A nonconserved Ala401 in the yeast Rsp5 ubiquitin ligase is involved in degradation of Gap1 permease and stress-induced abnormal proteins

Abstract

A toxic l-proline analogue, l-azetidine-2-carboxylic acid (AZC), causes misfolding of the proteins into which it is incorporated competitively with l-proline, thereby inhibiting the growth of the cells. AZC enters budding yeast Saccharomyces cerevisiae cells primarily through the general amino acid permease Gap1, not through the proline-specific permease Put4. We isolated an AZC-hypersensitive mutant that cannot grow even at low concentrations of AZC because of the accumulation of intracellular AZC. By screening through a yeast genomic library, the mutant was found to carry an allele of RSP5 encoding an E3 ubiquitin ligase. A single amino acid change replacing Ala (GCA) at position 401 with Glu (GAA) showed that Ala-401 in the third WW domain (a protein interaction module) is not conserved in the domain. The addition of to yeast cells growing on l-proline induced rapid ubiquitination, endocytosis, and vacuolar degradation of the plasma membrane protein Gap1. However, immunoblot and permease assays indicated that Gap1 in the rsp5 mutant remained stable and active on the plasma membrane probably with no ubiquitination, leading to AZC accumulation and hypersensitivity. The rsp5 mutants also showed hypersensitivity to various stresses (toxic amino acid analogues, high temperature in a rich medium, and oxidative treatments) and defects in spore growth. These results suggest that Rsp5 is involved in selective degradation of abnormal proteins and specific proteins for spore growth, in addition to nitrogen-regulated degradation of Gap1. Furthermore, Ala-401 of Rsp5 was considered to have an important role in the ubiquitination of targeted proteins.

Addition of some amino acid analogues can induce a transient physiological stress response in cells comparable to that of heat shock stress (1–3). Most analogues are transported into cells via amino acid permeases and cause misfolding of the proteins as they compete with naturally occurring amino acids. The accumulation of abnormal proteins, in turn, inhibits cell growth. Recently, Trotter and colleagues (4, 5) found that l-azetidine-2-carboxylic acid (AZC), a toxic four-membered ring analogue of l-proline, arrests proliferation in the G₁ phase of the cell cycle by the same mechanism as temperature up-shift. AZC is an unusual imino acid found only in several plants belonging to the Lilaceae family (6, 7), but can replace l-proline in proteins of bacteria and animal cells (8, 9), presumably in those of yeast cells (5). When AZC was added to cells of yeast Saccharomyces cerevisiae growing in minimal medium, cell viability gradually decreased, causing cell death (M. Nomura and H.T., unpublished work).

The accumulation of abnormal or misfolded proteins in cells under stress is a serious problem. To overcome it, the following two strategies can be considered: (i) to degrade the proteins through a ubiquitin-proteasome system or (ii) to refold the proteins by molecular chaperones. We isolated AZC-resistant mutants and strains and elucidated the mechanisms of AZC resistance. Large quantities of l-proline are believed to dilute AZC, its toxic competitor (10–13). We recently discovered the MPR1 gene and its homologues on the chromosome of yeast strains; these genes encode N-acetyltransferases that detoxify AZC by acetylating it (14–17). It is also possible that a recessive mutation occurs in the proline permease, the general amino acid permease encoded by GAP1, or the proline-specific permease encoded by PUT4 (18, 19); it may also occur in the membrane composition.

AZC is a valuable compound for analyzing cellular response to protein misfolding, because substituting AZC in place of l-proline in the protein will change the normal α-helical structure of a polypeptide to one having an angle of ≈15° smaller, which results in an altered tertiary structure of the protein (12). In this study, we isolated an AZC-hypersensitive mutant that cannot grow on medium containing low concentrations of AZC compared with its wild-type strain. We then identified the RSP5 gene encoding an E3 ubiquitin ligase involved in cell growth in the presence of AZC. The Rsp5 protein is known to ubiquitinate plasma membrane permeases followed by endocytosis and vacuolar degradation (20–22). In the Rsp5 sequence of this mutant, one amino acid substitution found in the nonconserved residue was shown to impair the specific interaction with the Gap1 permease. In addition, we propose a novel function of Rsp5 under various stresses that induce protein misfolding.

Materials and Methods

Strains and Plasmids. All yeasts used in this study were the S. cerevisiae strains with a S288C background. CKY8 (MATα ura3-52 leu2-3,112) and ERX-17C (MATa his4-N met4 ade8) were supplied by C. Kaiser (Massachusetts Institute of Technology, Boston). An AZC-hypersensitive mutant CHT81, which has a mutation in RSP5, was isolated from CKY8 after ethyl methanesulfonate treatment. The rsp5 mutant INV-A401rsp5 (MATa his3-Δ1 leu2 trp1-289 ura3-52 rsp5::TRP1 p366-A401rsp5[LEU2]) and its control strain INV-RSP5 (MATa his3-Δ1 leu2 trp1-289 ura3-52 rsp5::TRP1 p366-RSP5[LEU2]) were constructed from strain INVSc1 (MATa his3-Δ1 leu2 trp1-289 ura3-52) (Invitrogen). Escherichia coli strain DH5α [F^–λ^–Φ80lacZΔM15 Δ(lacZYA argF)U169 deoR recA1 endA1 hsdR17(r_k^–m_k⁺) supE44 thi-1 gyrA96] was used to subclone the yeast gene.

A yeast genomic library was purchased from American Type Culture Collection (ATCC 77162). This library was constructed by ligation of yeast genomic DNA partially digested with Sau3AI into the single BamHI site of YCp50-based p366. Centromere plasmid p366 (ATCC 77163) was derived from YCp50 containing CEN4 and LEU2 (11). Plasmid pCHT-1 containing RSP5 was isolated from a suppressor of CHT81 by introducing a yeast genomic library. Plasmids p366-RSP5 and p366-A401rsp5 contains the wild-type and the Ala401Glu mutant RSP5, respectively. Three centromere plasmids pRS414, pRS415, and pRS416 (Stratagene) harboring TRP1, LEU2, and URA3, respectively, were used for complementing the auxotrophic markers.

Culture Media. The media used for growth of S. cerevisiae were yeast extract/peptone/dextrose (YPD) (2% glucose, 1% yeast extract, 2% peptone) and SD (2% glucose, 0.67% Bacto Yeast Nitrogen Base without ammonium sulfate and amino acids; Difco). SD medium contained 10 mM (NH₄)₂SO₄ (SD+Am) or 10 mM l-proline (SD+Pro) as the sole nitrogen source. Yeast strains were also cultured on SD+Am agar plates containing AZC, l-canavanine (l-arginine analogue), O-f luoro-dl-phenylalanine (l-phenylalanine analogue), dl-β-hydroxynorvaline (l-valine analogue), dl-norleucine (l-methionine analogue), β-(2-thienyl)-dl-alanine (l-alanine analogue) or S-2-aminoethyl-l-cysteine (l-lysine analogue). All of the analogues were obtained from Sigma. The E. coli recombinant strains were grown in Luria–Bertani medium (23) containing ampicillin (50 μg/ml). If necessary, 2% agar was added to solidify the medium.

Gene Disruption. For GAP1 and PUT4 disruption, a DNA fragment containing URA3 and LEU2, respectively, was amplified by PCR performed with genomic DNA from strain CKY8 and oligonucleotide primers 5′-CGC TCT GGA TGA GAC ATA TAA AGA TGA AGG TGA AGT CCA CTG ATT CGG TAA TCT CCG AAC-3′, 5′-AAT CCC CAA AGT TCA ATA AGG AAA TGG GGA GTA AGG AGG CAT CAT TAC GTC CGA GAT TCC-3′, 5′-GCA CAG CAT ATC TCC ATC CAT GTA CTG ATA CAG ACG CAT AGT CCT GTA CTT CCT TGT TCA-3′, and 5′-TTA CCC CTG TTT CCT ATG CTG CGT CGC TAA ATA AAA CCG ATC GCG CGT TTC GGT GAT GAC-3′. The underlining indicates the sequence upstream of the initiation codon and downstream of the termination codon of GAP1 and PUT4, respectively. The unique 1.1- and 1.9-kb amplified band containing URA3 and LEU2, respectively, was purified and then integrated into the GAP1 and PUT4 locus, respectively, in strain CKY8 by transformation. The Ura⁺ and Leu⁺ phenotype was selected, and the correct disruption event was confirmed by using the chromosomal PCR.

Construction of the rsp5 Mutant. Aside from the mutant CHT81, we constructed another rsp5 mutant INV-A401rsp5 by disrupting an essential RSP5 gene on the chromosome of strain INVSc1 containing plasmid p366-A401rsp5. First, plasmid p366-RSP5 was constructed by blunt-end ligation of the 4.9-kb NcoI-AccIII fragment from pCHT-1 with the BamHI site of p366 and used as a template DNA for site-directed mutagenesis. The replacement of Ala-401 by Glu in Rsp5 was performed by PCR with oligonucleotide primers 5′-TAC GGA* ACG TGT ATA TTT CG-3′, 5′-TAT ACA CGT T*CC GTA TTG GT-3′, 5′-AAC TAC CTT CGT CAA GTC CG-3′, and 5′-ATC TTT CAT CGT CGG CAC TG-3′. The asterisks show the locations of mismatches. The unique amplified band of 1.4-kb was digested with PvuII and AvaII to recover the 1.1-kb fragment and ligated to the 12.3-kb fragment of plasmid p366-RSP5 digested with PvuII and AvaII. The mutation in the resultant plasmid p366-A401rsp5 was confirmed by DNA sequencing.

To disrupt RSP5 on the chromosome, a DNA fragment containing TRP1 was amplified by PCR performed with genomic DNA from strain INVSc1 and oligonucleotide primers 5′-ATG CCT TCA TCC ATA TCC GTC AAG TTA GTG GCT GCA GAG TTA CTA TTA GCT GAA TTG CCA-3′ and 5′-TCA TTC TTG ACC AAA CCC TAT GGT TTC TTC CAC GGC CAA TGG TTT TTC GCC CTT TGA CGT-3′. The underlining indicates the sequence upstream of the initiation codon and downstream of the termination codon of RSP5. The unique 1.5-kb amplified band containing RSP5 was purified and then integrated into the RSP5 locus in strain INVSc1 containing p366-A401rsp5 and p366-RSP5 to construct INV-A401rsp5 and INV-RSP5, respectively, by transformation. The Trp⁺ phenotype was selected, and the correct disruption event was confirmed by using the chromosomal PCR.

Northern Blot Analysis. Northern blot analysis was carried out by using a Gene Images Random-Prime Labeling and Detection system (Amersham Pharmacia). Total RNA from S. cerevisiae was isolated by the method of Köhrer and Domdey (24). RNA was separated in 1.0% agarose gel and transferred to nylon membrane. As a DNA probe, the DNA fragments of GAP1, PUT4, and ACT1 were prepared by PCR with oligonucleotide primers 5′-TTG ACG AAA CAG GTT CAG GG-3′, 5′-TGC TGG GAT GAA AAG CTT CC-3′, 5′-GAT TAT GGA CGT GGA CTT GG-3′, 5′-AAT GAG AGA GAA CCA CAC GG-3′, 5′-CGG AAT TCC TCT CCC ATA ACC TCC TA-3′, and 5′-CGG GAT CCG GGC TCT GAA TCT TTC GT-3′, respectively. Each unique amplified band was purified from agarose gel, denatured, and labeled according to the protocol recommended by the supplier.

Western Blot Analysis. Yeast cells were cultured to the exponential growth phase (OD₆₀₀ of 1.0) in SD+Pro medium, and the plasma membrane-enriched fractions before addition of 10 mM (NH₄)₂SO₄ and at intervals thereafter were prepared as described by Galan et al. (22). The cells were harvested and washed, and the whole-cell extracts were prepared by vortexing the cells with glass beads. The membrane-enriched fraction was collected by centrifugation for 45 min at 15,000 × g, suspended in 0.1 M Tris·HCl buffer (pH 7.5) containing 0.15 M NaCl, 5 mM EDTA, and 5 M urea, kept on ice for 30 min, and sedimented as above. The resulting pellets were resuspended in 0.1 M Tris·HCl buffer (pH 7.5) containing 0.15 M NaCl, 5 mM EDTA, and 2% SDS, and boiled for 10 min.

For Western blot analysis, the supernatant (30 μg of solubilized proteins) after centrifugation (10 min at 15,000 × g) was loaded on a 6% SDS-polyacrylamide gel in a Tricine system (25). For the experiment showing the shift of the ubiquitinated form of Gap1, SDS/PAGE was carried out on a 12.5% gel in Laemmli's system (26). Gap1 was detected by using a ECL Plus Western Blotting Detection System (Amersham Pharmacia) and anti-Gap1p polyclonal antibody at 1:30,000 dilution (a gift from B. André, Université Libre de Bruxelles, Brussels). The plasma membrane H⁺-ATPase Pma1 was used as an internal control. Protein concentrations were determined with a BCA Protein Assay Reagent Kit (Pierce).

Proline Permease Assays. Yeast cells were cultured to the exponential growth phase (OD₆₀₀ of 1.0) in SD+Pro medium and proline permease activities were measured by incorporation of ¹⁴C-radiolabeled l-citrulline and l-proline (New England Nuclear) before addition of 10 mM (NH₄)₂SO₄ and at intervals thereafter. The cells were harvested, washed, and suspended in 10 mM citric acid buffer (pH 4.5) containing 2% glucose to OD₆₀₀ of 2.0. One milliliter of cell suspension was incubated with 89 μM l-[¹⁴C]citrulline (56 mCi/mmol; 1 Ci = 37 GBq) for Gap1 or with 22 μM l-[¹⁴C]proline (229 mCi/mmol) for Put4 for 5 min at 25°C. The cells were then collected on a 0.8-μm nitrocellulose filter (Whatman) and washed twice with ice-cold water. The filters were dried, and intracellular labeled amino acid was determined by liquid scintillation counting. Permease activities were calculated as nmol per min per mg dry weight.

Intracellular Content of AZC. Yeast cells were cultured in SD+Am medium containing AZC, and 5 ml of cell suspension (≈5 × 10⁸ cells) were removed, washed twice with 0.9% NaCl, and suspended in 0.5 ml of distilled water. The 1.5-ml microcentrifuge tube containing cells was transferred to a boiling water bath, and intracellular amino acids were extracted by boiling for 10 min. After centrifugation (5 min at 15,000 × g), each supernatant was subsequently quantitated with an amino acid analyzer (l-8500A, Hitachi). The AZC content was expressed as a percentage of the dry weight.

Results

The Toxic AZC Enters Yeast Cells Primarily Through Gap1. l-Proline enters yeast cells via two transporters, the Gap1 general amino acid permease and the Put4 proline-specific permease. The gap1-and put4-disrupted strain failed to grow on a SD+Pro agar plate, whereas the wild-type and the single-gene disrupted strains used l-proline as the sole nitrogen source (Fig. 1). We examined which of the two permeases was mainly involved in AZC uptake. Fig. 1 shows that the gap1-disrupted strain grew on the AZC-containing medium and accumulated smaller amounts of intracellular AZC than the wild-type and the put4-disrupted strains. These results indicate that toxic AZC is transported into cells primarily through Gap1.

Fig. 1.

AZC was mainly transported into yeast cells via Gap1. The growth phenotype and AZC content of each strain were examined. Approximately 10⁶ cells of each strain and serial dilutions of 10^–1 to 10^–4 (from left to right) were spotted onto SD+Pro medium or SD+Am plus AZC medium. The plates were incubated at 30°C for 3 days. The intracellular AZC content was measured 2 h after addition of 5 mM AZC in SD+Am liquid medium. The values are means from three independent experiments. Variations in the values are <10%.

Isolation and Analysis of the AZC-Hypersensitive Mutant. To analyze the gene involved in growth when yeast cells are exposed to AZC, we isolated mutant CHT81, which is incapable of growing on SD+Am medium containing a low concentration of 0.1 mM AZC (10 μg/ml). This low level of AZC does not inhibit growth of the parent strain CKY8 (Fig. 2A). In this mutant, the intracellular content of AZC at 2 h after addition of 5 mM AZC was much higher than that of the wild-type strain (1.74% vs. 0.40% of dry weight), probably because a larger amount of AZC had been transported into cells. A heterozygous diploid (CHT81 × ERX-17C) crossed with a wild-type strain of the opposite mating type (ERX-17C) grew on 0.1 mM AZC-containing medium, indicating that the AZC-hypersensitive growth phenotype of the mutant was recessive (Fig. 2 A). Although sporulation of diploid cells was detectable, no ascospore from the diploid grew on YPD plates.

Fig. 2.

The recessive mutation of RSP5 enhanced sensitivity to AZC. (A) Various strains were cultivated on SD+Am medium containing 0.1 mM AZC at 30°C for 3 days. (B) Sequence alignment of Rsp5 WW domains. WW1, amino acids 229–266. WW2, amino acids 331–368. WW3, amino acids 387–424. The filled circles at the top indicate the highly conserved residues of the WW domain. The arrow at the bottom shows the mutation found in this study.

The AZC-Hypersensitive Mutant Carried an Allele of the RSP5 Gene. We then transformed the mutant CHT81 with a yeast genomic library and incubated the transformants on Leu^– agar plates containing 0.1 mM AZC at 30°C for several days. One plasmid (pCHT-1) was recovered from the only Leu⁺ transformant that was resistant to AZC. When pCHT-1 was used to retransform strain CHT81, all of the Leu⁺ transformants were found to grow on the AZC-containing medium, suggesting that the mutated gene resided in the plasmid (Fig. 2 A).

The nucleotide sequence of the ends of the inserted DNA showed that pCHT-1 had an ≈12-kb fragment on chromosome V. Subcloning of this region indicated that the gene complement to the AZC-hypersensitivity was RSP5, which encodes an E3 ubiquitin ligase. In the Rsp5 sequence of mutant CHT81, we found a single-base change from C to A at position 1,202, leading to replacement of alanine with glutamate at position 401 within the WW3 domain (Fig. 2B). It is noteworthy that Ala-401 is a nonconserved residue among the domains and that its function is unknown. To examine the effect of one amino acid substitution (Ala401Glu) in the Rsp5 protein on AZC-hypersensitivity, we constructed another rsp5 mutant INV-A401rsp5 as described in Materials and Methods. In agreement with the results obtained with mutant CHT81, AZC hypersensitivity in strain INV-401rsp5 was clearly evident (Fig. 2 A). These results suggest that an Ala401Glu mutation in Rsp5 impairs the specific interaction with the targeted protein.

Gap1 in the Ala401Glu rsp5 Mutant Remains Stable and Active on the Membrane. Because AZC is mainly incorporated into the cell via Gap1 (Fig. 1), we examined the effect of Rsp5 on GAP1 expression and Gap1 activity. Northern blot analysis showed quantitatively similar amounts of GAP1 transcripts in strains CKY8 and CHT81 in the presence or absence of (NH₄)₂SO₄ (Fig. 3A). Springael and André reported that the addition of ammonium ions to yeast cells growing on l-proline as the sole nitrogen source induces rapid inactivation and degradation of Gap1 through a process requiring Rsp5 (27). They showed that ammonium ions cause internalization of Gap1 by endocytosis, delivering it into the vacuole for degradation. To determine the amount and the permease activity of Gap1 in the rsp5 mutant, we carried out Western blot analysis using anti-Gap1 antibody and permease assays using ¹⁴C-labeled amino acids. As shown in Fig. 3B, the wild-type Gap1 was removed from the plasma membrane and its permease activity gradually decreased after the addition of ammonium ions, as occurred in the previous report (27). However, Gap1 in the rsp5 mutant CHT81 was shown to remain stable and active on the plasma membrane with little degradation.

Fig. 3.

Gap1 in the rsp5 mutant remained active without ammonium-induced ubiquitination on the plasma membrane. (A) Total RNA (≈50 μg) from strains CKY8 and CHT81 was hybridized to GAP1 and ACT1 probes. Cells were grown on SD+Pro medium before (time = 0) and 60 min after addition of (NH₄)₂SO₄. The GAP1 and ACT1 probes detected the 2.2-kb and the 1.4-kb transcripts, respectively. (B) Immunoblot and permease activity of Gap1 in the membrane-enriched extracts prepared before (time = 0) and at several times after addition of (NH₄)₂SO₄. Gap1 and Put4 activity was measured by incorporation of ¹⁴C-citrulline and ¹⁴C-proline, respectively. To determine the Put4 activity, we used the gap1-disrupted strains derived form CKY8 and CHT81. Each activity before addition of (NH₄)₂SO₄ is given as 100%. (Lower) Levels of the plasma membrane H⁺-ATPase Pma1 as a protein-loading control. (C) Immunoblot of Gap1 from the membrane-enriched extracts prepared before (time = 0) and at several times after addition of (NH₄)₂SO₄. The positions of ubiquitinated forms are indicated with dots.

We further examined the effect of Rsp5 on Put4 activity measured by incorporation of [¹⁴C]proline. To eliminate the influence of Gap1 on l-proline uptake, the gap1-disrupted strains were constructed from strains CKY8 and CHT81. There was no significant change in Put4 activity between two strains, even after the addition of ammonium ions (Fig. 3B). This result suggests that Put4 is not a targeted protein of Rsp5.

Gap1 in the Ala401Glu rsp5 Mutant Might Be Not Ubiquitinated in the Presence of . Ubiquitin is generally believed to be used as a signal for the endocytosis of some cell surface proteins and their subsequent degradation in the lysosome and vacuole (28). Therefore, ubiquitination of Gap1 was tested by Western blot analysis of membrane-enriched extracts (Fig. 3C). In the wild-type strain CKY8, two minor bands were seen just above the major Gap1 signal at 5 and 10 min after the addition of ammonium ions. Although we did not detect ubiquitin directly, these signals were assumed to be the ubiquitin-conjugated forms of Gap1 judging from the previous report (27). However, it seems likely that no apparent ubiquitination occurred in the rsp5 mutant. These results suggest that the mutant Rsp5 fails to recognize and to ubiquitinate Gap1 even in the presence of ammonium ions.

The Ala401Glu rsp5 Mutant Shows Hypersensitivity to Various Stresses. Regarding the Rsp5 function, in addition to ubiquitination of various permeases and transcriptional regulators (20–22, 29), it is known that a mutant has increased sensitivity to stresses such as cadmium and l-canavanine (the toxic l-arginine analogue) even at permissive temperatures (30). We therefore tested the growth phenotype of the rsp5 mutants CHT81 and INV-A401rsp5 with various stresses that induce protein misfolding in the cell (Fig. 4). When yeast cells were exposed to stresses such as toxic amino acid analogues (l-canavanine and O-fluoro-dl-phenylalanine), high growth temperature (37°C or 39°C) combined with a YPD medium, and oxidative stresses including hydrogen peroxide and heat shock treatments, the rsp5 mutants were found to be much more sensitive to these stresses than the wild-type strains CKY8 and INV-RSP5. Similar results were obtained in the cases of other toxic analogues [dl-β-hydroxynorvaline (2 mg/ml), dl-norleucine (2 mg/ml), β-(2-thienyl)- dl-alanine (10 μg/ml), and S-2-aminoethyl-l-cysteine (10 μg/ml)], and ethanol treatment (8% vol/vol) (data not shown). It is interesting that temperature sensitivity of the mutant was observed only on a rich complex YPD medium, not on a synthetic minimal SD+Am medium. These results suggest that Rsp5 is required for selective degradation of abnormal proteins generated by these stresses and that Ala at position 401 is important for its function.

Fig. 4.

The rsp5 mutants showed hypersensitivity to stresses that induce protein misfolding. The growth phenotypes of various strains were examined. Approximately 10⁶ cells of each strain and serial dilutions of 10^–1 to 10^–4 (from left to right) were spotted and incubated onto SD+Am medium containing the toxic analogue at 30°C for 3 days (A), YPD or SD+Am medium at the indicated temperature for 3 days (B), and the H₂O₂-containing SD medium or SD medium after heat shock treatment (50°C, 6h) at 30°C for 3 days (C).

Discussion

Gap1 Permease Is a Primary Transporter for AZC. We found that the toxic AZC enters yeast cells primarily via Gap1. Gap1 permease can transport all of the naturally occurring l-amino acids found in proteins and related compounds, such as ornithine and citrulline, and several d-amino acids and toxic amino acid analogues (31). It is reasonable that gap1-disrupted strain was less sensitive to AZC than the wild-type and put4-disrupted strains (Fig. 1). However, intracellular AZC still remained even in the gap1 and put4 double disruptant (0.14% of dry weight). It was postulated that AZC is in part transported through another permease(s), that take up l-proline, such as the structural analogue of l-proline γ-aminobutyric acid-specific permease Uga4 (32) or through unidentified permease(s) (33), whose function is similar to that of Gap1 or Put4.

Rsp5 Causes Degradation of Gap1, but Put4 Is Not a Target of Rsp5. Gap1 permease is inactivated by the addition of a rich nitrogen source (ammonia) to cells growing on the poor nitrogen source (27). Regarding the transcriptional regulation of GAP1 in the S. cerevisiae strain with a S288C background, GAP1 was found to express at high levels caused by the Nil1 transcriptional activator when ammonia was used as a nitrogen source (34). Analysis of GAP1 expression levels showed that the decrease in Gap1 activity in the wild-type strain CKY8 was not caused by a decrease in GAP1 transcription and that the GAP1 transcription of the rsp5 mutant CHT81 is not affected by the addition of ammonium ions (Fig. 3). These results indicate that Gap1 activity is regulated at the level of posttranslation. Helliwell et al. (35) recently found that overexpression of either Bul1 or Bul2, which has been identified as a protein that binds to Rsp5 (36. 37), could confer AZC resistance caused by the reduction of Gap1 and Put4 activity. In contrast, the bul1 and bul2 double disruptant was hypersensitive to AZC, because the activity of both permeases may be increased (35). However, in the wild-type strain we used, it is unlikely that Put4 is down-controlled by Rsp5 activity after addition of ammonium ions (Fig. 3B).

Ala-401 in the WW3 Domain Is Essential for in Vivo Function of Rsp5. RSP5 is an essential gene in S. cerevisiae; Rsp5 activity was recently shown to affect a variety of cellular events regulated by ubiquitination, including an early stage of cell surface proteins endocytosis [Gap1 (27), Fur4 (22), Tat2 (38), Gal2 (39), Zrt1 (40), and Ste2 (41)], mitochondrial inheritance (42), activation of the Spt23 transcription factor (29), degradation of the large subunit of RNA polymerase II (43), and actin cytoskeleton dynamics (44). The amino acid sequence of Rsp5 reveals several domains: a C2 domain, which may bind membrane phospholipids in a calcium-regulated manner, three WW domains (WW1, WW2, and WW3), which are evolutionally conserved protein interaction modules to determine the substrate specificity and are composed of 40 amino acids each that constitutes a hydrophobic pocket for binding proline-rich ligands (PXY and PPXY motifs) in substrate proteins, and a HECT domain (homologous to E6-AP carboxyl terminus), which catalytically ligates ubiquitin in the process of target proteins degradation (45). WW domains are presumably located at cellular membranes and fold into three antiparallel β-sheets (46). In particular, the WW3 domain (amino acids 387–424) and the WW2 domain are required for the essential in vivo function of Rsp5, and preferentially bind to peptides of consensus sequence (A/P)PPPYE, as determined by phage-display experiments (43, 47). Ala-401 in the WW3 domain is a nonconserved residue and is situated in turn-like or polar region (48). The mutation at position 401 in Rsp5 is novel because each of the point mutations in conserved residues, including two Trp residues, two Tyr residues, one Phe residue, and one Pro residue (see asterisks in Fig. 2B), abolished binding of WW domains to proline-rich peptides or resulted in temperature-sensitive growth defects (49–51). Our results suggest that position 401 in Rsp5 is important for ligand binding, leading to subsequent ubiquitination of the targeted protein. Replacement of Ala-401 by the large polar residue Glu may cause structural changes because of steric hindrance or may impair the electrical environment with the adjacent residue Arg-402 in the ligand-binding pocket. In the sequence of Gap1, a PXY motif (Pro-201–Lys-202–Tyr-203) is present in a small hydrophilic region based on the hydropathy profile (19), as a potential ligand for Rsp5. It should be possible that the Gap1 PXY motif interacts with the WW3 domain of Rsp5 without binding to that of an Ala401Glu variant.

Rsp5 Is Considered to Participate in Selective Degradation of Stress-Induced Abnormal Proteins. Our results also suggest a novel function of Rsp5 involved in selective degradation of stress-induced abnormal proteins. Yeast strains carrying the rsp5 mutation showed enhanced sensitivity to stresses, such as addition of the amino acid analogues in a minimal medium, H₂O₂ and transient heat shock treatments, or a combination of a rich growth medium (YPD) and elevated temperature (37–39°C), by accumulating abnormal misfolded cell proteins. The growth phenotype of the rsp5 mutants at elevated temperatures is not a result of the instability of Rsp5, because they proved not to be temperature-sensitive in either a SD medium (Fig. 4) or a synthetic complete medium (data not shown).

It should be noted that hypersensitivity of the rsp5 mutants to toxic amino acid analogues is attributable to two independent defects: one being the ammonia-induced inactivation of Gap1, which presumably uptakes the analogues (Fig. 3), and the other the rapid degradation of misfolded proteins into which the analogues are incorporated. We further examined the effect of GAP1 (data not shown). The gap1-disrupted mutant was still hypersensitive to these analogues, suggesting that the rsp5 mutation also affects degradation of misfolded proteins. These observations may support a role for Rsp5 in metabolism of misfolded and abnormal proteins.

Oxidative stress during exposure to H₂O₂ and to heat shock produces oxygen-free radicals and other reactive oxygen species that could attack vulnerable proteins containing iron sulfur centers (52). Ethanol can also disrupt protein folding by a mechanism distinct from that of AZC (5). In general, an appropriate rate of protein synthesis is essential to produce properly folded proteins. At high temperatures in a rich growth medium, it seems likely that the high rate of protein synthesis disturbs the correct folding of nascent polypeptides and results in the accumulation of denatured proteins. In S. cerevisiae, E2 ubiquitin-conjugating enzymes Ubc4 and Ubc5, the gene expression of which is heat inducible, were known to mediate selective degradation of l-canavanine-derived short-lived and abnormal proteins (4, 53). Loss of Ubc4 and Ubc5 activity impairs cell growth and leads to inviability at elevated temperatures or in the presence of l-canavanine. It is also a matter of course that the yeast cytosolic proteasome is deeply involved in stress-induced proteolysis in which l-canavanineincorporated proteins are ubiquitinated and degraded for cell survival (54). Our findings suggest that stress-induced abnormal proteins are selectively degraded through a process requiring Rsp5 E3 ubiquitin ligase, but we do not yet demonstrate this interesting event in vivo. Trotter et al. (5) have recently reported that misfolded proteins mediate heat shock activation of heat shock factor and its target genes. To obtain the evidence for our hypothesis, we now investigate whether the rsp5 mutant results in its stronger or more prolonged activation in response to heat shock or ethanol treatment.

Spore Growth May also Require the Rsp5-Mediated Degradation. Kanda (30) reported that an rsp5-uby mutant had defects in sporulation of the diploid, suggesting that Rsp5 participates in sporulation via degradation of a specific group of proteins. The heterozygous RSP5/rsp5 diploid sporulated, but none of the ascospore grew on YPD plates (data not shown). This result shows that Ala-401 in the Rsp5 protein is important for spore growth, although the sporulation process is normal. One possible mechanism is that Rsp5 participates in spore growth through degradation of specific or nonspecific protein(s) and through supplying amino acids for new protein synthesis in the cells.

In summary, it is suggested that Rsp5 participates in selective degradation of stress-induced abnormal proteins and specific proteins for spore growth in addition to nitrogen-regulated degradation of Gap1 permease (Fig. 5). In this process, Ala-401 of Rsp5 plays a significant role in ubiquitination of targeted proteins. We must further analyze the rsp5 mutants isolated here to investigate the mechanism. This approach could be a useful method for studying stress responses and for breeding novel stress-resistant yeast strains.

Fig. 5.

Proposed model for Rsp5-regulated proteolysis. Rsp5 is believed to ubiquitinate Gap1 in an ammonium-induced manner followed by endocytosis and vacuolar degradation (27). Various stress conditions probably lead to accumulate abnormal cellular proteins. The addition of amino acid analogues is believed to cause misfolding of proteins into which it is incorporated. Exposure to H₂O₂, heat shock, and ethanol denatures vulnerable proteins (5, 52). Cell growth by a combination of a rich medium and elevated temperature may disturb the correct folding of newly synthesized proteins. Our results suggest that Rsp5 is involved in degradation of the targeted proteins in these cellular events through ubiquitination. Yeast strains carrying the rsp5 mutation show hypersensitivity to these stresses and defects in spore growth, presumably because of failure of ubiquitination-triggered degradation of these proteins.