SNARE (soluble N-ethylmaleimide-sensitive fusion protein-attachment protein receptor) proteins contribute to specific recognition between transport vesicles and target membranes and are required for fusion of membranes. The SNARE Vti1p is required for several transport steps between late Golgi, endosomes and the vacuole in the yeast Saccharomyces cerevisiae. Here, we identified the late Golgi membrane protein TVP23 as a multicopy suppressor of the growth defect in vti1-2 cells. By contrast, the growth defect in vti1-11 cells was not suppressed by TVP23 overexpression. Deletion of TVP23 aggravated the growth defect in vti1-2 cells. Genetic interactions between TVP23 and vti1-2 were not found in transport from the late Golgi via the late endosome to the vacuole or in transport from the Golgi directly to the vacuole. These results suggest that Tvp23p is not involved in forward transport from the late Golgi. Therefore retrograde traffic to the late Golgi was analysed. vti1-2 cells accumulated GFP (green fluorescent protein)–Snc1p within the cell, indicating that retrograde transport from the early endosome to the late Golgi was defective in these cells. Deletion of TVP23 in vti1-2 cells resulted in a synthetic defect in GFP–Snc1p recycling, whereas tvp23Δ cells had a slight defect. These results indicate that Tvp23p performs a partially redundant function in retrograde transport from the early endosome to the late Golgi. This transport step was unaffected in vti1-11 cells, providing an explanation for the allele-specific multicopy suppression by TVP23.

INTRODUCTION

Membrane traffic in eukaryotic cells is mediated by transport vesicles, which bud from a donor organelle and fuse with a target membrane. Several protein–protein interactions are required for recognition between membranes. Members of the Rab/Ypt family of small GTPases have to be recruited to the vesicle membrane and activated by GEFs (guanine-nucleotide-exchange factors), which exchange GDP for GTP [1]. Tether proteins are associated with target membranes. They are hetero-oligomeric complexes or large coiled-coil proteins [2]. Some tether proteins act as GEFs for Rabs. On the other hand, some tether proteins function as Rab effectors, which bind only to the active GTP form. These interactions are required for the tethering of transport vesicles to target membranes. In addition, tether proteins form complexes with SNARE (soluble N-ethylmaleimide-sensitive fusion protein-attachment protein receptor) proteins. Specific SNAREs are found on the transport vesicle and target membrane [3]. SNAREs contain a C-terminal transmembrane domain or a post-translational hydrophobic modification. They form SNARE complexes via the conserved SNARE motifs next to the membrane anchor. These SNARE complexes consist of a bundle of four different parallel α-helices. Side chains from all four helices interact in 16 stacked layers in the core of the bundle. SNAREs can be classified as R-, Qa-, Qb- and Qc-SNAREs according to sequence similarities [4]. Functional SNARE complexes are composed of one member of each class. SNARE complex formation bridges the gap between both membranes and is required for fusion of both membranes.

The baker's yeast Saccharomyces cerevisiae is a powerful model to study the molecular mechanism of membrane traffic. Several trafficking pathways leave from the late Golgi in addition to the secretory route to the plasma membrane [5]. Most vacuolar proteins are transported from the late Golgi to the late endosome, also called multivesicular body or prevacuolar compartment, and in a second transport step to the vacuole, the equivalent of the mammalian lysosome. The vacuolar CPY (carboxypeptidase Y) serves as a marker for this pathway. By contrast, the vacuolar ALP (alkaline phosphatase) traffics from the late Golgi to the vacuole without a passage through the late endosome. In addition, there is traffic from the Golgi to the early endosome. Endocytosed proteins destined for degradation are transported from early endosomes via late endosomes to the vacuole. On the other hand, there is retrograde transport from the early endosome to the late Golgi to recycle proteins from the plasma membrane. Escaped late Golgi proteins are retrieved from the early endosome to maintain the identity of the late Golgi. Retrograde traffic is also observed from the late endosome to the late Golgi. This recycling route is taken by the CPY receptor Vps10p. The SNARE protein Vti1p- is essential for these transport steps as part of three different SNARE complexes [6,7]. The SNARE complex consisting of Ykt6p, Pep12p, Vti1p and Syn8p is required for transport to the late endosome in addition to the Rab/Ypt GTPase Vps21p and the CORVET (core vacuole/endosome tethering) tethering complex [810]. Ypt7p, the HOPS tethering complex and a SNARE complex consisting of Ykt6p, Vam3p, Vti1p and Vam7p are required for fusion with the vacuole [11]. The hetero-oligomeric GARP (Golgi-associated retrograde protein complex)–VFT (Vps fifty-three complex) tethering complex interacts with Ypt6p and the SNARE Tlg1p in retrograde traffic from endosomes to the late Golgi [12,13]. The SNARE complex in retrograde traffic to the late Golgi consists of Snc1p or Snc2p, Tlg2p, Vti1p and Tlg1p [14,15]. Snc1p/Snc2p is also required on secretory vesicles for secretion together with the plasma-membrane SNAREs Sso1p/Sso2p and Sec9p, which contributes two helices to the SNARE complex [16]. After incorporation into the plasma membrane, Snc1p/Snc2p is endocytosed and recycled via early endosomes to the late Golgi for another round of secretory vesicle formation [17].

In addition, Vti1p is involved in retrograde transport to the cis-Golgi, probably in a complex with Ykt6p, Sed5p and Sft1p [6,18]. We have been studying the function of Vti1p with the help of different temperature-sensitive mutants. vti1-11 cells do not grow at the restrictive temperature and are blocked in transport to the late endosome, to the vacuole and in retrograde traffic to the cis-Golgi. vti1-2 cells are defective in transport to the late endosome and in fusion with the vacuole as well as in growth at the restrictive temperature [6,7]. The growth defect of vti1-2 cells was used for a multicopy suppressor screen to identify additional components of these transport pathways.

Here, we identified the late Golgi protein TVP23 as an allele-specific genetic interaction partner of VTI1. We found that Tvp23p is involved in retrograde transport from the early endosome to late Golgi. We describe that this transport step is also defective in vti1-2 cells.

EXPERIMENTAL

Materials

Reagents were used from the following sources: enzymes for DNA manipulation from New England Biolabs (Beverly, MA, U.S.A.), [35S]methionine and [35S]cysteine (EXPRE35S35S protein labelling mix) from PerkinElmer (Waltham, MA, U.S.A.), fixed Staphylococcus aureus cells (Pansorbin) from Calbiochem (San Diego, CA, U.S.A.), zymolyase from Seikagaku (Tokyo, Japan) and FM4-64 from Invitrogen (Carlsbad, CA, U.S.A.).

Plasmid manipulations were performed in the Escherichia coli strain DH5α using standard media. Yeast strains were grown in rich media YEPD (1% yeast extract/1% peptone/2% dextrose) or SD minimal medium (standard minimal medium) with appropriate supplements.

Strains and plasmids

Yeast strains were constructed using standard genetic techniques. The strains FvMY7, FvMY24 and FvMY21 carrying the vti1-1, vti1-2 and vti1-11 mutations respectively in the SEY6210 strain background have been described previously [6,7,19]. tvp23Δ and tlg2Δ strains in BY4742 background were obtained from the Euroscarf strain collection [20]. The tvp23Δ::kanMX4 mutant locus was PCR-amplified from genomic DNA with approx. 200 bp of flanking regions and integrated into SEY6210 to yield ISY5, into FvMY24 resulting in ISY6 and into SNY18 (pho8Δ::ADE2 [21]) to obtain AGY4. A 2μ plasmid encoding TVP23 was generated by subcloning a 1.7 kb HindIII fragment containing nt 612.924–614.653 from chromosome IV from a suppressor plasmid into Yep352 yielding pIS2. pSN55 encoded the A-ALP (cytosolic domain of dipeptidyl aminopeptidase A fused to the transmembrane and luminal domain of alkaline phosphatase) fusion protein [22], pGS416 encoded the GFP (green fluorescent protein)–Snc1p fusion protein [17] and DsRed–Sec7p was expressed from pTPQ128 [23]. The yeast strain AHY41 carries the mutant alleles vps5Δ::HIS3 pho8Δ::ADE2 [27].

Suppressor screen

The growth defect of vti1-2 cells at 37 °C was used to identify multicopy suppressors after transformation with a YEp24 2μ genomic library [24,25].

Transport assays

Yeast cells were grown in exponential phase at 24 °C and 0.5 absorbance units, pre-incubated at the indicated temperature for 15 min or grown and labelled at 30 °C for A-ALP assays. Cells were labelled with [35S]methionine and [35S]cysteine (100 μCi; 0.5 absorbance units) for 10 min and chased for 30 min, and CPY was immunoprecipitated with a rabbit antiserum from cellular extracts (I) and medium (E), whereas ALP was only from cellular extracts. For A-ALP stability assays, the cells were labelled for 25 min and chased for indicated time points, and A-ALP was immunoprecipitated with antiserum directed against ALP from cellular extracts (I) as described in [7,22,26]. Cells lacking the retromer subunit Vps5p were used as a positive control (AHY41, vps5Δ::HIS3 pho8Δ::ADE2 [27]). ALP antiserum was kindly provided by T.H. Stevens (University of Oregon). Immunoprecipitates were analysed by SDS/PAGE and autoradiography. A BAS-1800 II (Fuji) was used for quantification.

Fluorescence microscopy

Yeast cells were grown to an early logarithmic stage in appropriate SD minimal medium at 24 °C for FM4-64 staining, DsRed–Sec7p and GFP–Snc1p localization. For FM4-64 endocytosis, approx. 0.7 absorbance units of cells were pelleted and resuspended in 120 μl of YEPD, and 2 μl of 16 mM FM4-64 was added. After a 5 min incubation at 24 °C, cells were washed once with YEPD and immediately viewed under a fluorescence microscope (Leica DM-5000 B) with the Leica DFC350 FX CCD camera (charge-coupled-device camera). tlg2Δ and tvp23Δ cells were shifted for 1 h to 30 °C and vti1-2 cells to 37 °C for the GFP–Snc1p–DsRed–Sec7p co-localization experiment. For GFP–Snc1p and DsRed–Sec7p localization, cells were washed once with PBS and immediately viewed under a fluorescence microscope.

RESULTS

Identification of TVP23 as a suppressor of the growth defect in vti1-2 cells

vti1-2 cells are defective in transport from the late Golgi to the late endosome, in direct transport from the late Golgi to the vacuole and in homotypic vacuolar fusion. These cells grow very slowly at the restrictive temperature of 37 °C. In order to identify genes co-operating with VTI1, a multicopy suppressor screen was performed utilizing the growth defect of vti1-2 cells at 37 °C. A 2μ plasmid containing nt 609360–617576 of chromosome IV improved the growth of vti1-2 cells at 37 °C (Figure 1A). Five full-length genes are encoded in this segment: the telomere-binding protein Stn1p, the nucleolar protein Rrp8p, Tvp23p, a regulator of pheromone receptor signalling (Afr1p) and the Sss1p subunit of the Sec61p translocation complex. Tvp23p was identified as a membrane protein in the late Golgi or early endosome [28]. As Tvp23p is the only protein encoded on this suppressor plasmid with a possible connection to post-Golgi membrane trafficking, a 2μ plasmid was generated, which encoded only Tvp23p. Over-expression of TVP23 alone was sufficient to improve growth of vti1-2 cells (Figure 1A). Suppression was not observed with fragments of the suppressor plasmid without intact TVP23 (results not shown). This indicates that TVP23 is a multicopy suppressor of the growth defect in vti1-2 cells. tvp23Δ cells grew as well as wild-type cells at all temperatures (Figure 1B). Deletion of TVP23 in vti1-2 cells reduced the slow growth of vti1-2 cells at 37 °C after prolonged incubation even further (Figure 1B). These results show that combination of vti1-2 and tvp23Δ mutations resulted in a synthetic growth defect at high temperatures as additional genetic interaction. Suppression may be due to direct or more general effects. Therefore vti1-11 cells were tested, which are defective in an additional transport step, in retrograde transport to the cis-Golgi. The growth defect of vti1-11 cells at 37 °C was not suppressed by overexpression of TVP23 (Figure 1C). This allele-specific effect suggests that Tvp23p functions in one of the trafficking steps defective in vti1-2 cells.

Overexpression of TVP suppressed and the absence of Tvp23p aggravated the growth defect in vti1-2 cells at 37 °C

Figure 1
Overexpression of TVP suppressed and the absence of Tvp23p aggravated the growth defect in vti1-2 cells at 37 °C

Dilutions of cells were incubated at 24 or 37 °C on plates with rich medium. (A) vti1-2 cells (FvMY24) overexpressing TVP23 either from a 8.2 kb library 2μ plasmid or as the only open reading frame on a 2μ plasmid grew better at 37 °C than vti1-2 cells. Abbreviation: WT, SEY6210 wild-type cells. (B) vti1-2 tvp23Δ double mutant cells (ISY6) grew slower at 37 °C than either vti1-2 or tvp23Δ (ISY5) single-mutant cells after prolonged incubation. (C) The growth defect in vti1-11 cells (FvMY21) was not suppressed by overexpression of TVP23, indicating that the suppression is allele-specific. All mutations were introduced into the SEY6210 background.

Figure 1
Overexpression of TVP suppressed and the absence of Tvp23p aggravated the growth defect in vti1-2 cells at 37 °C

Dilutions of cells were incubated at 24 or 37 °C on plates with rich medium. (A) vti1-2 cells (FvMY24) overexpressing TVP23 either from a 8.2 kb library 2μ plasmid or as the only open reading frame on a 2μ plasmid grew better at 37 °C than vti1-2 cells. Abbreviation: WT, SEY6210 wild-type cells. (B) vti1-2 tvp23Δ double mutant cells (ISY6) grew slower at 37 °C than either vti1-2 or tvp23Δ (ISY5) single-mutant cells after prolonged incubation. (C) The growth defect in vti1-11 cells (FvMY21) was not suppressed by overexpression of TVP23, indicating that the suppression is allele-specific. All mutations were introduced into the SEY6210 background.

Tvp23p was not involved in anterograde traffic to the vacuole

Effects of the tvp23Δ mutation on trafficking were tested. Transport from the late Golgi via the late endosome to the vacuole was investigated first. The vacuolar hydrolase CPY is modified to the pro-form p1CPY in the ER (endoplasmic reticulum) and receives further glycosylation in the Golgi to a slightly larger p2CPY. p2CPY travels via the late endosome to the vacuole; there it is proteolytically cleaved to mCPY with lower molecular mass [5]. p2CPY is secreted into the medium upon block in transport from the late Golgi to the vacuole. Newly synthesized proteins were labelled with [35S]cysteine and [35S]methionine, chased with unlabelled cysteine and methionine, and CPY was immunoprecipitated from cellular extracts (I) and the medium (E). Samples were analysed by SDS/PAGE and autoradiography to follow transport of newly synthesized CPY. CPY reached the vacuole as mCPY in wild-type cells (Figure 2A). By contrast, nearly all CPY was secreted into the medium (E) in vti1-2 cells at 33 °C. Overexpression of TVP23 from either the 8.2 kb 2μ suppressor plasmid or as the only gene on a 2μ plasmid did not improve sorting of CPY (Figure 2A). Deletion of TVP23 did not result in a CPY transport defect (Figure 2B). To investigate whether lack of Tvp23p caused a synthetic defect, CPY transport was analysed at a semi-permissive temperature of 28 °C. At this temperature, only a partial CPY sorting defect was observed in vti1-2 cells (Figure 2B). This defect would be aggravated in the case of a synthetic genetic interaction. However, sorting of CPY was comparable in vti1-2 cells and in vti1-2 tvp23Δ cells. These results indicate that Tvp23p does not function in CPY transport from the late Golgi via the late endosome to the vacuole.

Lack of genetic interaction between vti1-2 and TVP23 in transport of the vacuolar hydrolase CPY from the late Golgi via the late endosome to the vacuole

Figure 2
Lack of genetic interaction between vti1-2 and TVP23 in transport of the vacuolar hydrolase CPY from the late Golgi via the late endosome to the vacuole

(A) vti1-2 cells were defective in CPY transport via the late endosome to the vacuole at 33 °C, as indicated by the lack of mCPY in cellular extracts (I). Overexpression of TVP23 either from a 8.2 kb library 2μ plasmid or as the only open reading frame on a 2μ plasmid did not suppress the CPY transport defect. (B) CPY transport was analysed at the semi-permissive temperature of 28 °C. CPY transport was unaffected in tvp23Δ cells. The partial CPY sorting defect in vti1-2 cells was not aggravated in vti1-2 tvp23Δ double-mutant cells. CPY traffic was analysed by pulse–chase labelling, immunoprecipitation from cellular extracts (I) and the medium (E), SDS/PAGE and autoradiography. All strains were derived from an SEY6210 background. Abbreviations: WT, wild-type cells; p2CPY, Golgi pro CPY; mCPY, mature vacuolar CPY.

Figure 2
Lack of genetic interaction between vti1-2 and TVP23 in transport of the vacuolar hydrolase CPY from the late Golgi via the late endosome to the vacuole

(A) vti1-2 cells were defective in CPY transport via the late endosome to the vacuole at 33 °C, as indicated by the lack of mCPY in cellular extracts (I). Overexpression of TVP23 either from a 8.2 kb library 2μ plasmid or as the only open reading frame on a 2μ plasmid did not suppress the CPY transport defect. (B) CPY transport was analysed at the semi-permissive temperature of 28 °C. CPY transport was unaffected in tvp23Δ cells. The partial CPY sorting defect in vti1-2 cells was not aggravated in vti1-2 tvp23Δ double-mutant cells. CPY traffic was analysed by pulse–chase labelling, immunoprecipitation from cellular extracts (I) and the medium (E), SDS/PAGE and autoradiography. All strains were derived from an SEY6210 background. Abbreviations: WT, wild-type cells; p2CPY, Golgi pro CPY; mCPY, mature vacuolar CPY.

The vacuolar membrane protein ALP and CPY reach the vacuole via different transport routes. ALP is transported from the late Golgi to the vacuole without passage through the late endosome [5]. Arrival in the vacuole results in cleavage of the pro-form pALP to two different forms of mALP with lower molecular masses. Newly synthesized protein was labelled with [35S]cysteine and [35S]methionine, and ALP was immunoprecipitated. Overexpression of TVP23 from either 2μ plasmid did not increase the amount of mALP in vti1-2 cells at 36 °C (Figure 3A). Only mALP and no pALP were detected in tvp23Δ cells, indicating that ALP transport was not affected by the absence of Tvp23p (Figure 3B). The ratio between pALP and mALP was not different in vti1-2 cells and in vti1-2 tvp23Δ cells at a semi-permissive temperature of 30 °C (Figure 3B). The lack of a multicopy suppression of as well as a synthetic defect with the vti1-2 mutation indicates that Tvp23p is not involved in ALP transport from the late Golgi directly to the vacuole.

Lack of genetic interaction between vti1-2 and TVP23 in transport of the vacuolar phosphatase ALP from the late Golgi directly to the vacuole

Figure 3
Lack of genetic interaction between vti1-2 and TVP23 in transport of the vacuolar phosphatase ALP from the late Golgi directly to the vacuole

(A) Similar amounts of pALP indicated that overexpression of TVP23 either from a 8.2 kb library 2μ plasmid or as the only open reading frame on a 2μ plasmid did not suppress the ALP transport defect in vti1-2 cells at 36 °C. (B) ALP transport was analysed at the semi-permissive temperature of 30 °C. ALP reached the vacuole in tvp23Δ cells. The partial ALP transport defect in vti1-2 cells was not aggravated in vti1-2 tvp23Δ double-mutant cells. ALP traffic was analysed by pulse–chase labelling, immunoprecipitation, SDS/PAGE and autoradiography. All strains were derived from an SEY6210 background. Abbreviations: WT, wild-type cells; pALP, proALP; mALP, mature vacuolar ALP (two bands).

Figure 3
Lack of genetic interaction between vti1-2 and TVP23 in transport of the vacuolar phosphatase ALP from the late Golgi directly to the vacuole

(A) Similar amounts of pALP indicated that overexpression of TVP23 either from a 8.2 kb library 2μ plasmid or as the only open reading frame on a 2μ plasmid did not suppress the ALP transport defect in vti1-2 cells at 36 °C. (B) ALP transport was analysed at the semi-permissive temperature of 30 °C. ALP reached the vacuole in tvp23Δ cells. The partial ALP transport defect in vti1-2 cells was not aggravated in vti1-2 tvp23Δ double-mutant cells. ALP traffic was analysed by pulse–chase labelling, immunoprecipitation, SDS/PAGE and autoradiography. All strains were derived from an SEY6210 background. Abbreviations: WT, wild-type cells; pALP, proALP; mALP, mature vacuolar ALP (two bands).

tvp23Δ and vti1-2 cells were defective in transport from the early endosome to the late Golgi

As we did not find any indication of a role for Tvp23p in anterograde transport from the late Golgi to the vacuole, we analysed retrograde traffic from the early endosome to the late Golgi. GFP–Snc1p serves as a marker for this transport pathway [17]. Snc1p is an R-SNARE on secretory vesicles targeted from the late Golgi to the plasma membrane [29]. After fusion it is incorporated into the plasma membrane. Snc1p is endocytosed and recycled from early endosomes to the late Golgi for further sorting into secretory vesicles. GFP–Snc1p is found in the plasma membrane in most wild-type cells, but there is also an intracellular pool (Figure 4A). GFP–Snc1p accumulates in intracellular structures in cells defective in transport from the early endosome to the late Golgi such as tlg2Δ cells [17], which lack the SNARE Tlg2p required for this transport step (Figure 4B). Slightly fewer tvp23Δ cells than wild-type cells accumulated GFP–Snc1p at the plasma membrane (Figure 4C). GFP–Snc1p localized to the plasma membrane in 68.4% of 790 tvp23Δ cells grown at 24 °C analysed in four independent experiments (S.D.=3.4 between experiments) compared with 74.1% of wild-type cells (S.D.=1.3, 730 cells, three experiments). The difference was small but reproducible and statistically significant (P=0.043, Student's t test). These results suggest that recycling of GFP–Snc1p may be slightly affected in the absence of Tvp23p.

Slightly impaired recycling of GFP–Snc1p to the late Golgi in tvp23Δ cells

Figure 4
Slightly impaired recycling of GFP–Snc1p to the late Golgi in tvp23Δ cells

Cells expressing GFP–Snc1p were analysed by fluorescence microscopy. (A) GFP–Snc1p was localized predominantly to the plasma membrane (PM) in SEY6210 wild-type cells (WT). GFP–Snc1p was found in internal structures in tlg2Δ cells (BY4742 background) known for their block in transport from the early endosome to the late Golgi. (B) GFP–Snc1p was found on the PM and in internal structures in tvp23Δ cells (ISY5, 6210 background, C). (D) Quantification of cells with a GFP–Snc1p PM localization: 68.4% of tvp23Δ cells had PM staining (795 cells in four independent experiments, S.D.=3.4 between experiments) compared with 74.1% of wild-type cells (738 cells in three experiments, S.D. 1.3). The differences are statistically significant (P=0.043, Student's t test).

Figure 4
Slightly impaired recycling of GFP–Snc1p to the late Golgi in tvp23Δ cells

Cells expressing GFP–Snc1p were analysed by fluorescence microscopy. (A) GFP–Snc1p was localized predominantly to the plasma membrane (PM) in SEY6210 wild-type cells (WT). GFP–Snc1p was found in internal structures in tlg2Δ cells (BY4742 background) known for their block in transport from the early endosome to the late Golgi. (B) GFP–Snc1p was found on the PM and in internal structures in tvp23Δ cells (ISY5, 6210 background, C). (D) Quantification of cells with a GFP–Snc1p PM localization: 68.4% of tvp23Δ cells had PM staining (795 cells in four independent experiments, S.D.=3.4 between experiments) compared with 74.1% of wild-type cells (738 cells in three experiments, S.D. 1.3). The differences are statistically significant (P=0.043, Student's t test).

Transport from early endosomes to the late Golgi has not been investigated in vti1 mutant cells. vti1-1 cells display the mildest phenotype and are not temperature-sensitive for growth. vti1-1 cells are defective in transport from the late Golgi to the late endosome at 37 °C. vti1-1, vti1-2 and vti1-11 cells were grown at 24 °C and shifted to 30 or 37 °C for 1 h. GFP–Snc1p was concentrated at the plasma membrane in vti1-1 cells at all temperatures, indicating that recycling of GFP–Snc1p was not defective (Figure 5). vti1-11 cells displayed a similar distribution of GFP–Snc1p to wild-type cells. Intracellular GFP–Snc1p was observed in some vti1-11 cells at 37 °C. However, increased temperature also resulted in intracellular GFP–Snc1p in wild-type cells. This indicates that GFP–Snc1p recycling was normal or only slightly affected in vti1-11 cells. By contrast, GFP–Snc1p distribution was abnormal in vti1-2 cells at all temperatures. Even at 24 °C GFP–Snc1p was shifted from the plasma membrane to internal stores in 50% of the vti1-2 cells. The severity of the phenotype was increased at elevated temperatures (Figure 5). At 30 °C only 27% of the vti1-2 cells displayed GFP–Snc1p at the plasma membrane, 25% of over 200 cells counted in this set of experiments at 37 °C. These results indicate that transport from the early endosome to the late Golgi was somewhat affected even at 24 °C in vti1-2 cells and that the transport block was stronger at 30 and 37 °C.

Defective recycling of GFP–Snc1p to the late Golgi in vti1-2 cells

Figure 5
Defective recycling of GFP–Snc1p to the late Golgi in vti1-2 cells

Cells expressing GFP–Snc1p were grown at 24 °C, shifted to the indicated temperature for 1 h and analysed by fluorescence microscopy. Most GFP–Snc1p was localized to the plasma membrane at all temperatures in vti1-1 (FvMY7) and vti1-11 (FvMY21) cells. By contrast, some GFP–Snc1p was localized to internal structures at 24 °C in vti1-2 cells (FvMY24). The mislocalization was more pronounced at 30 and 37 °C, indicating that recycling to the late Golgi was blocked in vti1-2 cells.

Figure 5
Defective recycling of GFP–Snc1p to the late Golgi in vti1-2 cells

Cells expressing GFP–Snc1p were grown at 24 °C, shifted to the indicated temperature for 1 h and analysed by fluorescence microscopy. Most GFP–Snc1p was localized to the plasma membrane at all temperatures in vti1-1 (FvMY7) and vti1-11 (FvMY21) cells. By contrast, some GFP–Snc1p was localized to internal structures at 24 °C in vti1-2 cells (FvMY24). The mislocalization was more pronounced at 30 and 37 °C, indicating that recycling to the late Golgi was blocked in vti1-2 cells.

vti1-2 cells and vti1-2 tvp23Δ cells were examined for synthetic defects in GFP–Snc1p recycling at 24 °C. GFP–Snc1p was localized to the plasma membrane of fewer vti1-2 tvp23Δ cells compared with vti1-2 cells (Figures 6B and 6C). Some 35% of over 900 vti1-2 tvp23Δ cells analysed in three independent experiments (S.D.=1.2 between experiments) displayed a GFP–Snc1p localization at the plasma membrane compared with 46% of the vti1-2 cells (S.D.=2.0, 720 cells in three experiments) at 24 °C (Figure 6D). These differences were statistically significant (P=0.0012, Student's t test). The synthetic defect in GFP–Snc1p recycling in vti1-2 tvp23Δ cells provides further evidence for a role of Tvp23p in retrograde traffic from the early endosome to the late Golgi.

Synthetic defect in recycling of GFP–Snc1p to the late Golgi in vti1-2 tvp23Δ cells

Figure 6
Synthetic defect in recycling of GFP–Snc1p to the late Golgi in vti1-2 tvp23Δ cells

tvp23Δ (ISY5, A), vti1-2 (FvMY24, B) and vti1-2 tvp23Δ cells (ISY6, C) expressing GFP–Snc1p were grown at 24 °C and analysed by fluorescence microscopy. (D) Quantification of cells with a GFP–Snc1p plasma membrane (PM) localization: 35% of vti1-2 tvp23Δ cells had PM staining (909 cells in three independent experiments, S.D.=1.2 between experiments) compared with 46% of vti1-2 cells (726 cells in three experiments, S.D.=2.0). The differences are statistically significant (P=0.0012, Student's t test).

Figure 6
Synthetic defect in recycling of GFP–Snc1p to the late Golgi in vti1-2 tvp23Δ cells

tvp23Δ (ISY5, A), vti1-2 (FvMY24, B) and vti1-2 tvp23Δ cells (ISY6, C) expressing GFP–Snc1p were grown at 24 °C and analysed by fluorescence microscopy. (D) Quantification of cells with a GFP–Snc1p plasma membrane (PM) localization: 35% of vti1-2 tvp23Δ cells had PM staining (909 cells in three independent experiments, S.D.=1.2 between experiments) compared with 46% of vti1-2 cells (726 cells in three experiments, S.D.=2.0). The differences are statistically significant (P=0.0012, Student's t test).

It has been reported that GFP–Snc1p accumulates in early endosomes and transport vesicles mediating traffic between early endosomes and late Golgi upon block of recycling to the late Golgi [17]. These structures also contain late Golgi proteins because late Golgi proteins can escape to early endosomes and recycle from there to the late Golgi. In order to investigate whether GFP–Snc1p was localized to early endosomes the cells were allowed to endocytose the membrane dye FM4-64 for 5 min. As expected, GFP–Snc1p and FM4-64 co-localized in tlg2Δ cells (Figure 7A). A similar co-localization was observed in vti1-2 as well as in tvp23Δ cells, indicating that GFP–Snc1p accumulated in endocytic structures. DsRed–Sec7p was expressed as a late Golgi enzyme [23]. Considerable co-localization was seen between DsRed–Sec7p and GFP–Snc1p in tlg2Δ cells (Figure 7B). This could be due to a mislocalization of DsRed–Sec7p to early endosomes, because recycling from early endosome to late Golgi is blocked. On the other hand, GFP–Snc1p-containing transport vesicles could be tethered to the late Golgi. DsRed–Sec7p and GFP–Snc1p showed comparable co-localization in tlg2Δ, vti1-2 and tvp23Δ cells selected for internal GFP–Snc1p accumulation. Taken together these results demonstrate that GFP–Snc1p accumulated in similar internal structures in tvp23Δ, vti1-2 and tlg2Δ cells. We conclude that tvp23Δ as well as vti1-2 cells were defective in transport from early endosomes to the late Golgi.

GFP–Snc1p co-localized with endosomal and late Golgi markers in tvp23Δ and vti1-2 cells

Figure 7
GFP–Snc1p co-localized with endosomal and late Golgi markers in tvp23Δ and vti1-2 cells

(A) tlg2Δ (BY4742 background), tvp23Δ (ISY5) and vti1-2 cells (FvMY24) expressing GFP–Snc1p were analysed by fluorescence microscopy after incubation with the endocytosis marker FM4-64 for 5 min. Substantial overlap between intracellular GFP–Snc1p and FM4-64 was observed in these strains. (B) tlg2Δ, tvp23Δ and vti1-2 cells expressing GFP–Snc1p and DsRed–Sec7p were analysed by fluorescence microscopy. Sec7p localizes to the late Golgi and to early endosomes on block of recycling to the late Golgi. Many GFP–Snc1p-positive intracellular structures contained DsRed–Sec7p in tlg2Δ, tvp23Δ and vti1-2 cells. A slight gamma-correction was applied for clarity of the merged images.

Figure 7
GFP–Snc1p co-localized with endosomal and late Golgi markers in tvp23Δ and vti1-2 cells

(A) tlg2Δ (BY4742 background), tvp23Δ (ISY5) and vti1-2 cells (FvMY24) expressing GFP–Snc1p were analysed by fluorescence microscopy after incubation with the endocytosis marker FM4-64 for 5 min. Substantial overlap between intracellular GFP–Snc1p and FM4-64 was observed in these strains. (B) tlg2Δ, tvp23Δ and vti1-2 cells expressing GFP–Snc1p and DsRed–Sec7p were analysed by fluorescence microscopy. Sec7p localizes to the late Golgi and to early endosomes on block of recycling to the late Golgi. Many GFP–Snc1p-positive intracellular structures contained DsRed–Sec7p in tlg2Δ, tvp23Δ and vti1-2 cells. A slight gamma-correction was applied for clarity of the merged images.

Certain late Golgi proteins are retrieved from the late endosome back to the late Golgi. An example is DPAP A (diaminopeptidase A), which binds via the cytosolic domain to the retromer complex and to sorting nexins [27,30]. This transport pathway can be investigated using the fusion protein A-ALP, which consists of the cytosolic tail of DPAP A, required for trafficking, fused to the transmembrane and luminal part of ALP, used for immunoprecipitation with antiserum directed against ALP [22]. A-ALP recycles between the late Golgi, early endosomes and late endosomes. If retrograde traffic from the late endosome is blocked, unprocessed A-ALP (pA-ALP) will be transported to the vacuole and cleaved to a smaller form mA-ALP. Therefore loss of pA-ALP and generation of mA-ALP can serve as an assay for block in retrograde transport from the late endosome to the late Golgi. Wild-type, tvp23Δ and vps5Δ cells lacking the retromer subunit Vps5p [27] were incubated with [35S]cysteine and [35S]methionine, chased with unlabelled cysteine and methionine for different time points, and ALP was immunoprecipitated. In vps5Δ cells, approx. 50% of A-ALP was found as mA-ALP after a 60 min chase, and processing was more prominent after a 120 min chase (Figure 8) as described in [27]. The amounts of pA-ALP remaining were similar in wild-type and tvp23Δ cells chased for 60 and 120 min, whereas a band with a mobility similar to that of mA-ALP did not increase with time (Figure 8). This indicates that retrograde transport from the late endosome to the late Golgi was not significantly affected in tvp23Δ cells.

Transport from the late endosome to the late Golgi was not affected in tvp23Δ cells

Figure 8
Transport from the late endosome to the late Golgi was not affected in tvp23Δ cells

Retrograde transport from the late endosome to the late Golgi was monitored using A-ALP, which contained a retrieval signal for recycling from the late endosome to the late Golgi. Block in retrograde traffic from the late endosome, for example in vps5Δ cells, results in a destabilization of pA-ALP and processing to the slightly smaller mA-ALP, because it is transported to the vacuole. A-ALP consists of the cytosolic domain of DPAP A, which localizes to the late Golgi, fused to the transmembrane and luminal domain of ALP for detection. Cells were pulse-labelled at 30 °C and chased for indicated time periods before immunoprecipitation. A-ALP was immunoprecipitated with antisera directed against ALP in cells deleted for PHO8, the gene encoding ALP, to avoid immunoprecipitation of ALP. The stability of pA-ALP was comparable in wild-type (WT, SNY18) and tvp23Δ cells (AGY4), while a band with a mobility similar to mA-ALP did not increase with time. By contrast, the amounts of pA-ALP and mA-ALP were comparable after a 60 min chase, and mA-ALP was the predominant form after 120 min in vps5Δ cells (AHY41). Abbreviations: pA-ALP, Golgi pro A-ALP; mA-ALP, vacuolar mature A-ALP.

Figure 8
Transport from the late endosome to the late Golgi was not affected in tvp23Δ cells

Retrograde transport from the late endosome to the late Golgi was monitored using A-ALP, which contained a retrieval signal for recycling from the late endosome to the late Golgi. Block in retrograde traffic from the late endosome, for example in vps5Δ cells, results in a destabilization of pA-ALP and processing to the slightly smaller mA-ALP, because it is transported to the vacuole. A-ALP consists of the cytosolic domain of DPAP A, which localizes to the late Golgi, fused to the transmembrane and luminal domain of ALP for detection. Cells were pulse-labelled at 30 °C and chased for indicated time periods before immunoprecipitation. A-ALP was immunoprecipitated with antisera directed against ALP in cells deleted for PHO8, the gene encoding ALP, to avoid immunoprecipitation of ALP. The stability of pA-ALP was comparable in wild-type (WT, SNY18) and tvp23Δ cells (AGY4), while a band with a mobility similar to mA-ALP did not increase with time. By contrast, the amounts of pA-ALP and mA-ALP were comparable after a 60 min chase, and mA-ALP was the predominant form after 120 min in vps5Δ cells (AHY41). Abbreviations: pA-ALP, Golgi pro A-ALP; mA-ALP, vacuolar mature A-ALP.

DISCUSSION

Here, we studied different vti1 mutant cells for defects in retrograde transport from the early endosome to the late Golgi. A marker protein recycling between endosomes and late Golgi was not useful because anterograde transport from the Golgi to endosomes is blocked in vti1 mutant cells. Therefore we had to use a marker protein, which is endocytosed from the plasma membrane and recycled to the late Golgi, such as GFP–Snc1p [17]. GFP–Snc1p trafficking was not affected in vti1-1 cells and vti1-11 cells. By contrast, vti1-2 cells were defective in GFP–Snc1p recycling. Each of the three mutant alleles contains two amino acid exchanges in different positions in the N-terminal half of the SNARE motif or just before the SNARE motif [19]. The amino acid exchanges in the three vti1 mutant proteins have to affect different SNARE complex partners in different ways in order to cause defects in various numbers and combinations of transport pathways (Figure 9). The crystal structure of the mammalian late endosomal SNARE complex provides some hints due to the conserved structure [31]. One of the mammalian homologues of the Qb-SNARE Vti1p, vti1b is part of this complex together with syntaxin 7, syntaxin 8 and VAMP-8 (vesicle-associated membrane protein-8). Aspartic acid residue (Asp157) of vti1b forms a salt bridge with arginine residue (Arg164) of syntaxin 8. The equivalent amino acid residue in Vti1p is Glu145, which is changed to E145K in vti1-1p and to E145G in vti1-11p. The position of the Qc-SNARE syntaxin 8 in the SNARE complex is taken by Syn8p in transport to the late endosome, by Vam7p in transport to the vacuole and by Tlg1p in retrograde transport to the late Golgi. The salt bridge is only conserved in Syn8p with arginine residue (Arg183) in the position equivalent to Arg164 of syntaxin 8, but filled by Asn151 in Tlg1p and Ala269 in Vam7p. Therefore it is not surprising that vti1-1 and vti1-11 cells are defective in the Syn8p- but not in the Tlg1p-dependent transport step. The second amino acid exchange in both alleles is found in amino acid residues, which form layers of interacting side chains in the hydrophobic core of the helical bundle. vti1-1 cells with the G148R exchange in the –3 layer in vti1-1p are only defective in transport to the late endosome. By contrast, the L155F exchange in the –1 layer of vti1-11p has more severe consequences, because vti1-11 cells are defective in transport to the vacuole and in retrograde transport to the cis-Golgi in addition to a block in transport to the late endosome [6,7]. vti1-2p contains the amino acid exchanges S130P just four amino acid residues N-terminal of the SNARE motif (–7 layer Leu134) and I151T in the –2 layer. vti1-2 cells accumulated some intracellular GFP–Snc1p already at 24 °C and a clear defect was observed at 30 °C. Interestingly, transport of CPY to the late endosome was normal at 24 °C and defective at 31 °C in vti1-2 cells. The restrictive temperature for ALP transport to the vacuole was 37 °C. These results indicate that the SNARE partners of Vti1p in these three complexes react slightly differently to the same amino acid exchange.

Amino acid exchanges in the different vti1 mutant alleles result in different effects in membrane transport

Figure 9
Amino acid exchanges in the different vti1 mutant alleles result in different effects in membrane transport

(A) Vti1p consists of an N-terminal domain, which forms a three-helix bundle (grey rods), the SNARE motif (amino acid residues 134–186) and a C-terminal transmembrane domain (black rod). Boxes indicate the amino acid residues which participate in layers of interacting side chains within the centre of the SNARE complex. Amino acid exchanges are indicated: vti1-1 E145K, G148R (–3 layer); vti1-2 S130P, I151T (–2 layer); vti1-11 E145G, L155F (–1 layer). (B) Names of vti1 mutants next to arrows indicate transport steps defective in this allele. Defect in vti1-1 Golgi to LE; vti1-2 Golgi to LE, fusion with the vacuole, EE to Golgi; vti1-11 Golgi to LE, fusion with the vacuole, retrograde to cis-Golgi. Four different SNARE complexes are required for these trafficking steps (boxes). Tvp23p is involved in EE to Golgi transport. Abbreviations: LE, late endosome; EE, early endosome.

Figure 9
Amino acid exchanges in the different vti1 mutant alleles result in different effects in membrane transport

(A) Vti1p consists of an N-terminal domain, which forms a three-helix bundle (grey rods), the SNARE motif (amino acid residues 134–186) and a C-terminal transmembrane domain (black rod). Boxes indicate the amino acid residues which participate in layers of interacting side chains within the centre of the SNARE complex. Amino acid exchanges are indicated: vti1-1 E145K, G148R (–3 layer); vti1-2 S130P, I151T (–2 layer); vti1-11 E145G, L155F (–1 layer). (B) Names of vti1 mutants next to arrows indicate transport steps defective in this allele. Defect in vti1-1 Golgi to LE; vti1-2 Golgi to LE, fusion with the vacuole, EE to Golgi; vti1-11 Golgi to LE, fusion with the vacuole, retrograde to cis-Golgi. Four different SNARE complexes are required for these trafficking steps (boxes). Tvp23p is involved in EE to Golgi transport. Abbreviations: LE, late endosome; EE, early endosome.

The growth defect of vti1-2 cells at elevated temperatures was aggravated in the absence of Tvp23p and suppressed on overproduction of Tvp23p. By contrast, overexpression of TVP23 did not suppress the growth defect of vti1-11 cells at elevated temperatures. These results demonstrate an allele-specific genetic interaction of TVP23 with VTI1 and suggest that Tvp23p functions in one or more of the pathways, which are defective in vti1-2 cells. Absence of Tvp23p did not result in defects of transport of CPY from the Golgi via the late endosome to the vacuole or in transport of ALP directly to the vacuole, as observed in the present study and by Inadome et al. [32]. This does not exclude a role of Tvp23p in these transport pathways, because a redundant protein could substitute for a lack of Tvp23p. As we observed a synthetic growth defect in vti1-2 tvp23Δ cells, we should be able to identify a synthetic defect in one of the transport pathways, which require Vti1p. However, we did not observe synthetic defects in CPY or ALP transport. The problem of redundant proteins can be circumvented if overexpression results in the suppression of a phenotype. However, overexpression of TVP23 did not result in a suppression of the CPY or ALP sorting defects in vti1-2 cells, even though the growth defect was suppressed. Therefore we conclude that Tvp23p does not function in transport from the late Golgi to the late endosome or from the late Golgi directly to the vacuole. The lack of Tvp23p did not result in increased vacuolar degradation of the late Golgi protein A-ALP, indicating that retrograde transport from the late endosome was not defective. Slightly fewer cells accumulated GFP–Snc1p at the plasma membrane in the absence of Tvp23p compared with wild-type cells. In addition, the absence of Tvp23p increased the GFP–Snc1p recycling defect in vti1-2 cells. These results indicate that Tvp23p has a role in retrograde transport from the early endosome to the late Golgi. However, the block in transport was weak, suggesting redundancy in the pathway or a modifying role.

Our results show the first transport defects due to the lack of Tvp23p and are in agreement with earlier observations. Tvp23p was identified as a membrane protein of vesicles which were immunoisolated with antibodies directed against the late Golgi and early endosome SNARE Tlg2p [28]. Tagged Tvp23p co-localizes with proteins of the late Golgi and early endosome [28]. Therefore Tvp23p is found in the right subcellular localization to function in retrograde transport from the early endosome to the late Golgi. A synthetic growth defect was observed in tvp23Δ ypt6Δ cells [28]. As the Rab/Ypt GTPase Ypt6p functions in retrograde transport from the early endosome to the late Golgi [33], this is consistent with a function of Tvp23p in the same pathway. However, a synthetic defect between two deletion mutants could also indicate a function in two parallel pathways. Co-immunoprecipitation revealed that Tvp23p binds to Yip4p and Yip5p [32]. Yip4p and Yip5p are Rab GTPase interacting factors [34]. Yip4p is able to bind all Rab GTPases found in yeast. Yip5p interacts with all yeast Rab GTPases except for Ypt6p. As Yip4p and Yip5p can bind each other, the complex may function on Ypt6p. Yip4p and Yip5p belong to the YIP1 family of proteins. Yip3, a mammalian member of this protein family, recruits Rab9 to the membrane by dissociating the soluble GDI (guanine-nucleotide-dissociation inhibitor)–Rab complex [35] and is therefore called a GDF (GDI displacement factor). It is possible that Yip4p and Yip5p serve as a GDF for Ypt6p. Tvp23p may play a role in this process. Recruitment of Ypt6p-GDP to the membrane increases the amount of substrate available for activation by the Ric1p–Rgp1p complex, which acts as GEF for Ypt6p [36]. Overproduction of Tvp23p may increase the efficiency of the tethering between Ypt6p and the GARP–VFT tethering complex. As GARP–VFT binds directly to the SNARE Tlg1p [13], this may allow for an efficient use of the residual activity of the Ykt6p, Tlg2p, Vti1p and Tlg1p SNARE complex in vti1-2 cells. An alternative explanation would be a direct interaction between Tvp23p and Vti1p.

We thank Claudia Prange and Christiane Wiegand for excellent technical assistance. Many thanks to Dr T.H. Stevens (Institute of Molecular Biology, University of Oregon, Eugene, OR, U.S.A.), Dr H.R.B. Pelham (MRC Laboratory of Molecular Biology, Cambridge, U.K.), Dr S.F. Nothwehr (Division of Biological Sciences, University of Missouri, Columbia, MO, U.S.A.) and T.J. Proszynski (Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany) for the gift of antiserum, yeast strains and plasmids.

Abbreviations

     
  • A-ALP

    cytosolic domain of dipeptidyl aminopeptidase A fused to the transmembrane and luminal domain of alkaline phosphatase

  •  
  • ALP

    alkaline phosphatase

  •  
  • CPY

    carboxypeptidase Y

  •  
  • DPAP A

    diaminopeptidase A

  •  
  • GDI

    guanine-nucleotide-dissociation inhibitor

  •  
  • GDF

    GDI displacement factor

  •  
  • GEF

    guanine-nucleotide-exchange factor

  •  
  • GFP

    green fluorescent protein

  •  
  • SD minimal medium

    standard minimal medium

  •  
  • SNARE

    soluble N-ethylmaleimide-sensitive fusion protein-attachment protein receptor

FUNDING

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

References

References
1
Cai
H. Q.
Reinisch
K.
Ferro-Novick
S.
Coats, tethers, Rabs, and SNAREs work together to mediate the intracellular destination of a transport vesicle
Dev. Cell
2007
, vol. 
12
 (pg. 
671
-
682
)
2
Whyte
J.
Munro
S.
Vesicle tethering complexes in membrane traffic
J. Cell Sci.
2002
, vol. 
115
 (pg. 
2627
-
2637
)
3
Jahn
R.
Scheller
R. H.
SNAREs – engines for membrane fusion
Nat. Rev. Mol. Cell Biol.
2006
, vol. 
7
 (pg. 
631
-
643
)
4
Fasshauer
D.
Sutton
R. B.
Brunger
A. T.
Jahn
R.
Conserved structural features of the synaptic fusion complex: SNARE proteins reclassified as Q- and R-SNAREs
Proc. Natl. Acad. Sci. U.S.A.
1998
, vol. 
95
 (pg. 
15781
-
15786
)
5
Bowers
K.
Stevens
T. H.
Protein transport from the late Golgi to the vacuole in the yeast Saccharomyces cerevisiae
Biochim. Biophys. Acta
2005
, vol. 
1744
 (pg. 
438
-
454
)
6
Fischer von Mollard
G.
Nothwehr
S. F.
Stevens
T. H.
The yeast v-SNARE Vti1p mediates two vesicle transport pathways through interactions with the t-SNAREs Sed5p and Pep12p
J. Cell Biol.
1997
, vol. 
137
 (pg. 
1511
-
1524
)
7
Fischer von Mollard
G.
Stevens
T. H.
The Saccharomyces cerevisiae v-SNARE Vti1p is required for multiple membrane transport pathways to the vacuole
Mol. Biol. Cell
1999
, vol. 
10
 (pg. 
1719
-
1732
)
8
Lewis
M. J.
Pelham
H. R. B.
A new yeast endosomal SNARE related to mammalian syntaxin 8
Traffic
2002
, vol. 
3
 (pg. 
922
-
929
)
9
Kweon
Y.
Rothe
A.
Conibear
E.
Stevens
T. H.
Ykt6p is a multifunctional yeast R-SNARE that is required for multiple membrane transport pathways to the vacuole
Mol. Biol. Cell
2003
, vol. 
14
 (pg. 
1868
-
1881
)
10
Peplowska
K.
Markgraf
D. F.
Ostrowicz
C. W.
Bange
G.
Ungermann
C.
The CORVET tethering complex interacts with the yeast Rab5 homolog Vps21 and is involved in endo-lysosomal biogenesis
Dev. Cell
2007
, vol. 
12
 (pg. 
739
-
750
)
11
Wickner
W.
Yeast vacuoles and membrane fusion pathways
EMBO J.
2002
, vol. 
21
 (pg. 
1241
-
1247
)
12
Conibear
E.
Stevens
T.
Vps52p, Vps53p, and Vps54p form a novel multisubunit complex required for protein sorting at the yeast late Golgi
Mol. Biol. Cell
2000
, vol. 
11
 (pg. 
305
-
323
)
13
Siniossoglou
S.
Pelham
H. R. B.
An effector of Ypt6p binds the SNARE Tlg1p and mediates selective fusion of vesicles with late Golgi membranes
EMBO J.
2001
, vol. 
20
 (pg. 
5991
-
5998
)
14
Holthuis
J. C. M.
Nichols
B. J.
Dhruvakumar
S.
Pelham
H. R. B.
Two syntaxin homologues in the TGN/endosomal system in yeast
EMBO J.
1998
, vol. 
17
 (pg. 
113
-
126
)
15
Brickner
J.
Blanchette
J.
Sipos
G.
Fuller
R.
The Tlg SNARE complex is required for TGN homotypic fusion
J. Cell Biol.
2001
, vol. 
155
 (pg. 
969
-
978
)
16
Brennwald
P.
Kearns
B.
Champion
K.
Keranen
S.
Bankaitis
V.
Novick
P.
Sec9 is a SNAP-25-like component of a yeast SNARE complex that may be the effector of Sec4 function in exocytosis
Cell
1994
, vol. 
79
 (pg. 
245
-
258
)
17
Lewis
M. J.
Nichols
B. J.
Prescianotto-Baschong
C.
Riezman
H.
Pelham
H. R. B.
Specific retrieval of the exocytic SNARE Snc1p from early yeast endosomes
Mol. Biol. Cell
2000
, vol. 
11
 (pg. 
23
-
38
)
18
Lupashin
V. V.
Pokrovskaya
I. D.
McNew
J. A.
Waters
M. G.
Characterization of a novel yeast SNARE protein implicated in Golgi retrograde traffic
Mol. Biol. Cell
1997
, vol. 
8
 (pg. 
2659
-
2676
)
19
Fischer von Mollard
G.
Stevens
T. H.
A human homolog can functionally replace the yeast v-SNARE Vti1p in two vesicle transport pathways
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
2624
-
2630
)
20
Winzeler
E.
Shoemaker
D. D.
Astromoff
A.
Liang
H.
Anderson
K.
Andre
B.
Bangham
R.
Benito
R.
Boeke
J. D.
Bussey
H.
, et al. 
Functional characterization of the S. cerevisiae genome by gene deletion and parallel analysis
Science
1999
, vol. 
285
 (pg. 
901
-
906
)
21
Nothwehr
S. F.
Conibear
E.
Stevens
T. H.
Golgi and vacuolar membrane proteins reach the vacuole in vps1 mutant yeast cells via the plasma membrane
J. Cell Biol.
1995
, vol. 
129
 (pg. 
35
-
46
)
22
Nothwehr
S. F.
Roberts
C. J.
Stevens
T. H.
Membrane protein retention in the yeast Golgi apparatus: dipeptidyl aminopeptidase A is retained by a cytoplasmic signal containing aromatic residues
J. Cell Biol.
1993
, vol. 
121
 (pg. 
1197
-
1209
)
23
Proszynski
T. J.
Klemm
R. W.
Gravert
M.
Hsu
P. P.
Gloor
Y.
Wagner
J.
Kozak
K.
Grabner
H.
Walzer
K.
Bagnat
M.
, et al. 
A genome-wide visual screen reveals a role for sphingolipids and ergosterol in cell surface delivery in yeast
Proc. Natl. Acad. Sci. U.S.A.
2005
, vol. 
102
 (pg. 
17981
-
17986
)
24
Carlson
M.
Botstein
D.
Two differentially regulated mRNAs with different 5′ ends encode secreted proteins with intracellular forms of yeast invertase
Cell
1982
, vol. 
28
 (pg. 
145
-
154
)
25
Dilcher
M.
Köhler
B.
Fischer von Mollard
G.
Genetic interactions with the yeast Q-SNARE VTI1 reveal novel functions for the R-SNARE YKT6
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
34537
-
34544
)
26
Vater
C. A.
Raymond
C. K.
Ekena
K.
Howald
S. I.
Stevens
T. H.
The VPS1 protein, a homolog of dynamin required for vacuolar protein sorting in Saccharomyces cerevisiae, is a GTPase with two functionally separable domains
J. Cell Biol.
1992
, vol. 
119
 (pg. 
773
-
786
)
27
Nothwehr
S. F.
Hindes
A. E.
The yeast VPS5/GRD2 gene encodes a sorting nexin-1-like protein required for localizing membrane proteins to the late Golgi
J. Cell Sci.
1997
, vol. 
110
 (pg. 
1063
-
1072
)
28
Inadome
H.
Noda
Y.
Adachi
H.
Yoda
K.
Immunoisolation of the yeast Golgi subcompartments and characterization of a novel membrane protein, SvP26, discovered in the Sed5-containing compartments
Mol. Cell. Biol.
2005
, vol. 
25
 (pg. 
7696
-
7710
)
29
Protopopov
V.
Govindan
B.
Novick
P.
Gerst
J. E.
Homologs of the synaptobrevin/VAMP family of synaptic vesicle proteins function on the late secretory pathway in S. cerevisiae
Cell
1993
, vol. 
74
 (pg. 
855
-
861
)
30
Nothwehr
S. F.
Bryant
N. J.
Stevens
T. H.
The newly identified yeast GRD genes are required for retention of late-Golgi membrane proteins
Mol. Cell. Biol.
1996
, vol. 
16
 (pg. 
2700
-
2707
)
31
Antonin
W.
Fasshauer
D.
Becker
S.
Jahn
R.
Schneider
T.
Crystal structure of the endosomal SNARE complex reveals common structural principles of all SNAREs
Nat. Struct. Biol.
2002
, vol. 
9
 (pg. 
107
-
111
)
32
Inadome
H.
Noda
Y.
Kamimura
Y.
Adachi
H.
Yoda
K.
Tvp38, Tvp23, Tvp18 and Tvp15: Novel membrane proteins in the Tlg2-containing Golgi/endosome compartments of Saccharomyces cerevisiae
Exp. Cell Res.
2007
, vol. 
313
 (pg. 
688
-
697
)
33
Tsukada
M.
Gallwitz
D.
Isolation and characterization of SYS genes from yeast, multicopy suppressors of the functional loss of the transport GTPase Ypt6p
J. Cell Sci.
1996
, vol. 
109
 (pg. 
2471
-
2481
)
34
Calero
M.
Winand
N. J.
Collins
R. N.
Identification of the novel proteins Yip4p and Yip5p as Rab GTPase interacting factors
FEBS Lett.
2002
, vol. 
515
 (pg. 
89
-
98
)
35
Sivars
U.
Aivazian
D.
Pfeffer
S. R.
Yip3 catalyses the dissociation of endosomal Rab–GDI complexes
Nature
2003
, vol. 
425
 (pg. 
856
-
859
)
36
Siniossoglou
S.
Peak-Chew
S. Y.
Pelham
H. R. B.
Ric1p and Rgp1p form a complex that catalyses nucleotide exchange on Ypt6p
EMBO J.
2000
, vol. 
19
 (pg. 
4885
-
4894
)

Author notes

1

These authors contributed equally to this work.