Involvement of Ypt7p, a small GTPase, in traffic from late endosome to the vacuole in yeast

Small GTPases of the ras superfamily, including members of the ypt family in yeast and the rab family in mammalian cells, have been shown to play a pivotal role in intracellu-lar membrane trafficking (for reviews see Balch, 1990; Goud and McCaffrey, 1991; Gruenberg and Clague, 1992; Pfeffer, 1992). In yeast, the role of small GTPases in trans-port vesicle fusion is exemplified by Ypt1p and Sec4p, two essential components of the secretory pathway. Ypt1p is involved in vesicular transport between the endoplasmic reticulum (ER) and the cis-Golgi (Segev et al., 1988; Segev, 1991; Schmitt et al., 1988; Rexach and Schekman, 1991) whereas Sec4p is required for delivery of Golgi-derived secretory vesicles to the plasma membrane (Goud et al., 1988; Walworth et al., 1989). In both cases, temperature-sensitive mutants accumulate transport vesicles in vivo at the non-permissive temperature (Salminen and Novick, 1987; Becker et al., 1991). Another group of small GTPases is required for the correct formation of transport vesicles from the donor membrane (Nakano and Muramatsu, 1989; Stearns et al., 1990; D’Enfert et al., 1991).

In mammalian cells, several of the proteins belonging to the rab family have been localized to distinct subcellular compartments of both the exocytic and endocytic pathways (Segev et al., 1988; Chavrier et al., 1990; Goud et al., 1990; Plutner et al., 1991; van der Sluijs et al., 1991, Lombardi et al., 1993). rab5, rab7 and rab9 are associated specifically with the endocytic membrane system. Rab5 is localized to both the plasma membrane and early endosomes, whereas rab7 and most of rab9 are found in late endosomes (Chavrier et al., 1990; Lombardi et al., 1993). rab5 has been shown to be involved in fusion of early endosomes in vitro and to function as a regulatory factor influencing the kinet-ics of membrane traffic early in the endocytic pathway (Gorvel et al., 1991; Bucci et al., 1992). rab9, but not rab7, is involved in transport between late endosomes and the trans-Golgi network as shown in an in vitro transport assay (Lombardi et al., 1993). The role of rab7, in contrast, is still unknown in animal cells.

The best characterized ligand for receptor-mediated endocytosis in yeast is the pheromone α-factor, a tride-capeptide that is synthesized and secreted by α-cells. Upon binding of α-factor to its receptor on a cells the pheromone elicits a number of responses leading to the formation of diploid cells (for a review see Cross et al., 1988). The α-factor receptor, a plasma membrane protein of seven trans-membrane domains (Burkholder and Hartwell, 1985; Nakayama et al., 1985; Cartwright and Tipper, 1991) is down-regulated upon pheromone binding (Jenness and Spatrick, 1986), suggesting that α-factor is internalized by a receptor-mediated mechanism. The uptake of α-factor is time-, temperature- and energy-dependent, and degradation in the vacuole requires the PEP4 gene product (Chvatchko et al., 1986; Singer and Riezman, 1990). Upon internalization, α-factor passes through at least two separable mem-brane-bound compartments, early and late endosomes, as defined by kinetic analysis. These two endosomal com-partments have been further purified and characterized, and putative endosomal marker proteins have been isolated (Singer and Riezman, 1990; Singer-Krüger et al., 1993; Singer-Krüger et al., unpublished).

Soluble lysosomal enzymes in mammalian cells are sorted from secreted proteins in the trans-Golgi network by the mannose 6-phosphate receptor and are transported to the lysosome via an endosomal compartment (Griffiths and Simons, 1986; Dahms et al., 1989). Proteins destined for the yeast lysosome-like vacuole are segregated from secreted proteins in a late Golgi compartment (Graham and Emr, 1991; Stevens et al., 1982). Early steps in the sorting of vacuolar proteins have been elucidated by multiple studies on vps mutants (Rothman and Stevens, 1986; Banta et al., 1988; Robinson et al., 1988; summarized by Ray-mond et al., 1992). Subsequent transport steps of both soluble and membrane-associated vacuolar proteins have not been extensively characterized to date. It has been postulated that vacuolar proteins, in analogy with mammalian lysosomal enzymes, are delivered to the vacuole via an endosomal compartment (Wichmann et al., 1992; Raymond et al., 1992). Vida and colleagues (1993) provided evidence for the sorting of vacuolar hydrolases through an endosome-like prevacuolar compartment by showing a partial overlap in the fractionation of the Golgi-modified precursor of car-boxypeptidase Y (p2CPY) and endocytosed α-factor.

YPT7 encodes a yeast homolog of the mammalian rab7 protein. Δypt7 mutant cells exhibit fragmentation of the vacuole, a strong inhibition in the processing of membrane-bound vacuolar alkaline phosphatase and a delay in the maturation of soluble vacuolar hydrolases. α-Factor internalization is normal in the mutant, but degradation of the pheromone is slower (Wichmann et al., 1992). It is not known whether the delay of α-factor degradation is due to a true endocytic block or to a defect in vacuolar protease localization and it is therefore not resolved where in the cell Ypt7p functions.

We show here that endocytosed α-factor accumulates in late endosomes in Δypt7 mutant cells, providing direct evi-dence that the delay in α-factor degradation is not entirely due to the previously described defect in protease matura-tion but to a defect in α-factor transport to the vacuole. Taking advantage of the Δypt7 mutant phenotype, we pro-vide independent evidence supporting the idea that path-ways of vacuole biogenesis and endocytosis intersect in the endosomal membrane system in yeast.

The strains used in these experiments were RH144-3D (Mata, his4, ura3, leu2, bar1-1), RH289-1D (Mata, cI::URA3, ura3, leu2, lys2, pep4-3, bar1-1) as YPT7 strains and RH301-2C (Mata, ypt7::URA3, cI::URA3, his4, ura3, leu2, lys2, bar1-1) and RH301-1B (Mata, ypt7::URA3, cI::URA3, his4, ura3, leu2, lys2, pep4-3, bar1-1) as the corresponding Δypt7 mutants. RH301-2C and RH301-1B were obtained by backcrossing strain RH1 (Mata, ypt7::URA3, his4, ura3, leu2, bar1-1, kindly provided by D. Gall-witz) with RH289-7A (Matα, cI::URA3, his4, ura3, leu2, lys2, pep4-3, bar1-1), which was derived from the same cross as RH289-1D. cI::URA3 denotes the fragment of the human invari-ant chain gene (coding for an 82 amino acid c-myc-tagged miniprotein) that had been integrated at the URA3 locus in the original strains using an integration plasmid. The protein has no effect on endocytic traffic. RH449 (Matα, his4, ura3, leu2, lys2, bar1-1) was used for the production of α-factor. All yeast strains were grown in complete medium YPUAD (1% yeast extract, 2% peptone (both Gibco Ltd., Paisley, Great Britain), 40 μg/ml ade-nine (Sigma Chemical Co., St Louis, MO), 40 μg/ml uracil (E. Merck, Darmstadt, Germany), 2% glucose) to exponential phase (1.5×10⁷ to 2.5×10⁷/ml) at 30°C in a rotary shaker.

Lyticase was prepared as described by Scott and Schekman (1980). [³⁵S]sulfate was from Amersham (Zürich, Switzerland), silica gel 60 plates were from Merck, and solvents for thin-layer chromatography (TLC) were from both Fluka and Merck. Emul-sifier Safe was from Packard Instrument Co. (Groningen, Nether-lands) and Nycodenz was purchased from Nycomed Pharma AS (Oslo, Norway). Ultracentrifugation tubes were from Beckman Instruments Inc. (Palo Alto, CA). Chemicals for SDS-PAGE were obtained from Bio-Rad (Richmond, CA) or BDH (Poole, Great Britain), nitrocellulose filters were from Millipore Continental Water Systems (Bedford, MA) and from Schleicher & Schüll (Dassel, Germany). All other chemicals were either from Merck, Fluka or Sigma Chemical Co. Hexokinase antibody was kindly provided by G. Schatz (Biocenter, Basel, Switzerland). ³⁵S-labeled α-factor was prepared and isolated as described earlier (Dulic and Riezman, 1989; Dulic et al., 1991).

Yeast cell fractionation was performed according to a modified version of Singer and Riezman (1990). For each time course experiment, pep4-3 cells (∼2×10¹⁰ cells per time point) were har-vested from an exponentially growing culture of either RH289-1D or RH301-1B and were resuspended in 5 ml YPUAD medium per gram wet weight. ³⁵S-α-factor was added (∼5×10⁵cpm per time point) and allowed to bind for 1 hour on ice with shaking on a rocker. The cells were harvested at 4°C and α-factor was internalized at 30°C for 2, 7.5, 15, 20, 40, 60 or 90 minutes. Alter-natively, α-factor was internalized at 15°C for 20 minutes with subsequent shift to 32°C for 1, 15 and 30 minutes. At each time point, samples of equal size were removed on ice and diluted with two volumes of ice-cold sorbitol medium (0.6 M sorbitol, 5 mM 2-amino-2-methyl-1,3-propanediol-Pipes, pH 6.8) containing 15 mM NaN₃ and 15 mM NaF. The cells were harvested and resus-pended in 20 ml 0.14 M cysteamine-HCl, 0.6 M sorbitol, 25 mM 2-amino-2-methyl-1,3-propanediol-Pipes, pH 6.8, 5 mM EDTA, 20 mM NaN₃ and 20 mM NaF. After 20 minutes at 30°C with slow shaking, the cells were diluted with one volume of sorbitol medium, harvested and resuspended in 2.8 ml sorbitol medium containing 20 mM NaN₃ and 20 mM NaF. Lyticase (∼1.5×10³ units) was added and cells were converted into spheroplasts for 1 hour at 30°C. Spheroplasts were diluted with 30 ml sorbitol medium and harvested at 500 g at 4°C. Spheroplasts were resus-pended in 8 ml sorbitol medium and were lysed by 6 strokes in a Dounce glass homogenizer with a tightly fitting pestle (Wheaton Industries, Millville, NJ). The lysate was subjected to three dif-ferent centrifugation steps at 4°C. It was first centrifuged for 5 minutes at 3,500 g giving rise to pellet (P1) and supernatant (S1) fractions. S1 was further centrifuged for 5 minutes at 7,500 g resulting in P2 and S2. S2 was centrifuged for 1 hour at 100,000 g separating P3 from S3. Samples of each of the fractions were counted in a liquid scintillation analyzer (Packard Instrument Co.) to determine the amount of ³⁵S-α-factor or were subjected to SDS-PAGE for subsequent Western blot analysis (Towbin et al., 1979) using either anti-hexokinase- or anti-carboxypeptidase Y-antibodies. The P3 fraction was resuspended in 1.4 ml of 37% (w/v) Nycodenz, homogenized 4 times using a 5 ml tissue grinder, was overlayed with a discontinuous Nycodenz gradient and centrifuged as described by Singer and Riezman (1990). A total of 17 fractions of approximately 660 μl were collected from the top of the gradient and 580 μl were analyzed for ³⁵S-α-factor as described.

Correction for cell lysis was performed as described by Singer and Riezman (1990) using hexokinase as a cytosolic marker.

Pheromone degradation assays were performed as described earlier (Dulic et al., 1991) using strains RH144-3D, RH301-2C and RH301-1B. For each degradation time course, cells were harvested at 1×10⁷ to 1.5×10⁷/ml, washed twice with YPUAD, allowed to bind ³⁵S-α-factor (∼5×10⁵ cpm per 5 time points) and to internalize the pheromone for 7.5, 15, 30, 40 or 90 minutes at 30°C. Alterna-tively, uptake was performed for 20 minutes at 15°C with subse-quent shift to 30°C for 1, 15, 30, 40 or 90 minutes. After the indi-cated times, one sample was removed into 20 ml pH1 buffer and processed for TLC as described (Dulic et al., 1991), the second was allowed to continue incubation in the presence of metabolic inhibitors (50 mM NaN₃ and 50 mM NaF) for a total of 120 minutes at 30°C. After the chase, these samples were removed into pH 1 buffer and processed for TLC. Intact and degraded α-factor were resolved by TLC (using preparative 2 mm silica-gel 60 plates) as described, and fluorographs on Kodak XAR-5 films were quantified by densitometric scanning on a computing densitometer (Molecular Dynamics, Sunnyvale, CA). Quantification was done by scanning both the total lane and the intact α-factor. The percentage of degraded α-factor was obtained by determining the percentage of intact pheromone and subtracting that value from 100%.

To determine whether the delay of α-factor degradation in Δypt7 PEP4 cells is due to the defect in vacuole biogene-sis (Wichmann et al., 1992) or to the decrease in pheromone transport to the vacuole, we investigated whether α-factor was correctly delivered to the vacuole in the Δypt7 mutant. Therefore, we examined the subcellular distribution of α-factor as compared to the vacuolar marker carboxypeptidase Y (CPY) under internalization conditions that in wild-type cells result in the complete delivery of α-factor to the vacuole. For these studies, we performed cell fractionation experiments of YPT7 and Δypt7 strains in a pep4-3 back-ground as described in Materials and Methods. After the internalization of α-factor for 60 minutes at 30°C in Δypt7 cells, we observed 29% of the pheromone in the endosome-containing 100,000 g pellet P3. The majority of the CPY (75%), however, was found in the vacuole-containing 3,500 g pellet P1, suggesting firstly, that the Δypt7 vacuole, despite its fragmentation, behaves normally in fractionation and, secondly, that most of the intracellular CPY is correctly localized to vacuoles (Fig. 1A). The fractionation pat-tern of α-factor after internalization for 60 minutes at 30°C is clearly distinct from that of vacuolar CPY showing that the bulk of the two markers are not in the same organelle. In contrast, in YPT7 cells under identical internalization conditions, the pheromone cofractionated with CPY (Fig. 1B). These data suggested that lack of a functional Ypt7p resulted in a delay of endocytic transport to the vacuole.

We examined the kinetics of α-factor transport to the vacuole in both YPT7 and Δypt7 strains. For these studies, we performed fractionation experiments in a pep4-3 back-ground as described in Materials and Methods. Fig. 2 shows the percentage of total α-factor in both endosome-(P3; Fig. 2A) and vacuole-containing fractions (P1; Fig. 2B) as a function of internalization time. The amount of α-factor found in P3 remained relatively constant between 20 and 60 minutes in the Δypt7 mutant whereas the amount of α-factor decreased in the P3 derived from YPT7 cells (Fig. 2A). In YPT7 cells, the majority of the pheromone was shifted into P1 with time whereas in Δypt7 cells no accumulation in P1 was observed (Fig. 2B). These data suggested that α-factor accumulated in endosomes in the Δypt7 mutant when the majority of the pheromone had reached the vacuole in wild-type cells.

To examine transport through early and late endosomes of both Δypt7 mutant and wild-type, we compared α-factor profiles obtained from floatation of P3 fractions on Nyco-denz gradients (Singer and Riezman, 1990). Under the above internalization conditions, we observed that α-factor transport to the early endosome (fractions 8-10; Singer and Riezman, 1990; Singer-Krüger et al., 1993) occurred with comparable kinetics in both YPT7 and Δypt7 cells (Fig. 3). However, in cells lacking Ypt7p, the amount of α-factor detectable in early endosomes remained constant at times where in wild-type cells most of the pheromone had already passed through this organelle (see 15, 20 and 40 minute time points, Fig. 3A,B). The amounts of α-factor found in the late endosome (fractions 4-6) were very similar in both strains from 2 to 15 minutes. However, in Δypt7 mutants, we detected an accumulation of α-factor in late endosomes between 40 and 90 minutes (Fig. 3A). During that period, a constant amount of the pheromone was detectable in that part of the gradient. This suggested that transport from late endosome to vacuole was very slow in the mutant. In wild-type cells, approximately 80% of the α-factor is degraded by 30 minutes (see Fig. 5B), indicating that most of the pheromone was chased out of both early and late endosomes.

To determine clearly whether transport from early to late endosomes was normal in cells depleted of Ypt7p, we per-formed another time course experiment with both YPT7 and Δypt7 strains. α-Factor was accumulated in early endosomes and subsequently chased into late endosomes by allowing α-factor internalization for 20 minutes at 15°C and then shifting the cells to 32°C for 1, 15 and 30 minutes. The resulting α-factor profiles are shown in Fig. 4. After 20 minutes of preincubation at 15°C, the pheromone was found in early endosomes with a minor fraction having reached the late compartment. The ratio of α-factor in early compared to late endosomes was very similar in both YPT7 and Δypt7 cells. Shift to 32°C for 1 minute showed normal transport to late endosomes in cells deficient in Ypt7p (Fig. 4A). However, with time, α-factor accumulated in late endosomes whereas in wild-type cells, the majority of the pheromone was chased out of that compartment under the same conditions (Fig. 4).

The data presented above show that the kinetics of α-factor transport through the late endosome is significantly delayed in the Δypt7 mutant. Interestingly, at times corresponding to late time points after internalization (40 to 90 minutes), α-factor degradation had already begun in Δypt7 PEP4 cells (Wichmann et al., 1992; see below). Two possibilites exist to explain these findings. First, some of the α-factor could already have been transported to the vacuole in the Δypt7 mutant at these time points. Second, slow degradation could have occurred in an endosomal compartment if enzymes targeted to the vacuole and the α-factor are found in the same compartment. To test this hypothesis, we examined the degradation of α-factor during time course experiments with both wild-type and Δypt7 mutant strains (PEP4 strains) as described in Materials and Methods. After different times of α-factor internalization, two samples were removed. The first sample was directly processed for the detection of α-factor degradation. The second sample was further incubated to a total of 120 minutes at 30°C in the presence of NaN₃ and NaF before being processed for the detection of pheromone degradation. The metabolic inhibitors, NaN₃ and NaF, are known to block further transport of α-factor through the entire endocytic pathway and therefore prevent delivery of the pheromone to the vacuole (Singer and Riez-man, 1990). With this protocol, we detected additional α-factor degradation in the absence of further transport to the vacuole in Δypt7 cells (Fig. 5A), but not in wild-type cells (Fig. 5B). Additional energy-independent degradation in Δypt7 cells was observed after 15 minutes and before 90 minutes of primary incubation at 30°C. After 40 minutes of internalization, α-factor reached a compartment in Δypt7 cells where it could be completely degraded with time (Fig. 5A).

To be certain that the additional energy-independent α-factor degradation in Δypt7 mutants was specific and required vacuolar proteases, we determined whether the process was PEP4-dependent. Degradation of α-factor is known to occur in the vacuole under normal conditions and to require the PEP4 gene (Chvatchko et al., 1986; Singer and Riezman, 1990), which encodes proteinase A. Pro-teinase A is responsible for the processing of a number of other vacuolar proteases (Ammerer et al., 1986; Woolford et al., 1986), some of which are involved in the degradation of α-factor, for example proteinase B and CPY (B. Singer-Krüger, unpublished results). Therefore, the same degradation experiment was performed using a Δypt7 mutant that harbored a mutation in the PEP4 gene. Fig. 6 shows that degradation did not occur in cells that lacked a functional PEP4 gene product, irrespective of whether they had been further incubated with energy poisons or not.

In order to detect small amounts of energy-independent degradation in YPT7 cells, we performed another analysis of α-factor degradation. α-Factor was internalized for 20 minutes at 15°C, the cells were shifted to 30°C, were treated at various times with NaN₃ and NaF, and further incubated as described above, or were extracted directly for analysis of α-factor degradation. Under these conditions, we again observed a pronounced additional energy-independent degradation of α-factor in Δypt7 cells (Fig. 7A), but only perhaps a slight differential degradation behavior in wild-type cells (Fig. 7B).

Cells lacking the small GTPase Ypt7p show greatly reduced kinetics of transport through endosomal compartments and accumulate α-factor in the late endosome. We believe that Ypt7p is not required for initial steps of α-factor transport in the yeast endosomal membrane system because arrival in both early and late endosomes is not impaired in cells deficient in Ypt7p. It was speculated that Ypt7p regulates transport from the early to the late endosome (Wichmann et al., 1992), but no direct evidence has been provided yet to support this hypothesis. Our approach clearly addresses the question of at which stage in the endocytic pathway Ypt7p may function. Our data suggest that Ypt7p is involved in transport steps from the late endosome to the vacuole as α-factor mainly accumulates in late endosomes in Δypt7 cells under conditions in which the pheromone is completely degraded in wild-type cells. In wild-type cells under normal conditions, transport from late endosomes to the vacuole is probably not rate-limiting because the amount of α-factor detectable in early endosomes is higher than in late endosomes. Absence of Ypt7p may introduce a new rate-limiting step into the process, slowing down transport mainly from late endosome to vacuole. Absence of the protein probably does not completely block this trans-port step but delays it to a remarkable extent. The fact that in the Δypt7 mutant a considerable fraction of α-factor remains detectable in early endosomes for a prolonged period may be due to a backup caused by the accumulation of endocytic material in late endosomes.

The possibility that the delay of α-factor transport to the vacuole in Δypt7 mutant cells is due to the defect in vacuole biogenesis seems rather unlikely to us. Even though the vacuole exhibits an abnormal morphology, it seems to be a functional compartment. Soluble vacuolar hydrolases are eventually fully activated (Wichmann et al., 1992) and a significant fraction of CPY is correctly localized to vac-uoles in Δypt7 cells.

While yeast mutants defective in small GTPases of the secretory pathway, such as sec4 and ypt1 mutants, show a complete block in the transport steps they regulate, the phenotype of the Δypt7 mutant is leaky exhibiting a delay, but not a complete block in endocytic traffic. This difference may be due to the presence of two parallel endocytic pathways or to the activation of a salvage pathway upon disruption of YPT7. Disruption of clathrin heavy chain function in yeast induces an alternative pathway for delivery of vacuolar hydrolases from the Golgi to the vacuole (Seeger and Payne, 1992). This same alternative pathway could be acting here if the endocytic and vacuole biosynthetic path-ways truly overlap.

In the mammalian endocytic pathway, variations in the expression of the rab5 protein were shown to correlate with the rate of endocytic vesicle fusion (Bucci et al., 1992). No similar correlation has been described so far for small GTPases involved in secretion. It may be possible that there are fundamental differences between protein transport through the two pathways. In fact, parts of the endocytic pathway have been proposed to involve maturation rather than vesiculation steps (Murphy, 1991). Therefore, at this stage, we do not know whether Ypt7p is absolutely required for late endosome to vacuole traffic.

Vida et al. (1993) have provided fractionation data in favor of the idea that pathways of vacuole biogenesis and endocytosis intersect in a prevacuolar/endosomal compartment. Here we use completely different techniques to pro-vide additional evidence supporting this hypothesis. Our data strongly suggest that endocytosed α-factor and vacuolar hydrolases are in direct contact in a prevacuolar/endosomal compartment because the internalized α-factor can be slowly degraded in a PEP4-dependent (depending specifically on vacuolar hydrolases) and energy-independent manner (after further endocytic transport has been blocked) in Δypt7 cells. Degradation may be slow for two reasons. First, the concentration of the correponding vacuolar hydrolases is probably low in this compartment. The bulk of the CPY fractionates in P1, where vacuoles sediment. Second, the newly synthesized hydrolases are made as proforms and may not be fully activated yet. In wild-type cells, we did not detect any significant degradation that occurred in the absence of further transport to the vacuole. This could be because newly synthesized vacuolar proteins are so efficiently transported to the vacuole that the amount of proteases in late endosomes is too low. In contrast, in

Δypt7 cells, the steady-state concentration of vacuolar hydrolases should be higher in endosomes due to the trans-port defect from late endosomes to the vacuole. Conse-quently, this leads to a prolonged exposure of α-factor accumulating in late endosomes to an abnormally high concentration of the hydrolases involved in pheromone degradation. This is consistent with the difference in the subcellular distribution of vacuolar CPY between YPT7 and Δypt7 strains. In wild-type cells, the hydrolase is exclusively localized to the vacuole whereas in cells lacking a functional Ypt7p a minor but considerably increased per-centage of CPY is found in the endosome-containing 100,000 g pellet.

Although it is difficult to determine unequivocally at what stage in the endosomal membrane system α-factor comes into contact with vacuolar hydrolases, the energy-independent degradation that is observed in the Δypt7 mutant coincides with the time when the pheromone begins accumulating in late endosomes. However, this does not rule out the possibility that the two pathways intersect ear-lier. It is possible that the pathways intersect at the early endosome, but that endocytic content and hydrolases are rapidly transported to late endosomes. We would not have necessarily detected energy-independent degradation in the early endosome because transport out of this compartment is nearly normal in the Δypt7 mutant. In mammalian cells, the late endosome/prelysosomal compartment is generally considered to be the point where endocytosed and newly synthesized lysosomal proteins intersect prior to being delivered to lysosomes (for a review see Kornfeld and Mell-man, 1989). It is still possible that the two pathways meet in the early endosome in animal cells, but this may not have been detected due to the experimental techniques used.

The data presented in this paper suggest that Ypt7p reg-ulates transport from late endosomes to the vacuole and that this transport step is shared by the endocytic and vacuole biosynthetic pathways.

We are grateful to Dieter Gallwitz providing the original Δypt7 mutant strain before publication. We thank Stephan Schröder, Marc Pypaert and especially Linda Hicke for critically reading the manuscript and for fruitful discussions, and Birgit Singer-Krüger for advice on fractionation. Thanks to Fabienne Crausaz and Thomas Aust for technical assistance. This work was supported by the Swiss National Science Foundation and by the Kanton Basel Stadt.