY101/pSKP24 gea2-610, and EGY101 ret1-1 (B); and strains CJY49-11-4 gea2, CJY62-10-3 gea1-4, APY022 gea1-6, and CJY50 gea1 (C) were grown overnight at 24°C, and then cultures were split and either kept at 24°C or shifted to 37°C for 1 or 2 h. Cell lysates were prepared and subjected to Western blotting analysis using anti-Sec22p antibodies. The full-length Sec22-myc- chimeric protein, the Kex2p cleavage product (Kex2 CP: Sec22-myc) and endogenous Sec22p are indicated. Western blots were also performed using anti-myc antibodies, which detected the top two bands only (our unpublished data).

Rer1p is a cargo receptor that localizes to the Golgi apparatus at steady state and cycles between the ER and the Golgi (Sato et al., 2001). It is transported from the Golgi to the ER by a COPI-dependent mechanism (Sato et al., 1997, 2001). In both COPI and Arf1p mutants, Rer1p mislocalizes to the vacuole because retrograde trafficking is blocked and the protein that is accumulated in the Golgi is transported to the vacuole for degradation, where the GFP moiety is cleaved and accumulates in the vacuolar lumen (Sato et al., 2001; Yahara et al., 2001). Hence Rer1p localization is a sensitive test for a defect in retrograde trafficking. At the permissive temperature of 24°C, the gea2-28, gea2-60, gea2-74 and gea2-610 mutants had a phenotype identical to that of the GEA2 control: scattered spots throughout the cytoplasm (Figure 8). At 37°C, GFP-Rer1p localized to Golgi elements in all the mutants, with some of the spots considerably larger in the mutants (Figure 8). In addition, all of the mutants showed a portion of GFP-Rer1p in the ER (Figure 8). These results are strikingly different from those obtained for the COPI mutants ret1-1 (-COP), sec21-2 (-COP), and sec27-1 ('-COP) and several ts mutants of Arf1p, all of which showed redistribution of GFP-Rer1p to the vacuole (Sato et al., 2001, Yahara et al., 2001). Our results suggest that retrograde transport of GFP-Rer1p is intact in all of the mutants and that anterograde transport of the protein is slowed or blocked at the nonpermissive temperature.

View larger version (90K): [in this window] [in a new window]   Figure 8. GFP-Rer1p and GFP-Emp46p do not accumulate in the vacuole in the gea2-ts mutants. Strains CJY101/pSKP60, CJY101/pSKP61, CJY101/pSKP62, CJY101/pSKP63, and CJY101/pSKP64 were transformed with plasmids expressing GFP-Rer1p or GFP-Emp46p were grown at 24°C overnight and either shifted to 37°C for 30 min or maintained at 24°C before imaging. Scale bar, 5 µm.

It is currently believed that the COPI coat complex in yeast functions primarily in retrograde trafficking from the Golgi back to the ER, and in intra-Golgi retrograde trafficking (Lee et al., 2004). Consistent with the results that the gea2-ts mutants have no defect in retrograde trafficking, COPI was found in membrane structures in these mutants even after 45 min of incubation at the nonpermissive temperature (Figure 4 and our unpublished data). COPI is normally recruited to and released from membranes in a rapid cycle that depends on Arf1 activation for the membrane recruitment step (Donaldson and Jackson, 2000; Presley et al., 2002). This cycle is blocked by the drug BFA, which binds to an Arf GEF-Arf1-GDP complex and prevents Arf1 activation (Peyroche et al., 1999). In mammalian cells, COPI release from membranes by BFA has been extensively studied, but this phenomenon has not been demonstrated in yeast. Wild-type yeast cells are resistant to BFA, but an erg6 mutant, defective in one of the final steps of ergosterol biosynthesis, is sensitive (Graham et al., 1993; Peyroche and Jackson, 2001). To determine the effect of BFA on COPI binding to membranes in yeast, we first tested a standard wild-type strain in which the BFA-sensitive erg6 mutation had been introduced and which carried a GFP-tagged version of -COP (Sec21p-GFP) integrated into the chromosome. In these cells, Sec21p-GFP was completely released from membranes within 1 min of treatment with BFA (Figure 9A, left panels). We have previously shown that GBF1, the mammalian homologue of Gea2p, remains associated with Golgi membranes in the presence of BFA (Niu et al., 2005), and hence we examined the localization of Gea2p-GFP after BFA treatment in yeast. As in the case of GBF1, Gea2p-GFP remained associated with Golgi membranes after BFA treatment (Figure 9A, right panels). Hence the dynamic association of COPI and Gea2p as revealed by BFA treatment is very similar in yeast and mammalian cells.

View larger version (99K): [in this window] [in a new window]   Figure 9. The COPI associated with membranes in gea2-ts and sec7-ts mutants is recruited by a BFA-sensitive Arf GEF. (A) CJY104 (BY4742 erg6 with SEC21-3xGFP integrated into the SEC21 chromosomal locus) and BY4742 erg6/pSKP20 GEA2–3xGFP were grown at 30°C, imaged as described in Materials and Methods (- BFA), and then treated with 100 µg/ml BFA and imaged 1 min later (+ BFA). (B) Strains CJY105/pSKP60 + pAP45, CJY105/pSKP60 + pSKP72, CJY105/pSKP60 + pSKP78, CJY105/pSKP63 + pAP45, CJY105/pSKP63 + pSKP72, CJY105/pSKP60 + pSKP78, CJY105/pSKP64 + pAP45 CJY105/pSKP64 + pSKP72, and CJY105/pSKP64 + pSKP78 were grown at 37°C for 30 min and imaged and then immediately treated with 100 µg/ml BFA for 1 min and imaged again. (C) Strain CJY105/pSKP63 CEN-LEU2-gea2-74 + pSKP72 CEN-TRP1-sec7-12 was grown at 37°C and imaged and then treated with 100 µg/ml BFA and the same field was imaged after the amount of time indicated. Cells were then washed by removing medium containing BFA, washing twice with fresh medium lacking BFA, and imaged again after 2 min (Washout). (Note that a different field of cells was imaged after BFA washout.) Scale bar, 5 µm.

We have shown previously that the only BFA-sensitive GEFs in the yeast secretory pathway are Gea1p, Gea2p, and Sec7p (Peyroche et al., 1999). Hence the BFA-sensitive GEF responsible for recruiting COPI to membranes in the gea2-ts mutants could either be the mutant gea2 protein itself (which has a wild-type Sec7 domain) or the wild-type Sec7p present in the strain. We attempted to distinguish these possibilities by constructing erg6 gea2-ts sec7-ts strains expressing Sec21p-GFP. We have isolated new sec7-ts alleles in a screen similar to that described above for the gea2-ts mutants (our unpublished data). Two of these mutants, sec7-12 and sec7-18, are tight ts alleles that show no growth at 37°C. sec7-12 contains a single mutation, I615N, in the B/G region of Sec7p, which is conserved in both the Gea/GBF and Sec7/BIG subfamilies of Arf GEFs. sec7-18 also contains a mutation in the B/G region (V636D), but in addition carries two downstream substitutions (L744S and N839D), the latter in a highly conserved residue near the beginning of the Sec7 domain. The erg6 gea1 gea2/pCEN-gea2-ts strain described above has the chromosomal copy of SEC7 deleted, and carries a wild-type SEC7 gene on a plasmid. We also constructed strains carrying plasmid-borne copies of sec7-12 and sec7-18 in place of wild-type SEC7 (see Materials and Methods), to create erg6 gea2-ts sec7-ts strains carrying Sec21p-GFP integrated into the chromosome. In the single sec7-ts mutants, as in the gea2-ts single mutants, Sec21p-GFP was associated with membrane structures at 37°C (Figure 9B, left panels), indicating that these sec7-ts mutations did not abolish the ability of cells to recruit COPI to membranes.

In the double mutants, Sec21p-GFP structures were still visible after a 30-min incubation at the nonpermissive temperature (although the pattern was altered in some of the double mutants; see below). In all four mutants, BFA caused release of Sec21p-GFP from the membrane structures present in these cells (Figure 9B). We imaged the same field of erg6 gea2-74 sec7-12 Sec21p-GFP double mutant cells before and after BFA treatment to monitor the time course of COPI release from Golgi membranes. As soon as 30 s after BFA treatment, all COPI was released from membranes into the cytoplasm (Figure 9C). When cells were then washed to remove the BFA, Sec21p-GFP fluorescence was restored to a punctate pattern within 2 min, indicating that COPI had regained its Golgi localization. These results demonstrate that in the mutant strains, Sec21p membrane localization requires activated Arf1/Arf2p to be maintained. However, in some of the double mutant strains, the pattern of spots was different from the wild-type control strain (Figure 9B). The change was particularly evident in the gea2-610 sec7-18 double mutant, which had considerably fewer Sec21p-GFP structures that were much smaller in size. This result shows that both Gea1/Gea2p and Sec7p contribute to maintenance of normal-sized COPI-containing membrane structures in yeast cells, although they do not allow us to conclude whether, at the molecular level, it is exclusively Gea2p or Sec7p that activates Arf1/Arf2p to recruit COPI to membranes.

The results above allow us to make several important conclusions. The first is that Gea2p and Sec7p do play redundant roles in maintaining COPI-associated structures in cells. This could either be through maintenance of the membrane domains that COPI associates with, or through direct recruitment of COPI via Arf1 activation. We are currently carrying out experiments to distinguish these two possibilities. The second conclusion is that the COPI present on membrane structures in the single and double gea2-ts sec7-ts mutants is recruited to membranes by activated Arf1/Arf2p, and hence they are not novel structures where the protein is aggregated or associated with membranes by some other mechanism. Hence in the gea2-74 and gea2-610 mutants, COPI is recruited to membranes by activated Arf1/Arf2p, despite the fact that anterograde transport is almost completely blocked. These results demonstrate that COPI recruitment to membranes and the functioning of at least some retrograde trafficking pathways can be uncoupled from anterograde transport.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
In this article, we provide the first in vivo functional data for one of the most highly conserved regions of the Gea/GBF subfamily of Arf GEFs outside of the Sec7 domain, the B/G domain. The Golgi-localized Arf GEFs are large proteins (200 kDa) with multiple homology domains flanking the catalytic Sec7 domain (Jackson and Casanova, 2000; Cox et al., 2004). Three of these homology domains also share similarity to the BIG/Sec7 subfamily of Arf GEFs and are called B/G, B/G, and B/G (Cox et al., 2004). The functions of the evolutionarily conserved domains in the Golgi-localized Arf GEFs are largely unknown, although some information about binding partners of these regions is emerging. The G domain of the GNOM protein of Arabidopsis thaliana was shown to be required for dimerization of the protein and for cyclophilin binding, although the functional significance of this interaction is not known (Grebe et al., 2000). Both Gea1p and Gea2p interact with the transmembrane protein Gmh1p via the conserved domain B/G, which lies immediately downstream of the Sec7 domain (Chantalat et al., 2003). A single amino acid within this domain is mutated in the gea1-6 temperature-sensitive mutant, and analysis of the mutant indicates that this region is essential for Gea1p function, both in anterograde and retrograde trafficking (Peyroche et al., 2001; Chantalat et al., 2003). These essential functions are not solely mediated by Gmh1p, as GMH1 itself is dispensable for growth and secretion (Chantalat et al., 2003). A proline-rich domain at the C-terminal end of the GBF1 protein has been shown to interact with p115, a peripherally associated membrane protein involved in SNARE function (Garcia-Mata and Sztul, 2003). In this article, we have demonstrated that the B/G domain of Gea2p, which lies just upstream of the catalytic Sec7 domain, is an important functional domain of this Arf GEF in yeast.

We have isolated and characterized four temperature-sensitive mutants in the B/G region of the S. cerevisiae Gea2p Arf GEF. Three of these four mutants have a complete or almost complete block in anterograde transport, but are not defective in retrograde trafficking of three proteins that normally cycle between ER and Golgi in a COPI-dependent manner. COPI is present on membranes in all four mutants and is released from membranes by BFA at the nonpermissive temperature. As determined by immunofluorescence and electron microscopy, both ER and early Golgi membranes accumulate.

The mutants examined here with amino acid substitutions in the B/G region of Gea2p have some similarities to COPI mutants, in that they continue to secrete COPI-independent cargo such as HSP150, and block anterograde trafficking of proteins traversing the secretory pathway (such as CPY) in the ER and early Golgi. However, there are a number of differences between the gea2-ts mutants that we have characterized in this article and previously characterized COPI, Arf1p, and gea1-ts mutants. Rer1p or Emp46 are not transported to the vacuole in the gea2-ts mutants as they are in COPI and arf1-ts mutants, Sec22-myc- is efficiently sorted away from Kex2p in the gea2-ts mutants, but not in COPI, arf1-ts or gea1-ts mutants, and Och1p accumulates in the ER in the gea2-ts mutants but not in COPI and arf1-ts mutants. Och1p normally cycles between the ER and the Golgi in yeast (Karhinen and Makarow, 2004), so the accumulation of Och1p-HA in the ER is indicative of an anterograde trafficking block. Hence secreted proteins, proteins destined for the vacuole, a glycosylation enzyme and cargo receptors are all slowed or blocked in their anterograde transport through the Golgi in the gea2-ts mutants.

At the morphological level, both ER and early Golgi membranes accumulate in the gea2-ts mutants. These membranes contain early Golgi markers, considerable amounts of Rer1p, but not late Golgi resident proteins. These results show that there is a block in trafficking within the Golgi, so that membranes beyond the block are evacuated in the normal process of secretion, and membranes upstream of the block accumulate. Although it is certainly possible that the B/G region of Gea2p that is defective in the mutants is involved in recycling of a subset of proteins that are essential for anterograde transport, another interesting possibility is that this region is involved directly in anterograde transport, perhaps through effects on membrane lipids. We have preliminary evidence that at least one lipid-modifying enzyme interacts with this portion of Gea2p, and current studies are aimed at testing this possibility.

The best-characterized role to date of the Arf1 GTPase and its regulators is the recruitment of COPI to Golgi membranes to promote retrograde trafficking within the Golgi and from the Golgi back to the ER (Bonifacino and Lippincott-Schwartz, 2003; Lee et al., 2004). The current model for transport through the Golgi, the cisternal maturation model, postulates that anterograde transport occurs as a result of COPI-mediated retrograde transport of Golgi glycosylation enzymes and resident proteins. Contrary to the simplest interpretation of the cisternal maturation model, our results indicate that the Arf1p activator Gea2p plays a role in anterograde transport that is not a consequence of a global defect in retrograde trafficking. These results rule out the simplest type of cisternal maturation model for trafficking through the Golgi, which postulates a common COPI-dependent mechanism and indicate either that Gea2p plays a direct role in anterograde transport or that there are multiple mechanisms of retrograde transport important for cisternal maturation. This latter possibility is consistent with recent results showing that there are different populations of COPI-coated vesicles in mammalian cells that mediate retrograde transport of distinct cargos (Malsam et al., 2005).

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
We thank Carolyn Phillips for excellent assistance in isolation of the gea2-ts mutants and their initial characterization and to Emily Rainey for isolation of the sec7-ts mutants. We thank Ben Glick, Hans Dieter Schmitt, Ken Sato, Matthias Seedorf, Chris Stefan, Scott Emr, and Gerry Waters for plasmids; François Letourneur for strains; Gerry Waters for Sec22p antiserum; and Dale Hailey for assistance with microscopy.

	Footnotes

 
This article was published online ahead of print in MBC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E05-04-0289) on June 1, 2005.

Address correspondence to: Catherine L. Jackson (cathyj{at}helix.nih.gov).

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Ballensiefen, W., Ossipov, D., and Schmitt, H. D. ((1998). ). Recycling of the yeast v-SNARE Sec22p involves COPI-proteins and the ER transmembrane proteins Ufe1p and Sec20p. J. Cell Sci. 111, (Pt 11), 1507-1520.[Abstract]

Bonfanti, L., Mironov, A. A., Jr., Martinez-Menarguez, J. A., Martella, O., Fusella, A., Baldassarre, M., Buccione, R., Geuze, H. J., Mironov, A. A., and Luini, A. ((1998). ). Procollagen traverses the Golgi stack without leaving the lumen of cisternae: evidence for cisternal maturation. Cell 95, , 993-1003.[CrossRef][Medline]

Bonifacino, J. S., and Lippincott-Schwartz, J. ((2003). ). Coat proteins: shaping membrane transport. Nat. Rev. Mol. Cell. Biol. 4, , 409-414.[CrossRef][Medline]

Chantalat, S., Courbeyrette, R., Senic-Matuglia, F., Jackson, C. L., Goud, B., and Peyroche, A. ((2003). ). A novel Golgi membrane protein is a partner of the ARF exchange factors Gea1p and Gea2p. Mol. Biol. Cell 14, , 2357-2371.[Abstract/Free Full Text]

Cox, R., Mason-Gamer, R. J., Jackson, C. L., and Segev, N. ((2004). ). Phylogenetic analysis of Sec7-domain-containing Arf nucleotide exchangers. Mol. Biol. Cell 15, , 1487-1505.[Abstract/Free Full Text]

de Graffenried, C. L., and Bertozzi, C. R. ((2004). ). The roles of enzyme localisation and complex formation in glycan assembly within the Golgi apparatus. Curr. Opin. Cell Biol. 16, , 356-363.[CrossRef][Medline]

Donaldson, J. G., and Jackson, C. L. ((2000). ). Regulators and effectors of the ARF GTPases. Curr. Opin. Cell Biol. 12, , 475-482.[CrossRef][Medline]

Duden, R. ((2003). ). ER-to-Golgi transport: COP I and COP II function (Review). Mol. Membr. Biol. 20, , 197-207.[CrossRef][Medline]

Garcia-Mata, R., and Sztul, E. ((2003). ). The membrane-tethering protein p115 interacts with GBF1, an ARF guanine-nucleotide-exchange factor. EMBO Rep. 4, , 320-325.[CrossRef][Medline]

Gaynor, E. C., and Emr, S. D. ((1997). ). COPI-independent anterograde transport: cargo-selective ER to Golgi protein transport in yeast COPI mutants. J. Cell Biol. 136, , 789-802.[Abstract/Free Full Text]

Glick, B. S., and Malhotra, V. ((1998). ). The curious status of the Golgi apparatus. Cell 95, , 883-889.[CrossRef][Medline]

Godi, A., Di Campli, A., Konstantakopoulos, A., Di Tullio, G., Alessi, D. R., Kular, G. S., Daniele, T., Marra, P., Lucocq, J. M., and De Matteis, M. A. ((2004). ). FAPPs control Golgi-to-cell-surface membrane traffic by binding to ARF and PtdIns(4)P. Nat. Cell Biol. 6, , 393-404.[CrossRef][Medline]

Graham, T. R., Scott, P. A., and Emr, S. D. ((1993). ). Brefeldin A reversibly blocks early but not late protein transport steps in the yeast secretory pathway. EMBO J. 12, , 869-877.[Medline]

Grebe, M., Gadea, J., Steinmann, T., Kientz, M., Rahfeld, J. U., Salchert, K., Koncz, C., and Jurgens, G. ((2000). ). A conserved domain of the Arabidopsis GNOM protein mediates subunit interaction and cyclophilin 5 binding. Plant Cell 12, , 343-356.[Abstract/Free Full Text]

Harris, S. L., and Waters, M. G. ((1996). ). Localization of a yeast early Golgi mannosyltransferase, Och1p, involves retrograde transport. J. Cell Biol. 132, , 985-998.[Abstract/Free Full Text]

Hua, Z., and Graham, T. R. ((2003). ). Requirement for neo1p in retrograde transport from the Golgi complex to the endoplasmic reticulum. Mol. Biol. Cell 14, , 4971-4983.[Abstract/Free Full Text]

Jackson, C. L., and Casanova, J. E. ((2000). ). Turning on ARF: the Sec7 family of guanine-nucleotide-exchange factors. Trends Cell Biol. 10, , 60-67.[CrossRef][Medline]

Karhinen, L., and Makarow, M. ((2004). ). Activity of recycling Golgi mannosyltransferases in the yeast endoplasmic reticulum. J. Cell Sci. 117, , 351-358.[Abstract/Free Full Text]

Kweon, H. S. et al. ((2004). ). Golgi enzymes are enriched in perforated zones of golgi cisternae but are depleted in COPI vesicles. Mol. Biol. Cell 15, , 4710-4724.[Abstract/Free Full Text]

Lee, M. C., Miller, E. A., Goldberg, J., Orci, L., and Schekman, R. ((2004). ). Bi-directional protein transport between the ER and Golgi. Annu. Rev. Cell. Dev. Biol. 20, , 87-123.[CrossRef][Medline]

Liou, W., Geuze, H. J., and Slot, J. W. ((1996). ). Improving structural integrity of cryosections for immunogold labeling. Histochem. Cell Biol. 106, , 41-58.[CrossRef][Medline]

Lupashin, V. V., and Waters, M. G. ((1997). ). Trichloroacetic acid-SNARE activation through transient interaction with a rab-like guanosine triphosphatase. Science 276, , 1255-1258.[Abstract/Free Full Text]

Malsam, J., Satoh, A., Pelletier, L., and Warren, G. ((2005). ). Golgin tethers define subpopulations of COPI vesicles. Science 307, , 1095-1098.[Abstract/Free Full Text]

Muhlrad, D., Hunter, R., and Parker, R. ((1992). ). A rapid method for localized mutagenesis of yeast genes. Yeast 8, , 79-82.[CrossRef][Medline]

Nakano, A. ((2004). ). Yeast Golgi apparatus— dynamics and sorting. Cell Mol. Life Sci. 61, , 186-191.[CrossRef][Medline]

Niu, T. K., Pfeifer, A. C., Lippincott-Schwartz, J., and Jackson, C. L. ((2005). ). Dynamics of GBF1, a brefeldin A-sensitive Arf1 exchange factor at the Golgi. Mol. Biol. Cell 16, , 1213-1222.[Abstract/Free Full Text]

Peyroche, A., Antonny, B., Robineau, S., Acker, J., Cherfils, J., and Jackson, C. L. ((1999). ). Brefeldin A acts to stabilize an abortive ARF-GDP-Sec7 domain protein complex: involvement of specific residues of the Sec7 domain. Mol. Cell 3, , 275-285.