OUP user menu

Sorting defects of the tryptophan permease Tat2 in an erg2 yeast mutant

Katsue Daicho, Nishiho Makino, Toshiki Hiraki, Masaru Ueno, Masahiro Uritani, Fumiyoshi Abe, Takashi Ushimaru
DOI: http://dx.doi.org/10.1111/j.1574-6968.2009.01722.x 218-227 First published online: 1 September 2009


Cholesterol (ergosterol in yeast) in conjunction with sphingolipids forms tight-packing microdomains, ‘lipid rafts,’ which are thought to be critical for intracellular protein sorting in eukaryotic cells. When the activity of Erg9 involved in the first step of ergosterol biogenesis, but not that of Erg6 involved in a late step, is compromised, vacuolar degradation of the tryptophan permease Tat2 is promoted. It is unknown whether this difference simply reflects the difference between the inhibition of early and late steps. Here, it is shown that the deletion in ERG2, which encodes sterol C8–C7 isomerase (the next enzymatic step after Erg6), promotes the vacuolar degradation of Tat2. It suggests that the accumulation of specific sterol intermediates may alter lipid raft structures, promoting Tat2 degradation. The erg2Δ-mediated Tat2 degradation required Tat2 ubiquitination. Lipid raft association of Tat2 is compromised in erg2Δ cells. The erg2Δ mutation showed a synthetic growth defect with the trp1 mutation, indicating that Tat2 sorting is preferentially compromised in these mutants. Consistent with this notion, the raft-associated protein Pma1 was associated with detergent-resistant membranes and sorted to the plasma membrane. This study suggests the potential for the pharmacological control of cellular nutrient uptake in humans by regulating enzymes involved in cholesterol biogenesis.

  • ergosterol
  • ERG2
  • tryptophan permease
  • Tat2


Cholesterol and sphingolipids together are thought to form tightly packed lipid microdomains, called ‘lipid rafts,’ which are resistant to detergent solubilization at low temperature. Hence, lipid rafts are also termed as detergent-insoluble glycolipid-enriched complexes (DIGs) or detergent-resistant membranes. Lipid rafts are critical for intracellular protein sorting (Simons & Ikonen, 1997, 2000; Brown & London, 1998, 2000; Ikonen, 2001). Glycosylphosphatidylinositol-anchored proteins and several specific transmembrane proteins are enriched in lipid rafts. Depletion of cholesterol or sphingolipids results in the mis-sorting of secretory proteins and blocking of endocytosis in mammalian cells.

In the case of the budding yeast Saccharomyces cerevisiae, ergosterol (the yeast functional equivalent of cholesterol in mammalian cells) and sphingolipids are enriched progressively along the secretory pathway, reaching their highest levels at the plasma membrane (Zinser et al., 1993). Other membranes, such as the membrane of the vacuole, are remarkably poor in sphingolipids and exhibit a low ratio of ergosterol to phospholipids (Zinser et al., 1991; Hechtberger et al., 1994; Schneiter et al., 1999). Ergosterol biosynthesis requires a large number of enzymes encoded by ERG genes in yeast (Fig. 1). Recently accumulated data have shown that lipid rafts are also important for membrane trafficking in yeast and are implicated in the secretion of raft-associated proteins: glycosylphosphatidylinositol-anchored protein Gas1; proton ATPase Pma1; various permeases, such as uracil permease Fur4, general amino acid permease Gap1, and tryptophan permease Tat2; and hexose transporter Hxt1 (Bagnat et al., 2000, 2001; Eisenkolb et al., 2002; Abe & Iida, 2003; Dupre & Haguenauer-Tsapis, 2003; Hearn et al., 2003; Malinska et al., 2003; Umebayashi & Nakano, 2003; Lauwers & Andre, 2006; Okamoto et al., 2006).

Figure 1

Sterols in yeast. (a) Sterol synthesis pathway in yeast. For details, see the text. (b) Structures of zymosterol, fecosterol, and ergosterol. For details, see Discussion.

Tat2 is synthesized in the endoplasmic reticulum and trafficked via the secretory pathway to the plasma membrane (Beck et al., 1999). However, Tat2 predominantly localizes in late endosomes and bulk lipids (nonraft lipids) under normal conditions (Abe & Iida, 2003; Umebayashi & Nakano, 2003). Vacuolar sorting and degradation of Tat2 is evoked in response to various environmental conditions, nutrient starvation, high tryptophan, and high pressure stress, or the anticancer drug 4-phenylbutyrate (Beck et al., 1999; Abe & Horikoshi, 2000; Abe & Iida, 2003; Umebayashi & Nakano, 2003; Liu et al., 2004; Miura & Abe, 2004). The protein kinase TOR (target of rapamycin), which is the central controller of cell growth in response to nutrient availability, regulates Tat2 stability. The specific TOR inhibitor rapamycin promotes the vacuolar degradation of Tat2 via the protein kinase Npr1 (Schmidt et al., 1998).

Tat2 vacuolar degradation requires Tat2 ubiquitination mediated by the HECT-domain E3 ubiquitin ligase Rsp5 (a homolog of mammalian Nedd4) (Beck et al., 1999; Abe & Iida, 2003; Umebayashi & Nakano, 2003). Rsp5-interacting proteins Bul1 and Bul2 are also involved in Tat2 ubiquitination (Abe & Iida, 2003; Umebayashi & Nakano, 2003). Inhibition of Tat2 ubiquitination stimulates association of Tat2 with lipid rafts (Abe & Iida, 2003).

Tryptophan uptake activity is decreased in the mutant strain defective in the ERG6 gene, which encodes an S-adenosylmethionine Δ-24-sterol-C-methyltransferase (Gaber et al., 1989). However, the protein level of Tat2 did not decrease in the erg6 mutant when the tryptophan level in the culture medium was maintained at a high level (Umebayashi & Nakano, 2003). Thus, targeting of Tat2 to the plasma membrane is compromised in erg6Δ cells, but vacuolar sorting of Tat2 is not stimulated. In contrast, inhibition of ERG9-encoded squalene synthetase (SQS) by zaragozic acid (ZA) promotes the vacuolar degradation of Tat2 (Daicho et al., 2007). SQS catalyzes the conversion of farnesyl pyrophosphate to squalene. This reaction represents the first committed step in the formation of cholesterol and related sterols and is thought to represent a major control point of isoprene and sterol biosynthesis in eukaryotes (Faust et al., 1979; Brown & Goldstein, 1980; Bruenger & Rilling, 1986; Goldstein & Brown, 1990; Robinson et al., 1993). It is suspected that this difference in the vacuolar degradation of Tat2 between Erg9 and Erg6 dysfunction reflects the difference in the inhibition of early and late steps of ergosterol biogenesis, because cells defective in Erg6 function still possess sterol intermediates, whereas cells defective in Erg9 activity have no such intermediates.

Here, it is shown that a mutant strain deficient in ERG2, which encodes a sterol C8–C7 isomerase (the next enzymatic step after Erg6), stimulates the vacuolar degradation of Tat2 in contrast to the erg6Δ strain. This suggests that the accumulation of specific sterol intermediates alters lipid raft structures and promotes Tat2 degradation. These findings will be discussed in combination with Tat2 ubiquitination.

Experimental procedures

Strains, plasmids, and media

The S. cerevisiae strains used are listed in Table 1. To detect Tat2, plasmids pSCU301 (pHA2-TAT2) and pSCU425 (pHA2-TAT2-5KR) (Beck et al., 1999) were used. The composition of adenine-containing rich medium (YPAD; yeast extract, peptone, adenine and 2% glucose) and synthetic minimal medium (SD) were described previously (Daicho et al., 2007). Normal SD medium contained 20 μg mL−1 of tryptophan.

View this table:
Table 1

Yeast strains used in this study

Name (alias)Description (reference)
SCU733 (RH2886)MATa trp1 ade2 ura3 leu2 bar1 (H. Riezman)
SCU734 (RH488)MATa his4 leu2 ura3 lys2 bar1 (Munn et al., 1999)
SCU719 (erg2Δ)SCU734 with erg2URA3 (Munn et al., 1999)
SCU730 (erg6Δ)SCU734 with erg6LEU2 (Munn et al., 1999)
SCU725 (erg3Δ)SCU734 with erg3LEU2 (Heese-Peck et al., 2002)
SCU727 (erg4Δ)SCU734 with erg4URA3 (Heese-Peck et al., 2002)
SCU729 (erg5Δ)SCU734 with erg5kanMX4 (Heese-Peck et al., 2002)
SCU720 (erg2Δ)SCU719 with erg2ura3loxP-kanMX-loxP (this study)
SCU443 (erg4Δ)SCU727 with erg4ura3loxP-kanMX-loxP (this study)
SCU809 (erg2Δtrp1)Descendant from SCU733 and SCU719 (this study)
SCU816 (erg6Δtrp1)Descendant from SCU733 and SCU730 (this study)
SCU811 (erg3Δtrp1)Descendant from SCU733 and SCU725 (this study)
SCU813 (erg4Δtrp1)Descendant from SCU733 and SCU727 (this study)
SCU814 (erg5Δtrp1)Descendant from SCU733 and SCU729 (this study)

Western blot analysis

Western blot analysis was performed in accordance with the previous report (Daicho et al., 2007), using anti-HA (HA, hemagglutinin) mouse monoclonal (16B12, BAbCo), anti-Pep12 mouse monoclonal (Molecular Probes Inc.), anti-Pma1 mouse monoclonal (40B7, EnCor Biotechnology Inc., Gainesville, Florida), and anti-Cdk rabbit polyclonal (PSTAIRE, Santa Cruz) antibodies.

Indirect immunofluorescence

Indirect immunofluorescence was performed as described previously (Honma et al., 2006). Immunofluorescent detection of HA-tagged Tat2 and Pma1 was performed with anti-HA mouse monoclonal antibody (clone 16B12; BAbCO) and anti-Pma1 rabbit polyclonal antibody (a gift from Maria Cardenas) (Cardenas et al., 1990), respectively, at a final dilution of 1 : 1000 in PBT (0.5% BSA and 10% Triton-X100 in phosphate buffered saline). The samples were treated with primary antibody for 2 h and subsequently with Cy3-conjugated secondary antibody (Molecular Probes Inc.), diluted 1 : 1000 in PBT for 90 min.

DIG isolation

DIG isolation was performed as described previously with slight modifications (Abe & Iida, 2003). Briefly, the cells (10 mL culture in SD grown to midlog phase) were lysed in 200 μL buffer A (50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 1 mM polymethylsulfonyl fluoride, and 10 mM NaN3) and protease inhibitors as described above. Crude extracts were cleared with a 5-min, 500 g centrifugation. The cell extract was subjected to a 10-min, 13 000 g centrifugation to yield a P13 (pellet) fraction. The pellet was incubated in 75 mL of TXNE1 buffer (50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 150 mM NaCl, and 1% Triton X-100) for 30 min in an ice bath. After extraction with Triton X-100, the lysates were adjusted to 40% OptiPrep (Nycomed, Oslo) by the addition of 150 mL of OptiPrep solution and were overlaid with 900 μL of 30% OptiPrep in TXNE0.1 buffer (50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 150 mM NaCl, and 0.1% Triton X-100) and 100 μL of TXNE0.1 buffer. The DIG fraction was floated through these layers by centrifugation (100 000g for 2 h). Four fractions (300 μL) were collected from the top and were mixed with sodium dodecyl sulfate (SDS)–2-mercaptoethanol-sample buffer. The protein sample was placed at 37 °C for 10 min to denature the membrane proteins with SDS.

Tryptophan import activity

Tryptophan import activity of cells was measured as described previously (Daicho et al., 2007). Experiments were repeated with essentially identical results. Only one set of data is shown here.


Several erg mutants are defective in tryptophan uptake

The yeast ERG genes encoding enzymes involved in the later steps of ergosterol biosynthesis (ERG6, ERG2, ERG3, ERG5, and ERG4) (Fig. 1a) are not essential for cell viability (Arthington et al., 1991; Ashman et al., 1991; Hardwick & Pelham, 1994; Lees et al., 1995). However, the erg6Δ strain shows synthetic lethality with trp1, which is alleviated by the addition of high levels of tryptophan (Umebayashi & Nakano, 2003) (see Fig. 2), because the erg6Δ strain is defective in Tat2 sorting to the plasma membrane and in tryptophan uptake. Similar synthetic lethality was found for other viable erg-null mutant strains erg2Δ, erg3Δ, erg5Δ, and erg4Δ. None of the mutants showed significant growth defects in the tryptophan prototrophic (Trp+) background (Fig. 2a, control and data not shown). The growth defect was alleviated by the addition of high concentrations of tryptophan to the medium (Fig. 2, high Trp, 30 °C). These observations strongly suggest that tryptophan uptake is compromised in these erg mutants.

Figure 2

Growth defects of tryptophan auxotrophic erg mutant strains. ergΔ strains show a severe synthetic growth defect with trp1. Cells were incubated on YPAD plates with or without 200 μg mL−1 of additional tryptophan at 30 or 37°C for 3 (left plates) or 4 (right plates) days. The following TRP1 strains were used: wild type (WT, SCU734), erg2Δ (SCU719), erg6Δ (SCU730), erg3Δ (SCU725), and erg4Δ (SCU727). The following trp1 strains were used: trp1 (SCU733), erg2Δtrp1 (SCU809), erg6Δtrp1 (SCU816), erg3Δtrp1 (SCU811), and erg4Δtrp1 (SCU813).

Various end mutants defective in endocytosis show temperature-sensitive growth for unknown reasons (Chvatchko et al., 1986; Benedetti et al., 1994; Horvath et al., 1994; Munn & Riezman, 1994). Although an erg2 mutation was isolated as end11 (Munn et al., 1999), erg2ΔTRP1 cells grew normally at 37 °C. However, the erg2Δtrp1 cell was lethal even though high concentrations of tryptophan were added (Fig. 2, High Trp, 37 °C). These findings suggest that the impedance of tryptophan uptake in the erg2Δ strain is exacerbated at high temperature.

The tryptophan permease Tat2 levels are decreased in erg2 mutants

The reduced tryptophan uptake in erg2Δ, erg3Δ, erg4Δ, and erg5Δ mutants suggests that these erg mutants are defective in Tat2 sorting. It was found that Tat2 levels were reduced markedly in erg2Δ cells (Fig. 3c). It indicates that Tat2 vacuolar degradation was enhanced. On the other hand, Tat2 levels were not decreased significantly in the erg3Δ, erg4Δ, or erg5Δ mutants similar to the erg6Δ mutant (Fig. 3a and b). Vacuolar degradation of Tat2 was not promoted in erg6Δ mutant cells in the presence of sufficient tryptophan, although targeting of Tat2 to the plasma membrane was compromised (Umebayashi & Nakano, 2003). It is likely that similar situations occur in erg3Δ, erg4Δ, and erg5Δ mutants. Thus, ERG2 is required specifically for the inhibition of Tat2 vacuolar sorting, unlike other ERG genes. Therefore, the present study focused on the effects of the erg2 deletion on Tat2 sorting.

Figure 3

Tat2 levels and tryptophan uptake decrease in erg2Δ cells. (a and b) Tat2 protein levels in erg mutants. Cells of wild-type (SCU734), erg2Δ (SCU720), erg6Δ (SCU730), erg3Δ (SCU725), erg4Δ (SCU443), and erg5Δ (SCU729) strains harboring the plasmid pHA2-TAT2 (pSCU301) were used. Protein levels of HA-tagged Tat2 were examined by Western blotting using anti-HA antibody. CDK was used as the loading control. (c) Decrease in tryptophan uptake in erg2Δ cells. Tryptophan uptake activity was measured using strains SCU734 (wild type) and SCU720 (erg2Δ) as described in Materials and methods. WT, wild type.

Tryptophan import activity was indeed reduced in the erg2Δ mutant (Fig. 3a). The decrease in the tryptophan uptake activity in the erg2Δ cells reflects the decrease in Tat2 localized to the plasma membrane. Although the total protein level of cellular Tat2 was reduced substantially in erg2Δ mutant cells, the decrease in the tryptophan uptake activity was not drastic. This result is not so surprising, because the bulk of Tat2 protein is localized to the late endosome but not to the plasma membrane (Beck et al., 1999; Umebayashi & Nakano, 2003; Daicho et al., 2007) (see also Fig. 4b, Tat2, WT). The erg2Δ-mediated decrease in Tat2 levels may mainly reflect the decrease in Tat2 localized to late endosomes. Similar situations were observed in the case of ZA-treated cells (Daicho et al., 2007).

Figure 4

Effects of the inhibition of Tat2 ubiquitination in erg2Δ cells. (a) Inhibition of Tat2 ubiquitination in erg2Δ cells suppresses Tat2 degradation. SCU734 (wild type) and SCU720 (erg2Δ) cells harboring the plasmid pHA-TAT2 (pSCU301) or pHA-TAT2-5KR (pSCU425) were used. Protein levels of HA-tagged Tat2 were examined by Western blotting using an anti-HA antibody. Coomassie brilliant blue (CBB)-stained proteins were used as the loading control. Note that immunoblotting signals were detected more rapidly as compared with those in Fig. 3a, because the Tat2-5KR signal in the wild-type strain was remarkably strong (see Fig. 5). (b) Inhibition of Tat2 ubiquitination in erg2Δ cells promotes Tat2 targeting to the plasma membrane. SCU734 (wild type) and SCU720 (erg2Δ) cells harboring the plasmid pHA-TAT2 (pSCU301) or pHA-TAT2-5KR (pSCU425) were observed with Cy3-stained Tat2 (red) and DAPI-stained DNA (blue) signals. Nomarski images (DIC) were also recorded. Note that the exposure time of the images was not equal in each sample. WT, wild type.

Tat2 degradation in the erg2Δ mutant is dependent on Tat2 ubiquitination

Tat2-5KR is a stabilized version of Tat2, in which five putative ubiquitination target lysine sites in the NH2-terminal region of Tat2 are substituted by arginine, and its protein level in the wild-type strain cells increases strikingly as compared with those of the wild-type Tat2 (Beck et al., 1999) (see Figs 4a and 5). Furthermore, Tat2-5KR is resistant to rapamycin, ZA, 4-phenylbutyrate, and high pressure at the protein level (Beck et al., 1999; Abe & Iida, 2003; Liu et al., 2004; Daicho et al., 2007).

Figure 5

Effect of erg2 deletion on raft- and bulk lipid-associated Tat2. Detergent-insoluble membrane fractions (Fraction 1) were isolated from wild-type (SCU734) and erg2 (SCU720) cells harboring the plasmid pHA2-TAT2 (pSCU301) or pHA2-TAT2-5KR (pSCU425). HA-tagged Tat2 in each fraction was detected using Western blotting. Pma1 and Pep12 were used as raft- and bulk lipid-associated protein markers, respectively.

In the case of erg2Δ cells, Tat2 levels were also increased by this mutation (Fig. 4a, erg2, 5KR; see also Fig. 5, erg2, 5KR, overexposed). This indicates that Tat2 degradation in the erg2Δ mutant is dependent on Tat2 ubiquitination and suggests that ubiquitination-dependent Tat2 degradation is promoted in erg2Δ cells. However, the level of Tat2-5KR in the erg2Δ cells was still lower than that found in wild-type cells. This may not be surprising, because Tat2-5KR has other ubiquitination site(s) (Abe & Iida, 2003; Umebayashi & Nakano, 2003). In addition, ergosterol depletion stimulates abnormal sorting pathways for the vacuolar degradation of Tat2 (Daicho et al., 2007).

Unlike the wild-type Tat2, Tat2-5KR is predominantly localized to the plasma membrane (Beck et al., 1999) (see Fig. 4b, Tat2-5KR, ERG2). Its localization is also resistant to rapamycin and ZA (Beck et al., 1999; Daicho et al., 2007). Inhibition of Tat2 ubiquitination promotes Tat2 targeting to the plasma membrane and represses the vacuolar sorting of Tat2, even under conditions in which vacuolar sorting of the wild-type Tat2 is stimulated. In addition, in the case of erg2Δ cells, Tat2-5KR was effectively sorted to the plasma membrane (Fig. 4b). Thus, inhibition of Tat2 ubiquitination repressed the erg2Δ-stimulated vacuolar degradation of Tat2 and still promoted the sorting of Tat2 to the plasma membrane.

Raft association of Tat2 in erg2Δ cells

The raft component ergosterol is required for the membrane trafficking of raft-associated protein Tat2 from the Golgi to the plasma membrane in yeast cells (Abe & Iida, 2003; Umebayashi & Nakano, 2003). It is also likely that the defect in Tat2 sorting to the plasma membrane results from abnormalities in the lipid raft structure in erg2Δ cells. To examine the effect of erg2 deletion on the raft association of Tat2, membrane lipid rafts (DIGs) were isolated and the associated proteins were detected using Western blotting. The raft marker protein Pma1 (plasma membrane proton ATPase) was found predominantly in the detergent-insoluble fraction (Fig. 5, Pma1, fraction 1), while the bulk lipid-associated protein Pep12 existed mainly in the detergent-soluble fraction (Pep12, fraction 4), as described previously (Abe & Iida, 2003). Tat2 in wild-type cells was contained mainly in the bulk lipid fraction and not in the lipid raft fraction (Fig. 5, ERG2, Tat2), as shown previously (Abe & Iida, 2003; Umebayashi & Nakano, 2003).

However, Tat2-5KR was found not only in the lipid raft fraction but also in the bulk lipid fraction (Abe & Iida, 2003) (Fig. 5, WT, Tat2-5KR). The fact that the RSP5 mutant strain HPG1-1 and bul1Δbul2Δ strain show similar Tat2 distributions (Abe & Iida, 2003) indicates that the 5KR mutations themselves do not increase Tat2 affinity to the lipid rafts. Rather, it is likely that Tat2 has an intrinsic ability to associate with lipid rafts, and that the association is normally repressed by Tat2 ubiquitination.

Pma1 was found in the DIG fraction from erg2Δ cells (Fig. 5). This finding indicates that the lipid raft structure is still present in this mutant (see Discussion). Tat2 was found predominantly in the bulk lipid fraction, not in the raft fraction, in erg2Δ cells, as found in the wild-type strain (Fig. 5, Tat2). However, the amount of Tat2 was lower in erg2Δ cells than in wild-type cells, which was consistent with the decrease in the total cellular amount of Tat2 (Fig. 3a). Tat2-5KR was found mainly in the nonlipid raft fraction in erg2Δ cells, and the level of the raft-associated Tat2-5KR was lower than that of non-raft-associated Tat2-5KR. This is in sharp contrast to the fraction profiles of Tat2-5KR found in the wild-type cells in which the amounts of raft- and non-raft-associated Tat2-5KR were approximately equal. These findings suggest that the ability of association of Tat2 to lipid rafts is impeded in abnormal lipid rafts. Similar observations were obtained in erg6Δ cells (Umebayashi & Nakano, 2003).

Raft association and plasma membrane targeting of Pma1 in erg2Δ cells

Pma1 is properly associated to lipid rafts in erg2Δ cells (Fig. 5). In addition, the fact that erg2Δ mutant cells are viable suggests that the essential plasma membrane protein Pma1 is sorted to the plasma membrane, although Tat2 sorting is impeded strongly. It was indeed the case that Pma1 was localized to the plasma membrane in erg2Δ cells, as also seen in wild-type cells (Fig. 6). Thus, the abnormality of lipid rafts in erg2Δ cells preferentially compromises the plasma membrane sorting of Tat2 but not Pma1.

Figure 6

Pma1 is localized in the plasma membrane in erg2Δ cells. Pma1 distribution in SCU734 (wild type) and SCU720 (erg2Δ) cells were visualized using anti-Pma1 primary and Cy3-stained secondary antibodies. WT, wild type.


Lipid rafts in the erg2Δ

It is shown here that Pma1 is found in the DIG fraction in erg2Δ cells. Similarly, the raft-associated proteins Gap1 and Yps1 were found in the DIG fractions in erg6Δ cells (Sievi et al., 2001; Umebayashi & Nakano, 2003). These findings suggest that these viable erg mutant cells possess lipid raft structures. However, Tat2 sorting is obviously compromised in these erg cells (this study) (Umebayashi & Nakano, 2003). This indicates that the structure in these mutants is aberrant due to the absence of ergosterol and the accumulation of sterol intermediates. Thus, these two mutants have similar features. However, vacuolar degradation of Tat2 is stimulated in erg2Δ cells but not erg6Δ cells (this study) (Umebayashi & Nakano, 2003). This suggests that, as for the function of Tat2 sorting, the structure of abnormal lipid rafts in erg2Δ cells is different from that in erg6Δ cells.

Association to lipid rafts and ubiquitination of Tat2 in erg2Δ cells

Gas1 and Pma1 are preferentially distributed to lipid rafts (Bagnat et al., 2000). In contrast, Tat2 was found in the nonlipid raft fractions (Abe & Iida, 2003; Umebayashi & Nakano, 2003). It is proposed that the relatively higher hydrophobicity of Tat2, as compared with other integral membrane proteins, including Pma1, is sufficient for the localization of Tat2 to the bulk lipid fraction via interaction with the aliphatic chain of phospholipids (Abe & Iida, 2003). Ubiquitination-resistant Tat2-5KR seems to be a good indicator of the affinity of Tat2 to lipid rafts. Tat2-5KR showed that the raft association of Tat2 was impeded in erg2Δ cells (Fig. 5), in a similar manner to that in erg6Δ cells (Umebayashi & Nakano, 2003). This is probably due to abnormalities in lipid rafts composed of sterol intermediates and sphingolipids.

Interestingly, Rsp5 is partially resistant to detergent extraction (Wang et al., 2001). This suggests that Rsp5 is resident in both lipid rafts and nonlipid rafts. However, it is plausible that Rsp5-mediated Tat2 ubiquitination occurs predominantly in nonlipid rafts, because ubiquitination-dependent vacuolar degradation of Tat2 and association of Tat2 to nonlipid rafts are both enhanced in the erg2Δ mutant. In this scenario, Tat2 that escapes from ubiquitination in the bulk lipid fraction could be translocated to lipid rafts, thereafter being targeted to the plasma membrane (Fig. 7).

Figure 7

Model for Tat2 sorting in erg mutant cells. (a–c) Tat2 sorting in wild-type, erg2Δ, and erg6Δ cells. Tat2 is attacked by the ubiquitin ligase Rsp5 in nonlipid rafts in the trans-Golgi to be transported to the late endosome (1), and some proteins are further sorted to the vacuole (2). Nonubiquitinated Tat2 is translocated to lipid rafts to be sorted to the plasma membrane (3). Plasma membrane sorting of Tat2 is impeded in both mutants, but vacuolar sorting is promoted in erg2Δ cells but not in erg6Δ cells. See text for further details.

Sorting of membrane proteins to the plasma membrane or the endosome (and the vacuole) is determined in the trans-Golgi, and the latter sorting is dependent on ubiquitination (for a review, see Umebayashi, 2003). This suggests that the Rsp5 responsible for Tat2 ubiquitination is resident in the trans-Golgi. When Tat2 is degraded in the vacuole, Tat2 in the endosome is further trafficked to the vacuole. Importantly, under normal conditions, Tat2 vacuolar degradation is increased in erg2Δ cells but not in erg6Δ cells (Fig. 2a), and Tat2 is accumulated in late endosomes in erg6Δ cells (Umebayashi & Nakano, 2003). It is most likely that the status of Rsp5-mediated ubiquitination of Tat2 is different in these two mutants, leading to different sorting destinations of Tat2. Because Rsp5 is located in the trans-Golgi, retention in the endosome or further transport to the vacuole of Tat2 may already be determined in the trans-Golgi (Fig. 7).

Zymosterol and cholesta-5,7,24-trienol are the dominant sterols in erg6Δ cells, while fecosterol and ergosta-8-enol are dominant in erg2Δ cells (Munn et al., 1999). Like ergosterol, zymosterol and cholesta-5,7,24-trienol, but not fecosterol nor ergosta-8-enol, have a double bond in a main alkyl chain, which might be critical for the raft association of Tat2 (Fig. 1b). The difference in these sterol structures in both mutants brings about the difference in lipid raft structures, Tat2 association to the lipid rafts, and the status of Rsp5-mediated ubiquitination. Moreover, it is conceivable that the raft association and function of Rsp5 are also affected by defects in ergosterol biosynthesis.

Other raft-associated proteins in the erg2Δ

The fact that the erg2 TRP1 mutant showed no growth defect suggested that ergosterol-depleted cells are not drastically defective in the membrane trafficking of other lipid raft-associated proteins. Indeed, in contrast to Tat2, Pma1 was also found on the plasma membrane in erg2Δ cells (Fig. 6). It is probable that other essential plasma membrane proteins, including Gas1, are also targeted to the plasma membrane. In addition, Pma1 was found in the lipid raft fraction in the erg2Δ mutant in a manner similar to the wild-type cells (Fig. 5). According to the raft model, sphingolipids are the main components of lipid rafts, while ergosterol/cholesterol is a subcomponent filling the space between sphingolipids (Harder & Simons, 1997; Simons & Ikonen, 1997). Sphingolipid depletion may entirely abrogate the structure of lipid rafts, causing cell death. In fact, the temperature-sensitive lcb1-100 mutant, which is defective in sphingolipid synthesis, is lethal at restrictive temperatures, and Pma1 and Gas1 are lost from the detergent-resistant fraction after transfer to restrictive temperatures (Bagnat et al., 2000). In contrast, ergosterol depletion by the deletion of ERG2 may partially compromise structure and function of lipid rafts. It is possible that sterol intermediates in erg2Δ cells substitute for ergosterol in lipid rafts, which allows targeting of Pma1 to the plasma membrane.

Other aspects

It was shown that nutrient uptake is specifically impeded in various viable erg mutant cells. The highly conserved membrane-trafficking system implies similar features in other organisms including humans. It is possible that cells with high cholesterol may guarantee high nutrient import, and drugs to inhibit human homologous enzymes of Erg6, Erg2, Erg3, Erg5, and Erg6 could specifically inhibit nutrient import in human cells.

It is plausible that there is a physiological system that regulates the sorting of nutrient transporters by altering sterol synthesis, without markedly affecting the functions of other plasma membrane proteins. In support of this notion, it was found that novel TOR-controlled transcriptional regulators, which are needed for the full expression of ERG genes, preferentially control the Golgi-to-plasma membrane sorting of Tat2 but not Pma1 (our unpublished data). It is suspected that TOR regulates Tat2 sorting via these factors, and further studies are now investigating this. This control seems to be different from TOR-regulated Tat2 sorting via protein kinase Npr1 proposed in an early study, although it is unknown why Npr1 regulates Tat2 ubiquitination (Schmidt et al., 1998). Similar regulatory processes may exist in other organisms including humans.


We thank Michael Hall, Howard Riezman, and Maria Cardenas for materials, and Reika Watanabe for helpful discussions. We especially thank the laboratory members of T.U. for helpful discussions. This work was supported in part by Grants-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Science, Sports, and Culture of Japan (grant no. 19370082). The research was partially carried out using an instrument at the Center for Instrumental Analysis of Shizuoka University.


  • Editor: Linda Bisson


View Abstract