Saccharomyces cerevisiae adapts to changing nutrient conditions by regulating its genome-wide transcription profile and cell-wide protein complement in correlation with the reigning nutrient conditions. The target of rapamycin (TOR) signalling pathway is one of the major control mechanisms within the cell that facilitates these changes. The transcription, intracellular trafficking, and protein turnover of nutrient transporters, including the hexose transporter proteins (Hxts), are regulated in response to nutrient conditions. The Vid and Gid proteins facilitate the nutrient-dependent degradation of the gluconeogenic enzymes FBPase and Mdh2p when glucose-starved cells are replenished with glucose. Three members of the VID and GID gene families, VID30/GID1, GID2, and VID28/GID5 are needed for the rapamycin or nitrogen starvation-induced degradation of the high-affinity hexose transporter Hxt7p is shown here. In addition, evidence that the functions of several Vid and Gid proteins are in close relation to the TOR signalling pathway is provided.
Vid28p and Vid30p
rapamycin-induced Hxt7p turnover
Saccharomyces cerevisiae senses and adapts to its changing nutrient environment by altering its transcriptional profile and modifying the protein complement within the cell. For example, yeast cells are starved of glucose when grown with either ethanol or acetate as a carbon source. When these cells are replenished with glucose, gluconeogenic enzymes such as fructose-1,6-bisphosphatase (FBPase) and malate dehydrogenase (Mdh2p) are degraded (Horak & Wolf, 1997; Huang & Chiang, 1997; Chiang & Chiang, 1998). The degradation of FBPase can occur in the vacuole or the proteasome. Upon glucose replenishment of 3 days glucose-starved cells, FBPase is targeted to novel vacuolar import and degradation (vid) vesicles, which in turn fuse with vacuoles to result in the vacuolar degradation of FBPase (Brown et al., 2003; Regelmann et al., 2003; Hung et al., 2004). A group of Vacuolar Import and Degradation (VID) genes, including VID22, VID24, VID27, VID28, and VID30, are needed for this function (Hoffman & Chiang, 1996; Huang & Chiang, 1997). Alternatively, upon glucose replenishment of one day glucose-starved cells, FBPase is ubiquitinated and targeted to proteasomes for degradation in a manner dependent on nine Glucose Induced Degradation (GID) genes (GID1 to GID9) (Schork et al., 1995; Regelmann et al., 2003). Some GID and VID genes (GID1/VID30, GID4/VID24 and GID5/VID28) encode the same protein product and therefore function in the proteasomal and vacuolar degradation of FBPase (Brown et al., 2003; Regelmann et al., 2003; Hung et al., 2004).
A systematic proteomics approach to identify protein–protein interactions in S. cerevisiae led to the identification of multiple in vivo protein complexes of which one complex contained seven Vid or Gid proteins (Ho et al., 2002). This complex, consisting of Vid30p, Gid2p, Vid24p, Vid28p, Gid7p, Gid8p, Gid9p, and Moh1p, was recovered from GID7 and MOH1 tandem affinity purification (TAP)-tagged strains (Ho et al., 2002). Independently, Regelmann et al., (2003) identified Gid2p as a member of a large cytoplasmic complex (c. ∼600 kDa), termed the Gid protein complex (GPC) and proposed that it contained those components identified in the systematic protein complex identification performed by Ho et al., (2002). All the components of the GPC, except Moh1p, are needed for FBPase degradation (Chiang & Chiang, 1998; Brown et al., 2003; Regelmann et al., 2003). Consistently, Vid30-TAP was used to identify the Vid30p protein complex (Vid30c) (Pitre et al., 2006), which contained all the interacting Vid and Gid proteins identified by Ho et al., (2002) with the exception of Gid7p and Moh1p, but included the uncharacterized protein Ydl176wp. Using a protein–protein interaction prediction algorithm to investigate interactions between components of the Vid30c, Pitre et al., (2006) proposed that the strongest interaction within the Vid30c existed between Vid30p and Vid28p to form the core components of the proposed Vid30c.
Several lines of evidence specific to individual components of the Vid30c indicate its involvement in the nutrient-related adaptation of S. cerevisiae. As stated, all the components of the Vid30c are needed for the degradation of FBPase when yeast cells adapt to glucose replenishment following glucose starvation (Chiang & Chiang, 1998; Brown et al., 2003; Regelmann et al., 2003). In addition, VID30 is needed for the effective transcription of nitrogen-regulated genes (van der Merwe et al., 2001), and Gid2p functions in the carbon-related ubiquitination and subsequent proteasomal degradation of FBPase (Regelmann et al., 2003). Also, Vid24p was identified as a peripheral membrane protein that participates in the fusion of vid vesicles with the vacuole to enable the degradation of FBPase (Chiang & Chiang, 1998; Brown et al., 2003). Very little is known about the specific role(s) of the remaining members of the Vid30c in nutrient-mediated adaptation of yeast.
Yeast cells use well-defined signalling pathways to respond to nutrient conditions. The target of rapamycin (TOR) pathway controls a variety of cellular processes, including cell cycle progression, ribosome biogenesis and tRNA synthesis, translation, nitrogen- and carbon-regulated gene expression, and autophagy (Pedruzzi et al., 2000, 2003; Schmelzle & Hall, 2000; Raught et al., 2001; Lorberg & Hall, 2004; Powers et al., 2004; Roosen et al., 2005). Rich nutrient conditions activate the pivotal TOR kinases while nutrient limitation or treatment of the yeast with the macrolide antibiotic rapamycin renders TOR inactive (Heitman et al., 1991). TOR regulates the activities of a series of intracellular phosphatases and kinases, such as PP2A, Sit4p and Npr1p, which in turn control a wide range of intracellular processes. For example, the nucleocytoplasmic translocation of many transcription factors, including the redundant stress response transcription factors Msn2/Msn4p, the activators Gln3p and Gat1p that are needed for nitrogen- and carbon-related gene activation, and Rgt1/3p that activates the transcription of respiratory genes, are controlled by TOR (Cooper, 2002; Crespo et al., 2002; Powers et al., 2004). In addition, the ubiquitination and turnover of several nutrient-regulated permeases, including Gap1p, Tat2p, Hip1p, and Hxt1p, are controlled by TOR via Npr1p and the E3 ubiquitin ligase Rsp5p (Schmidt et al., 1998; Beck et al., 1999; De Craene et al., 2001).
Hexose transporters mediate the uptake of hexose sugars in S. cerevisiae. Low-affinity hexose transporters, like Hxt1p and Hxt3p, are active when fermentable sugars such as glucose and fructose are abundant, while high-affinity transporters, such as Hxt6p and Hxt7p, are active when these sugars are limiting. The transcription of the HXT genes and the abundance of their protein products are regulated in response to the reigning nutrient conditions. The transcription of HXT7 is repressed by the presence of high concentrations of glucose and nitrogen starvation, but induced by nonfermentable carbon sources such as raffinose and ethanol (DeRisi et al., 1997; Gasch et al., 2000). Also, the presence of high concentrations of glucose and a limiting amount of nitrogen leads to the rapid internalization of Hxt7p from the plasma membrane and its targeting to the vacuole for degradation (Krampe et al., 1998; Krampe &Boles, 2002). Hxt7p contains three ubiquitination sites (Peng et al., 2003) and its internalization occurs in a mechanism dependent on Rsp5p, Doa4p and the endocytic pathway protein End4p (Krampe et al., 1998; Krampe & Boles, 2002). It therefore seems feasible to suggest that the endocytosis of Hxt7p occurs in an ubiquitin-dependent manner.
Hxt7p was identified as a binding partner of Gid7p (Ho et al., 2002). Also, the gid7Δ mutant has an increased ability to consume glucose in nitrogen-limited growth conditions (Gardner et al., 2005). At least one component of the Vid30c, Gid2p, is needed for the ubiquitination of a target protein, FBPase, and its subsequent degradation (Regelmann et al., 2003). It was therefore hypothesized that the GPC/Vid30c or individual members of this complex are needed for the regulation of Hxt7p turnover in response to changing nutrient conditions. In this study, the involvement of TOR and the GPC/Vid30c was investigated in Hxt7p turnover by analysing the degradation of Hxt7p in cells lacking Vid28p, Vid30p, Gid2p, or Gid7p.
Materials and methods
Strains and growth conditions
All the yeast strains used in this study are listed in Table 1. Yeast cells were grown in YPD medium (1% yeast extract, 2% Bacto peptone, 2% glucose; 2% agar for solid media) for preculturing and transformation purposes and were always incubated at 30 °C.
For growth analyses yeast cells were precultured in YPD to exponential growth phase, washed with sterile water and cell densities adjusted to an initial OD600 nm of 1.0. Tenfold dilutions were transferred to solid YPD media containing either rapamycin [1 mg mL−1 in drug vehicle (90% Ethanol and 10% Tween-20)] at final concentrations of 20 or 200 ng mL−1, or drug vehicle (DV) as a control. YPD+DV plates were incubated for 3 days and the YPD+rapamycin plates for 4 days.
Yeast strains used to monitor Hxt7-GFP localization and Hxt7-3HA degradation were precultured in YPD to generate biomass. Cells were washed and transferred to synthetic media [0.17% (w/v) Yeast Nitrogen Base without amino acids and ammonium sulphate] containing 3% (w/v) raffinose and 0.5% (w/v) ammonium sulphate (hereafter raffinose with ammonium). Amino acids were added to complement auxotrophic requirements. Following a 4 h incubation to stimulate HXT7 expression, cultures were divided in two equal volumes. Samples were collected for either fluorescence microscopy or protein extraction (time 0). Rapamycin was added to one volume to a final concentration of 200 ng mL−1 and drug vehicle added to the other. Samples were subsequently collected at the indicated times and used for fluorescence microscopy or protein extraction.
The nutrient-dependent turnover of Hxt7-GFP was performed by shifting strains cultured in raffinose with ammonium media for 4 h to YNB-based media containing either 5% glucose or 3% raffinose as carbon sources. In addition, these media either contained ammonium sulphate at 0.5% (w/v) (glucose with ammonium or raffinose with ammonium) as a nitrogen source or were devoid of ammonium sulphate (glucose without ammonium or raffinose without ammonium). Samples were collected for fluorescence microscopy before the shift (time 0) and at the indicated time intervals following the shift to fresh media.
The analysis of Hxt1-3HA degradation was performed by growing transformants of pTB380 (gift from Mike Hall) in glucose with ammonium media to exponential phase. Cells were washed and shifted to fresh glucose with ammonium media for 4 h before rapamycin was added to a final concentration of 200 ng mL−1. Samples for protein extraction were collected before (time 0) and 3 and 6 h following rapamycin addition.
The chromosomally tagged strains used in this study were created using the PCR-based integrative transformation procedure described previously (Longtine et al., 1998). The primers used contained 100 nt at the 5′ ends homologous to the native chromosomal locus and 20 nt at the 3′ ends homologous to the specific plasmid used to amplify the respective integration cassettes. For chromosomal tagging of HXT7, the forward and reverse primers were designed with homologous sequences upstream and downstream of the stop codon, respectively. Plasmids pFA6a-3HA-His3MX6 and pFA6a-GFP(S65T)-His3MX6 were used to generate the respective integration cassettes (Longtine et al., 1998). Following transformation, the correct integration events were verified by PCR.
The BYvid28 HXT7-GFP and BYgid2 HXT7-GFP strains were used to construct the BYvid28vid30 HXT7-GFP and BYgid2vid30 HXT7-GFP double mutants, respectively. The regions of the primers homologous to the yeast genome were 80 nt in length and were designed to replace the entire coding region of VID30 with the natMX6 cassette of pFA6a-natMX6 (Van Driessche et al., 2005). Using a similar approach, the BY4742 strain served as the parent strain to construct the BYure2 mutant. Homologous primers were designed to replace the URE2 gene with the kanMX6 cassette amplified from pFA6a-kanMX6 (Longtine et al., 1998).
RNA extraction and Northern analysis
Yeast strains were cultured in the indicated media before RNA extraction. The phenol-based RNA extraction procedure, electrophoresis and RNA transfer to nylon membrane were described previously (van der Merwe et al., 2001). A PCR digoxigenin Probe Synthesis Kit (Roche) was used according to the manufacturer's recommendations to generate digoxigenin-labelled probes with primers VID28F2 (TGGGAGACCAGTTGGCTAAG) and VID28R2 (AATGGATCAAACCACCAAGG) to detect VID28, ACT1F6 (ACCAACTGGGACGATATGGA) and ACT1R6 (TAATACGACTCACTATAG GGCCACCAATCCAGACGGAGTA) to detect ACT1, and GFP-FP (GAGAAGAACTTTTCACTGGAG) and GFP-RP (TAGTTCATCCATGCCATGTG) to detect HXT7-GFP transcripts. Prehybridization was performed at 50 °C using 10 mL digoxigenin Easy Hyb solution (Roche) for 30 min. Denatured probes were added to 5 mL digoxigenin Easy Hyb solution and hybridized for 18 h. Membranes were washed for 2 × 5 min in 2 × SSC/0.1 × sodium dodecyl sulphate (SDS) at room temperature and 2 × 15 min in 0.1 × SSC/0.1 × SDS at 50 °C. Membranes were blocked for 30 min, followed by a 30 min incubation with anti-digoxigenin antibodies (1 : 10 000 in blocking solution; Roche) and detected with CDP-Star (Roche) for 5 min. Membranes were exposed to autoradiography film for visualization.
The monitoring of the subcellular localizations of Hxt7-GFP were performed by preparing slides directly from the indicated cell cultures followed by immediate analysis using the × 100 objective lens of a Nikon Eclipse E600 microscope. Images were recorded using a Coolsnapfx monochrome CCD digital camera (Roper Scientific) and processed using Metamorph (Universal Imaging, Version 5.0).
FM4-64 staining was used to confirm Hxt7-GFP is internalized via the endocytic pathway. Wild type and vid28 mutant strains were incubated in raffinose with ammonium media and incubated for 210 min to induce HXT7 expression. The cells were concentrated (OD600 nm=1.25) and FM4-64 (16 mM stock in DMSO) was added to a final concentration of 80 μM. The cells were incubated with FM4-64 for 15 min, harvested and washed, and resuspended in fresh raffinose and ammonium medium. Rapamycin was added to a final concentration of 200 ng mL−1 followed by a 6 h incubation after which the cells were collected for analysis and imaging as described.
Protein extraction and Western blotting
Harvested cells were resuspended in a lysis buffer (0.7 M Sorbitol, 50 mM Tris pH 7.5, 2 mM phenylmethylsulphonyl fluoride, 5% SDS; plus protease inhibitor cocktail tablets, Roche), added to 0.3 g glass beads and vortexed for 2 min. Lysates were centrifuged at 16 000 g for 20 min to remove cell debris and the supernatants collected. Protein concentrations were determined using the DC Protein Assay (Biorad) according to the manufacturer's recommendations. Equal amounts of protein were separated by SDS-polyacrylamide gel electrophoresis (PAGE) using a 7.5% acrylamide gel and transferred to a nitrocellulose membrane for 1 h. The ECL Detection kit (GE) was used to detect the Hxt7-3HA or Hxt1-3HA protein according to the manufacturer's recommendations. Rabbit antihemagluttinin (HA) antibody (Sigma) was used as a primary antibody and donkey anti-rabbit immunoglobulin G conjugated to horseradish peroxidase (GE) as a secondary antibody. Equal amounts of protein in each lane were confirmed with rabbit antialdehyde dehydrogenase (ADH) (200-4144, Rockland) primary antibody as described previously (Onodera & Ohsumi, 2004). Membranes were exposed to autoradiography film for visualization.
Several gid and vid mutants have altered sensitivity to rapamycin
VID30 is needed for the turnover of FBPase in response to a carbon shift (Hoffman & Chiang, 1996; Huang & Chiang, 1997; Regelmann et al., 2003) and the vid30Δ mutant is hypersensitive to rapamycin (van der Merwe et al., 2001). In combination these lines of evidence suggested that the function(s) of the VID/GID genes were controlled by a nutrient-sensing pathway such as the TOR pathway. The growth of eleven mutants associated with either the GPC or VID30c in the presence of increasing concentrations of rapamycin was consequently analysed. The cells of the mutant (vid30Δ, gid2Δ, gid3Δ, vid24Δ, vid28Δ, gid6Δ to gid9Δ, ydl176wΔ, and moh1Δ) and wild-type strains were precultured as described and transferred to YPD plates containing either the drug vehicle or rapamycin. The isogenic gln3Δ and ure2Δ mutant strains were included as controls as they are known to be rapamycin resistant and hypersensitive, respectively. The experiment was performed in the presence of 20, 50, 100, and 200 nM rapamycin, but only the data of the lowest and highest rapamycin concentrations are shown. Although the wild-type and all the mutant strains showed normal growth in the presence of the drug vehicle, it was evident that seven of the mutant strains were hypersensitive to rapamycin treatment (Fig. 1). The vid30Δ, gid2Δ, gid3Δ, vid24Δ, vid28Δ, gid8Δ, and gid9Δ mutants were all hypersensitive to rapamycin treatment as these strains showed a growth pattern more similar to that of the ure2Δ mutant than the wild type strain. In contrast, the gid6Δ, gid7Δ, moh1Δ, and ydl176wΔ mutants showed a growth phenotype similar to the wild type strain. These strains were also tested for rapamycin hypersensitivity in YP media containing 3% (w/v) raffinose as the sole carbon source. Although all the strains tested were less affected by rapamycin in the presence of raffinose, the patterns of hypersensitivity to rapamycin were similar to that observed for all the strains with glucose as a carbon source (data not shown). These growth characteristics were consistent at 20, 50, 100, and 200 nM rapamycin, indicating that the mutants were affected similarly by low and high concentrations of the drug. These findings suggest that all the VID/GID genes analysed, with the possible exception of GID6, GID7, MOH1, and YDL176w, have a function associated with the TOR signalling pathway.
Components of the Vid30c function in a TOR-dependent manner. BY4742 (wild type) and the indicated deletion mutants were cultured in YPD to exponential growth phase (OD600 nm∼0.5), harvested, washed, and 10-fold dilutions of OD600 nm=1.0 prepared and transferred to YPD-based solid media containing either the drug vehicle or rapamycin at a final concentrations of 20 and 200 nM. Plates were incubated for 3 days (DV) and 4 days (rapamycin).
It is known that the transcription of VID30 is regulated by TOR as rapamycin treatment increases its expression in YPD (van der Merwe et al., 2001). Similarly, rapamycin treatment of the wild type cells in YPD activated the transcription of GID2, GID3, GID7, GID8, and GID9 (Huang et al., 2004). Interestingly, the transcription of VID24 and GID6 were down-regulated in response to rapamycin treatment (Huang et al., 2004). No data could be found that described VID28 transcription in response to rapamycin treatment. The rapamycin hypersensitivity of the vid28Δ mutant (Fig. 1) along with the presence of two potential Msn2/Msn4p DNA-binding sites (STREs) in the VID28 promoter, prompted us to investigate the effect of rapamycin on VID28 expression. Wild type cells were cultured in YPD to exponential growth phase and subjected to rapamycin treatment. Samples were collected for RNA extraction directly before (time 0), and as indicated following rapamycin addition. Rapamycin treatment clearly induced the transcription of VID28 (Fig. 2). After 30 min of treatment the expression of VID28 was induced and this level of expression remained consistent after 60 and 90 min of rapamycin treatment. The increased expression of VID28 was therefore not transient. These lines of evidence indicate that the transcriptional regulation of VID28, like many of the other VID/GID genes, is controlled by the TOR pathway.
Transcription of VID28 is regulated by TOR. BY4742 (wild type) was grown in YPD to exponential phase. Cells were collected for RNA extraction to represent time 0 (0). Rapamycin was added to the remaining culture to a final concentration of 200 ng mL−1 and incubated at 30°C with shaking. Samples were collected for RNA extractions after 30, 60 and 90 min as indicated. Membranes were probed for VID28 and ACT1 as internal control.
Rapamycin-induced Hxt7p turnover is dependent on Vid28p
Gid7p directly interacts with the high-affinity hexose transporter Hxt7p (Ho et al., 2002), which led to the hypothesis that GID7 is involved in Hxt7p turnover. The glucose utilization of wild type, vid28Δ, and gid7Δ strains grown in YPD media containing 20% (w/v) glucose was anslysed. Compared with the wild type, both mutant strains have increased abilities to consume glucose (data not shown). TOR is one of the major nutrient signalling pathways in yeast and vid28Δ is hypersensitive to rapamycin treatment. In combination these observations suggested that Vid28p, or other members of the Vid30c, were involved in hexose transporter turnover.
Three copies of hemagluttinin-encoding DNA (3HA) or the GFP gene was fused to HXT7 in both the wild type and vid28Δ strains. These strains were precultured in YPD to generate biomass. The expression of HXT7 is glucose-repressed and is therefore not activated in YPD. Cells were washed and shifted to fresh raffinose with ammonium media. Following a 4 h incubation to induce HXT7 expression, the cells were treated with rapamycin (see ‘Materials and methods’). Hxt7-GFP was highly expressed in raffinose with ammonium media and was clearly present in the plasma membranes of both the wild-type and vid28Δ mutant strains (Fig. 3a, time 0). Rapamycin treatment of the wild-type cells resulted in the internalization of Hxt7-GFP and ultimately its degradation with time. Loss of Hxt7-GFP from the plasma membrane, as seen in the decreased fluorescence of the plasma membrane, started after 3 h and continued through 6 h until its absence after 12 h of rapamycin treatment (Fig. 3a). Hxt7-GFP localized to internal, vesicle-like structures reminiscent of multivesicular bodies (MVBs) (Nikko et al., 2003; Luhtala & Odorizzi, 2004). A stain of wild type and vid28Δ mutant cells with FM4-64, a dye used to identify the vacuole and components of the endocytic pathway, showed the colocalization of Hxt7-GFP and FM4-64 confirming these structures as components of the endocytic pathway (Fig. 3c).
Hxt7p internalization is rapamycin induced and at least partially dependent on VID28. BY4742 HXT7-GFP (wild type) and BYvid28 HXT7-GFP (vid28Δ) were cultured as described in ‘Materials and methods’. Cultures were treated with (a) rapamycin at a final concentration of 200 ng mL−1, or (b) drug vehicle (DV; 90% ethanol+10% Tween-20). Samples were collected at the indicated times and analysed using DIC and fluorescence microscopy. (c) BY4742 HXT7-GFP (wild type) and BYvid28 HXT7-GFP (vid28Δ) cells were incubated with FM4-64 followed by rapamycin addition and a 6 h incubation period before cells were analysed using DIC and fluorescence microscopy. The red colour indicates the FM4-64 stain and green the Hxt7-GFP. Colocalization is indicated by orange–yellow colour.
In contrast, wild-type cells treated with drug vehicle still showed fluorescence in the plasma membrane after 24 h (Fig. 3b). Although the internalization of Hxt7p could be observed with the drug vehicle treatment alone as MVBs were formed, Hxt7-GFP was clearly present in the plasma membrane at each time point (Fig. 3b). In combination these data showed that the internalization and degradation of Hxt7-GFP is accelerated in the wild-type strain in response to rapamycin treatment, indicating that TOR inactivation contributed to Hxt7p internalization and ultimately its turnover in the cell. Thus, active TOR can potentially serve to stabilize Hxt7p in the plasma membrane.
The vid28ΔHXT7-GFP strain was analysed in parallel and it was found that Hxt7-GFP clearly remained associated with the plasma membrane for longer following rapamycin treatment. Internal structures started appearing after 6 h of rapamycin treatment and fluorescence was still clearly present in the plasma membrane following 24 h of rapamycin treatment (Fig. 3a). Analysis of the drug vehicle-treated strains showed similar fluorescence intensities in the plasma membranes of the wild-type and vid28Δ strains after 12 h of growth (Fig. 3b). Clear intracellular structures could be observed in the wild type strain after 18 and 24 h of growth. These structures were largely absent from the vid28Δ strain (Fig. 3b). In combination these results show that Hxt7p is partially stabilized in the plasma membrane in the vid28Δ mutant in response to rapamycin treatment.
To confirm these data, the same experiment was performed with wild-type and vid28Δ strains containing HXT7 chromosomally tagged with 3HA. Western analysis performed with these samples confirmed that the levels of Hxt7-3HA detected in the vid28Δ mutant were higher than that of the wild-type strain in response to rapamycin treatment (Fig. 4a). The amount of Hxt7-3HA detected in the wild-type strain was abundant in the absence of rapamycin, but decreased after 6 h and were very low after 12 h of rapamycin treatment. In contrast, the vid28Δ mutant still had clearly detectable levels of Hxt7-3HA present after 18 h of treatment (Fig. 4a). The levels of Hxt7-3HA did therefore decrease in the mutant, but the turnover was markedly delayed in comparison to the wild-type strain.
Degradation of Hxt7p is rapamycin induced and at least partially dependent on VID28. (a) BY4742 HXT7-3HA (wild type) and BYvid28 HXT7-3HA (vid28Δ) strains were grown in raffinose with ammonium media followed by rapamycin addition. Samples were collected at the times indicated for protein extractions. Samples were analysed by Western blotting using anti-HA antibodies to detect Hxt7-3HA. Equal amounts of protein in each lane were confirmed with anti-ADH antibodies. (b) BY4742 HXT7-GFP (wild type) and BYvid28 HXT7-GFP (vid28Δ) were grown in raffinose with ammonium media followed by rapamycin addition. RNAs were extracted from samples collected before and 12 h following rapamycin addition, and used for Northern analysis. A GFP probe was used to detect the transcripts of HXT7-GFP and the ACT1 probe was used as internal control.
The persisting fluorescence of Hxt7-GFP in the plasma membrane of the vid28Δ cells could be due to differences in the transcription of HXT7-GFP between the vid28Δ mutant and wild type strains in response to rapamycin treatment. To investigate this possibility wild type (VID28 HXT7-GFP) and vid28ΔHXT7-GFP strains were grown in raffinose with ammonium and RNA samples were collected before and 12 h following rapamycin addition. Northern blot analysis showed that the levels of HXT7-GFP transcript were similar in wild type and vid28Δ cells, both before and following rapamycin treatment (Fig. 4b). This data indicated that the transcription of HXT7-GFP was not regulated in a manner dependent on VID28. Collectively, these data indicate that Vid28p plays a role in the internalization of Hxt7p in response to rapamycin treatment.
Vid30p and Gid2p are involved in Hxt7p turnover
The clear involvement of Vid28p in rapamycin-induced Hxt7p internalization prompted us to investigate the roles of other Vid and Gid proteins in this process. The internalization and degradation of Hxt7-GFP in the vid30Δ, gid2Δ, and gid7Δ mutants compared with that of the wild type strain in response to rapamycin treatment. Hxt7-GFP internalization and turnover was delayed in the vid30Δ and gid2Δ strains, but not in the gid7Δ mutant. Following 12 h of rapamycin treatment Hxt7-GFP was still clearly associated with the plasma membranes of the vid30Δ and gid2Δ strains, but largely absent from the plasma membranes of the wild type and gid7Δ strains (Fig. 5). Western analysis of Hxt7-3HA degradation in the vid30Δ and gid2Δ strains showed the transporter was more stable in these mutant strains compared with the wild type (Fig. 6). In contrast, Hxt7-3HA was internalized and degraded similarly in both the wild type and gid7Δ strains (Fig. 6). In combination these data indicate that in addition to Vid28p, Vid30p and Gid2p are also involved in the degradation of Hxt7p in response to rapamycin treatment.
Other components of the Vid30c are needed for the rapamycin-induced internalization of Hxt7p. BY4742 HXT7-GFP (wild type), BYvid30 HXT7-GFP (vid30Δ), BYgid2 HXT7-GFP (gid2Δ), and BYgid7 HXT7-GFP (gid7Δ) were cultured as described in ‘Materials and methods’. Following rapamycin addition, samples were collected at the indicated times and analysed by fluorescence microscopy.
Degradation of Hxt7p is at least partially dependent on VID30 and GID2. a) BY4742 HXT7-3HA (wild type) and BYvid30 HXT7-3HA (vid30Δ), BYgid2 HXT7-3HA (gid2Δ), and BYgid7 HXT7-3HA (gid7Δ) strains were grown in raffinose with ammonium media followed by rapamycin addition. Samples were collected at the times indicated for protein extractions and were analysed by Western blotting using anti-HA antibodies to detect Hxt7-3HA. Equal amounts of protein in each lane were confirmed with anti-ADH antibodies.
Core components of the Vid30c are partially redundant in Hxt7p internalization
An increased stability of Hxt7p exists in both the vid28Δ and vid30Δ mutants. These observations suggested that Vid28p had a slightly larger role in the turnover of Hxt7p than Vid30p as Hxt7p was more stable in the vid28Δ mutant than the vid30Δ mutant. Vid30p and Vid28p were proposed to be the core components of the Vid30c (Pitre et al., 2006) and could be partially redundant in the functioning of this complex. This potential functional redundancy in the turnover of Hxt7-GFP was analysed using the vid28Δ, vid30Δ, and vid28Δvid30Δ strains. Fluorescence microscopy clearly indicated the increased stability of Hxt7-GFP in the vid28Δvid30Δ double mutant compared with either of the two single mutants (Fig. 7). These observations suggest that the functions of Vid28p and Vid30p overlap in Hxt7p internalization and turnover.
Vid28p and Vid30p have redundant roles in the rapamycin-induced degradation of Hxt7p. BY4742 HXT7-GFP (wild-type), BYvid30 HXT7-GFP (vid30Δ), BYgid2 HXT7-GFP (gid2Δ), BYvid28 HXT7-GFP (vid28Δ), BYvid30gid2 HXT7-GFP (vid30Δgid2Δ), and BYvid28vid30 HXT7-GFP (vid28Δvid30Δ) were cultured as described in ‘Materials and methods’. Following rapamycin addition, samples were collected at the indicated times and analysed by fluorescence microscopy.
Gid2p was needed for the ubiquitination of FBPase targeted for degradation (Regelmann et al., 2003). Also, Gid2p was predicted not to be part of the Vid30c core complex, but rather a member of the subcomplex due to its predicted weak interaction with the core components (Pitre et al., 2006). Analysis of Hxt7-GFP localization in the vid30Δ, gid2Δ, and gid2Δvid30Δ mutants following rapamycin treatment clearly showed an increased stability of Hxt7-GFP in the double mutant vs. either of the single mutants (Fig. 7). However, this increased stability was not as pronounced as that observed in the vid28Δvid30Δ double mutant. These observations indicate that at least three components of the Vid30c overlap functionally in the turnover of Hxt7p in response to TOR signalling.
Rapamycin-mediated degradation of Hxt1p is independent of the Vid30c
Hxt1p is degraded rapidly in response to rapamycin treatment (Schmelzle et al., 2004). To determine if both Hxt1p and Hxt7p were degraded in a Vid30c-dependent manner in response to rapamycin treatment the wild type and vid30Δ, gid2Δ, vid28Δ, and gid7Δ strains were transformed with pTB380 which contains HXT1 fused with 3HA at its 3′ end (Schmelzle et al., 2004). Proteins were extracted from transformants grown in glucose with ammonium media before and following rapamycin treatment. Similar levels of Hxt1p-3HA were detected in the wild type and respective vid/gidΔ mutant strains following 3 h of rapamycin treatment (Fig. 8). Hxt1p turnover is rapid and samples analysed following 6 h of rapamycin treatment showed no detectable Hxt1p-3HA (data not shown). These observations suggest that Hxt1p is degraded in a mechanism that is independent of the Vid30c.
Components of the Vid30c are not needed for the rapamycin-induced degradation of Hxt1p. BY4742 (WT), BYvid30 (vid30Δ), BYgid2 (gid2Δ), BYvid28 (vid28Δ), and BYgid7 (gid7Δ) were transformed with pTB380 (HXT1-3HA). Transformants were grown in synthetic glucose with ammonium media to exponential phase and samples were collected for protein extraction before (time 0) and 3 h after rapamycin treatment. Protein samples were subjected to Western analyses and probed with anti-HA antibodies to detect Hxt1-3HA. Equal amounts of protein in each lane were confirmed with anti-ADH antibodies.
Vid30c function in the nutrient-related turnover of Hxt7p
The rapamycin treatment of yeast cells is known to simulate nutrient starvation (Heitman et al., 1991). Previous studies identified specific nutrient conditions needed for the degradation of Hxt7p. The presence of high concentrations of glucose combined with nitrogen starvation resulted in the rapid degradation of Hxt7p (Krampe et al., 1998; Krampe & Boles, 2002). These findings, in combination with the present results showing that components of the Vid30c are needed for the rapamycin-induced internalization of Hxt7-GFP, implied a role for the Vid30c in the nutrient-dependent internalization and degradation of Hxt7p.
Wild-type HXT7-GFP and vid28Δvid30ΔHXT7-GFP strains were cultured in raffinose with ammonium media to induce Hxt7-GFP expression and shifted these cells to abundant glucose with or without ammonium sulphate and abundant raffinose with or without ammonium sulphate media. Samples were collected at 6 h intervals following the respective media shifts and the internalization of Hxt7-GFP was monitored by fluorescence microscopy. When cells were shifted to glucose-based media, with or without ammonium sulphate, Hxt7-GFP was rapidly internalized from the plasma membranes and degraded at similar rates in both the vid28Δvid30Δ and wild type strains (data not shown).
Next the link between nitrogen availability and the stability of Hxt7-GFP in the plasma membrane was investigated using raffinose as a carbon source to prevent glucose-mediated regulation of Hxt7p. When wild-type HXT7-GFP and vid28Δvid30ΔHXT7-GFP cells were grown in raffinose with ammonium media (time 0), a comparable presence of Hxt7-GFP in the plasma membranes of both wild type and vid28Δvid30Δ double mutant strains (Fig. 9a) was found. Following a shift of these cells into fresh raffinose with ammonium media, the fluorescence intensities remained similar for the first 12 h followed by a slow but gradually decrease in intensity as internal structures became visible. This observation is similar to that seen in the drug-vehicle treated cells (Fig. 3b) and was thought to be a result of nitrogen depletion. This possibility was investigated by analysing the presence of ammonia in the growth media at times 0, 6, 12, and 24 h and found significant levels (1.14±0.03 g L−1) of ammonia were still present in the media after 24 h of growth. The reason for this gradual loss of Hxt7p from the plasma membrane is therefore not due to the depletion of ammonia.
Nutrient-dependent internalization of Hxt7p is at least partially dependent on VID28. BY4742 HXT7-GFP (WT) and BYvid28vid30 HXT7-GFP (vid28Δvid30Δ) were grown in raffinose with ammonium to exponential phase, harvested, washed, and shifted to (a) raffinose with ammonium, and (b) raffinose without ammonium media. Following the nutrient shift, samples were collected at the times indicated and analysed by fluorescence microscopy.
In contrast, Hxt7-GFP was less stable in the plasma membrane of the wild-type strain when cells were shifted to raffinose without ammonium media. Internalization of Hxt7-GFP started occurring within 6 h postshift with a concurrent decrease in fluorescence from the plasma membrane (Fig. 9b). The vid28Δvid30Δ double mutant, however, showed a delayed turnover of Hxt7-GFP in these conditions. Hxt7-GFP was still clearly present with high intensity in the plasma membrane of the vid28Δvid30Δ double mutant after 12 h in the raffinose without ammonium media. In addition, 24 h postshift Hxt7-GFP was still present in the plasma membrane of the vid28Δvid30Δ cells while it was not detectable in the wild type strain. These results mimic the rapamycin-induced internalization of Hxt7p (compare Figs 7 and 9b). Rapamycin treatment might therefore mimic nitrogen starvation-induced Hxt7p turnover. In combination these findings suggest that the nitrogen-mediated internalization and ultimate degradation of Hxt7p is dependent on the Vid30c.
The roles of individual components of the GPC/Vid30c in carbon-related protein turnover is known (Brown et al., 2003; Regelmann et al., 2003; Hung et al., 2004). However, the nutrient-related roles of the VID and GID genes are not necessarily restricted to adaptation to preferred carbon sources. To this end, VID30 is needed for nitrogen-regulated activation of gene expression, it is transcriptionally regulated by TOR and the vid30Δ mutant is hypersensitive to rapamycin treatment (van der Merwe et al., 2001). Here, the nutrient response in yeast is better understood by providing evidence that the functions of the VID/GID genes are related to the TOR signalling pathway. The involvement of the TOR signalling pathway in the regulation of Hxt7p presence in the plasma membrane was investigated and it was found that rapamycin treatment induced the internalization and degradation of this high-affinity hexose transporter. Linking the TOR signalling pathway to the regulation of Hxt7p is consistent with the known role of TOR in the turnover of other nutrient permeases, including Gap1p, Tat2p, and Hip1p (Schmidt et al., 1998; Beck et al., 1999; De Craene et al., 2001; Schmelzle et al., 2004).
It was shown that at least three components of the Vid30c, Vid30p, Gid2p, and Vid28p were needed for the efficient internalization of Hxt7p in response to rapamycin treatment. To the authors' knowledge this is the first report in which components of the Vid30c have been linked to the internalization and degradation of a plasma membrane protein. The colocalization of Hxt7-GFP and FM4-64 indicates that the Vid and Gid proteins not only function in the turnover of proteins via the vid or proteasomal pathways as described previously (Brown et al., 2003; Regelmann et al., 2003; Hung et al., 2004), but also have a role in protein turnover via the endocytic pathway. In combination these observations predict a role for the Vid30c that is pivotal to a variety of protein turnover mechanisms within the cell.
Vid28p and Vid30p were proposed to be the core components of the Vid30c, while Gid2p was suggested to form part of a subcomplex of the Vid30c that interacts weakly with the core complex (Pitre et al., 2006). The loss of Vid28p function delayed Hxt7p turnover more severely than the loss of either Vid30p or Gid2p. A double mutant lacking both VID28 and VID30 showed significant delay in Hxt7p internalization, indicating that these two proteins have some degree of functional redundancy. These data support the proposed model in which all the components of the Vid30c interacts with both Vid28p and Vid30p (Pitre et al., 2006). Thus, deleting either VID28 or VID30 could still support the existence of a complex containing most of the components of the Vid30c. However, the vid28Δvid30Δ double mutant should prevent the formation of the Vid30c. It is important to note that turnover of Hxt7p still occurred in the vid28Δvid30Δ double mutant, albeit at a significantly delayed rate. This indicates that either the remaining components of the Vid30c can still support the degradation observed in the vid28Δvid30Δ double mutant or another pathway that functions in parallel to the Vid30c still enables the internalization and degradation of Hxt7p.
Gid2p has previously been shown to be involved in the ubiquitination of an enzyme targeted for degradation following a nutrient shift (Regelmann et al., 2003). Hxt7p contains multiple ubiquitination sites and is degraded in response to either glucose replenishment or nitrogen starvation or both. These findings show a prolonged presence of Hxt7p in the plasma membrane following rapamycin treatment in cells lacking Gid2p, but internalization and degradation of Hxt7p did occur. The delayed turnover was prolonged in cells lacking both Gid2p and Vid30p, further supporting the model that individual components of the Vid30c function in concert to mediate the turnover of Hxt7p.
The internalization of Hxt7-GFP in the vid28Δ (Fig. 3a and b) and vid28Δvid30Δ mutants (Fig. 7) revealed that the intracellular trafficking of Hxt7-GFP to the vacuole might be affected by the Vid30c. Similar to the wild type strain, these mutant strains have clear intracellular structures, possibly MVBs, visible after 12 h. However, these structures disappear in the mutant strains with a concomitant increase in fluorescence in the plasma membranes. This effect is more pronounced in the vid28Δvid30Δ double than the vid28Δ single mutant. These observations indicate that the targeting of Hxt7-GFP from the MVB to the vacuole might be dependent on the Vid30c. Understanding this phenomenon will require further investigation.
Gardner (2005) showed a gid7Δ mutant ferments glucose and fructose more effectively than its wild-type strain in nitrogen-limited broths. The fermentation characteristics of a gid7Δ mutant were independently studied and had an increased ability to metabolize glucose (data not shown), consistent with the findings of Gardner (2005). The Gid7p-Hxt7p interaction provided a potential explanation for the increased sugar utilization and suggested Gid7p could be involved in the nutrient-dependent degradation of Hxt7p. However, investigations of this study showed that TOR-regulated Hxt7p turnover was not dependent on Gid7p. The specific function of Gid7p remains unclear.
In contrast to HXT7, the low affinity transporter HXT1 is highly expressed in the presence of glucose (DeRisi et al., 1997; Gasch et al., 2000). Upon rapamycin treatment, Hxt1p is degraded in a mechanism dependent on End4p (a component of the endocytic pathway) and Rsp5p (an E3 ubiquitin ligase) (Schmelzle et al., 2004), both of which have been shown to be involved in Hxt7p degradation (Krampe et al., 1998; Krampe & Boles, 2002). The stability of Hxt1p in the vid/gid mutants after 90 min (data not shown) and 3 h of rapamycin treatment (Fig. 8) was tested and it was found that Hxt1p degradation occurred similarly in the mutant and wild-type strains. Although Hxt7p and Hxt1p are needed in the plasma membranes of yeast cells in different carbon conditions, the internalization and degradation of both these permeases are regulated by the TOR pathway and require the proper functioning of similar components of cellular ubiquitination and endocytosis machineries. It is anticipated that the signalling mechanisms downstream of TOR resulting in these two endocytic events would be different. Here it was shown that components of the Vid30c are needed for the TOR-mediated degradation of Hxt7p, but not for Hxt1p degradation.
The investigation of the nitrogen starvation-dependent internalization and degradation of Hxt7p clearly showed that the Vid30c is needed for this process to occur in the absence of glucose. Hxt7p remained present at similar levels in the plasma membranes of both wild type and vid28Δvid30Δ cells grown in raffinose with ammonia media, indicating that the signalling cascade needed for the activation of Hxt7p internalization was not activated in these growth conditions. However, Hxt7p endocytosis was activated in raffinose media lacking nitrogen and this event was dependent on Vid28p and Vid30p. Interestingly, no impact of Vid28p and Vid30p was found on the internalization of Hxt7p in abundant glucose conditions, irrespective of the nitrogen availability. The combination of these two nutrient conditions, high glucose and nitrogen starvation, leads to the rapid degradation of Hxt7p and appears to occur independent of the Vid30c. These findings suggest that Hxt7p degradation occurs in both Vid30c-dependent and Vid30c-independent mechanisms. The precise signalling mechanism that governs the nitrogen-induced Vid30c-dependent internalization of Hxt7p is currently unclear.
The precise molecular mechanism by which Vid28p, and ultimately the Vid30c, function in the TOR-mediated internalization of Hxt7p is currently unclear. However, there are multiple stages in the process of Hxt7p internalization and degradation in response to nutrient conditions where Vid28p and other components of the Vid30c could potentially exert their function. These include (1) regulating the activities of specific phosphatases and kinases, potentially Sit4p or Npr1p, in response to nutrient signals, and (2) activating components of the ubiquitination machinery, potentially Rsp5p, to enable the ubiquitination of Hxt7p, (3) initiating the endocytosis of Hxt7p, and (4) trafficking Hxt7p to the vacuole. Elucidating the precise molecular function of the Vid30c will increase the understanding of nutrient adaptation of S. cerevisiae.
We thank Dr Mark Longtine for supplying plasmids pFAa-kanMX6, pFA6a-3HA-His3MX6 and pFA6a-GFP(S65 T)-His3MX6, Dr Mike Hall for supplying pTB380, and Dr Dev Mangroo for the use of the fluorescence microscope. This research was supported by a National Science and Engineering Research Council (NSERC) Discovery Grant to GvdM.
(1998) Vid24p, a novel protein localized to the fructose-1, 6-bisphosphatase-containing vesicles, regulates targeting of fructose-1,6-bisphosphatase from the vesicles to the vacuole for degradation. J Cell Biol 140: 1347–1356.
et al. (2003) Catabolite degradation of fructose-1,6-bisphosphatase in the yeast Saccharomyces cerevisiae: a genome-wide screen identifies eight novel GID genes and indicates the existence of two degradation pathways. Mol Biol Cell 14: 1652–1663.
(2001) Ammonia regulates VID30 expression and Vid30p function shifts nitrogen metabolism toward glutamate formation especially when Saccharomyces cerevisiae is grown in low concentrations of ammonia. J Biol Chem 276: 28659–28666.