Organelle-localized NHEs (Na+/H+ exchangers) are found in cells from yeast to humans and contribute to organellar pH regulation by exporting H+ from the lumen to the cytosol coupled to an H+ gradient established by vacuolar H+-ATPase. The mechanisms underlying the regulation of organellar NHEs are largely unknown. In the present study, a yeast two-hybrid assay identified Mth1p as a new binding protein for Nhx1p, an organellar NHE in Saccharomyces cerevisiae. It was shown by an in vitro pull-down assay that Mth1p bound to the hydrophilic C-terminal half of Nhx1p, especially to the central portion of this region. Mth1p is known to bind to the cytoplasmic domain of the glucose sensor Snf3p/Rgt2p and also functions as a negative transcriptional regulator. Mth1p was expressed in cells grown in a medium containing galactose, but was lost (possibly degraded) when cells were grown in medium containing glucose as the sole carbon source. Deletion of the MTH1 gene increased cell growth compared with the wild-type when cells were grown in a medium containing galactose and with hygromycin or at an acidic pH. This resistance to hygromycin or acidic conditions was not observed for cells grown with glucose as the sole carbon source. Gene knockout of NHX1 increased the sensitivity to hygromycin and acidic pH. The increased resistance to hygromycin was reproduced by truncation of the Mth1p-binding region in Nhx1p. These results implicate Mth1p as a novel regulator of Nhx1p that responds to specific extracellular carbon sources.
In all eukaryotic cells, the lumina of organelles, such as the Golgi, endosomes, secretory vesicles and lysosomes (vacuoles in fungi and plants) are maintained at distinct acidic pHs. These distinct pH values are important for a variety of physiological processes, such as intracellular transport of small molecules [1,2], activation of lysosomal (vacuolar) hydrolases, post-translational modifications and processing of secreted proteins , and trafficking of endocytosed ligands or newly synthesized proteins to their functional sites .
V-ATPase (vacuolar H+-ATPase) is an ATP-driven proton pump that is found in most organellar membranes and plays an essential role in the acidification of the organelle . However, V-ATPase alone does not account for the differences in pH among various organelles. One possibility is that an organelle-specific proton leakage system regulates luminal pH. In fact, the addition of a V-ATPase inhibitor, such as bafilomycin or concanamycin, results in immediate alkalization of organelles [6,7], supporting the presence of a proton leak system across the organelle membranes. One candidate for this leakage system is an organellar-type NHE (Na+/H+ exchanger).
NHEs are ubiquitous proteins in the membranes of cells of various species from yeast to humans and higher plants . The NHE proteins are predicted to have 12 transmembrane segments and a hydrophilic tail region [9–16]. The NHEs exchange Na+ (or K+) for H+ across the cell membranes and play an important physiological role in the regulation of intracellular pH and Na+ concentration [9–15]. Nine isoforms (NHE1–NHE9) have been described in mammalian cells. NHE1–NHE5 are mainly localized to the plasma membranes [9–13] and NHE6–NHE9 are distributed to the organellar membranes: NHE6 and NHE9 are in early/recycling endosomes, NHE7 is in the trans-Golgi network and NHE8 is in the mid-Golgi stacks [14–16]. We found previously that organelles in mammalian cells transiently overexpressing NHE8 or NHE9 have higher luminal pH than control cells . Knockdown or overexpression of NHE6 causes a decrease or increase of the endosomal pH respectively . Knockdown of the gene encoding a NHE6-interacting protein, RACK1 (receptor for activated C-kinase 1), increases endosomal NHE6 levels and results in alkalization of endosomes . These findings suggest that mammalian organelle-type NHEs contribute to organellar pH homoeostasis by a proton-leak system from the organellar lumen in co-operation with the V-ATPase.
The budding yeast Saccharomyces cerevisiae also has an organelle-type NHE, Nhx1p. Nhx1p has a primary sequence similar to mammalian NHE6 and is localized in the late endosomes (pre-vacuolar compartments) [19,20]. Yeast cells lacking the NHX1 gene show impaired cell growth and acidification of their vacuolar lumen relative to the wild-type cells when the cells are grown in medium at an acidic pH (<4) . These findings suggest that yeast Nhx1p also contributes to the regulation of the luminal pH of the organelles, and that Nhx1p is a functional yeast orthologue of mammalian NHE6.
Organellar pH is always influenced by changes in intracellular metabolite levels, which in turn depend on carbon sources in the culture medium and uptake of extracellular solutes through endocytosis . Living yeast cells are believed to have mechanisms to maintain their pH homoeostasis against such changes. In this regard, it has been shown that V-ATPase activity is tightly regulated, depending on the carbon source in the culture medium [23,24]. One mechanism regulating the V-ATPase activity is reversible dissociation of its peripheral-membrane V1 subunit from its membrane-integrated Vo subunit, depending on the presence of carbon sources such as glucose or galactose [23–25]. This mechanism is a rapid and effective way to reduce or increase proton pump activity. Unlike V-ATPase, the dependence of the regulatory mechanisms for Nhx1p on carbon sources in the culture medium has not been investigated.
We have shown that the mammalian organellar-type NHE6 is regulated by binding of the scaffold protein RACK1 to the hydrophilic tail region . Thus this hydrophilic tail region is believed to play an important role in its regulation [26–30]. On the basis of this finding, we hypothesized that a putative protein interacts with the hydrophilic tail region of yeast Nhx1p to regulate its activity. By using a yeast two-hybrid technique, we found a new binding partner for Nhx1p. This protein, Mth1p, has been described as a negative transcriptional regulator of an extracellular glucose-sensing mechanism . In the present study we show that Mth1p regulates Nhx1p activity, probably by modulating its ion transport activity in response to extracellular carbon sources. This regulation of Nhx1p activity by Mth1p may contribute to maintain the luminal pH of endosomes regardless of changes in extracellular carbon sources.
Strains and culture conditions
Yeast S. cerevisiae strains MKY05121, MKY0813 and MKY0814, bearing deletions of NHX1, MTH1, and both NHX1 and MTH1 respectively, were derived from W303–1B (MATa ade2-1 trp1-1 can1-100 leu2-3/112 his3-11/15 ura3–1; ). Deletion of NHX1 or MTH1 was performed by one-step gene replacement . The LEU2 or Schizosaccharomyces pombe his5 gene was amplified by PCR from pUG73 or pUG27  using the following primer sets respectively: forward #1 (5′-ATGCTATCCAAGGTATTGCTGAATATAGCTTTCAAGGTGCCAGCTGAAGCTTCGTACGC-3′) and reverse #1 (5′-CTAGTGGTTTTGGGAAGAGAAATCTGCAGGTGATTGCGTAGCATAGGCCACTAGTGGATCTG-3′) for NHX1 knockout; and forward #2 (5′-GAATTTTATTCGAACGCATAGAGTACACACACTCAAAGGACAGCTGAAGCTTCGTACGC-3′) and reverse #2 (5′-TCTCCAAAAAAACCATCGGGAAGGTTTCTTTTTAGTATCTGCATAGGCCACTAGTGGATCTG-3′) for MTH1 knockout. The PCR products were introduced into yeast cells and the resulting Leu+ or His+ transformants were selected on synthetic medium lacking leucine or histidine. Gene knockout was confirmed by PCR amplification of the genomic DNA. Standard yeast culture medium and genetic manipulations were as described by Sherman . Transformation of yeast cells was performed by the lithium acetate method . All yeast strains were routinely cultured at 30 °C in YPAD medium [1% (w/v) yeast extract, 2% (w/v) peptone, 40 mg/l adenine and 2% (w/v) glucose], SD medium [0.17% yeast nitrogen base without ammonium sulfate or amino acids, 0.5% ammonium sulfate and 2% (w/v) glucose] or SCD medium [0.17% yeast nitrogen base without ammonium sulfate or amino acids, 0.5% ammonium sulfate, 0.5% casamino acids and 2% (w/v) glucose] supplemented with appropriate nucleotides and amino acids. SGal and SCGal medium have 2% (w/v) galactose to replace the glucose in SD and SCD medium. SRaf and SGly/Eth medium have 2% (w/v) raffinose, and 3% (v/v) glycerol and 2% (v/v) ethanol respectively to replace the glucose in SD and SCD medium. Escherichia coli strains JM109 and BL21(DE3) were used to propagate the plasmids and to express various proteins. E. coli cells were cultured in lysogeny broth with an antibiotic appropriate for selection of transformants, as described previously .
Plasmid construction for expression of recombinant peptides in E. coli cells
For expression of recombinant GST (glutathione transferase)–NHX1Ct (full length), GST–NHX1Ct (489–560), GST–NHX1Ct (561–633), GST–NHX1Ct (524–604) or MBP (maltose-binding protein)–MTH1 proteins, the corresponding DNA fragments were amplified by PCR from yeast genomic DNA using the following primers: forward #3 (5′-TTCGGATCCAAGACTGGTTGCATAAGTGAAG-3′) and reverse #3 (5′-GCAGGTCGACCTAGTGGTTTTGGGAAGAGAAATC-3′), forward #3 and reverse #4 (5′-AGACGTCGACCTATCCAAAAGTATTACCACCAG-3′), forward #4 (5′-CCGGATCCGGCCTTAATGAAACTGAGAA-3′) and reverse #3, forward #5 (5′-CCGGATCCGGCCCATATTCTGACAACAA-3′) and reverse #5 (5′-AGACGTCGACCTATACCTGTTCATCAAAATTTTG-3′), forward #6 (5′-CGGGATCCTTTGTTTCACCACCACCAGC-3′) and reverse #6 (5′-TGCACTGCAGTCAGGATACTGAATCCGGCTGCC-3′). The resulting PCR products were cloned into pGEX4T-2 (GE Healthcare) or pMALcri (New England Biolabs) using the BamHI/SalI or BamHI/PstI sites respectively.
Plasmid construction for expression of recombinant peptides in yeast cells
To obtain plasmids encoding GAL4-BD (binding domain)–NHX1Ct, GAL4-BD–NHX1Ct (489–560), GAL4-BD–NHX1Ct (561–633), GAL4-BD–NHX1Ct (524–604) and GAL4-AD (activation domain)–MTH1 for the yeast two-hybrid assay, corresponding DNA fragments were amplified by PCR from yeast genomic DNA using the following primers: forward #7 (5′-CGGAATTCAAGACTGGTTGCATAAGTG-3′) and reverse #3, forward #7 and reverse #4, forward #8 (5′-CGGAATTCGGCCTTAATGAAACTGAGAA-3′) and reverse #3, forward #9 (5′-CGGAATTCGGCCTTAATGAAACTGAGAA-3′) and reverse #5, and forward #6 and reverse #6 respectively. The PCR product was cloned into pGBT9 or pGAD424 (Clontech) using the EcoRI/SalI or BamHI/PstI sites. For expression of wild-type Nhx1p–GFP (green fluorescent protein) and Nhx1p–GFP mutants with the truncation of the C-terminus (Δ604, Δ560, Δ523 and Δ488), the corresponding DNA fragments were amplified by PCR from yeast genomic DNA using the following primers: forward #10 (5′-GACGGTACCATGCTATCCAAGGTATTG-3′) and reverse #7 (5′-AGCGCATGCCGTGGTTTTGGGAAGAGAAATC-3′), forward #10 and reverse #8 (5′-ACATGCATGCCTACCTGTTCATCAAAATTTTG-3′), forward #10 and reverse #9 (5′-ACATGCATGCCTCCAAAAGTATTACCACCAG-3′), forward #10 and reverse #10 (5′-ACATGCATGCCCAAATCTGTCTGAATAGAAC-3′), and forward #10 and reverse #11 (5′-AGCGCATGCCAATATTTAAAACTTCTAACA-3′). The resulting DNA fragments were cloned into pRS314-NHA1–GFP  using the KpnI/SphI sites.
Yeast two-hybrid screening
GAL4-based yeast two-hybrid screening was performed using a peptide from the Nhx1p C-terminus (residues 489–633) as bait. A plasmid encoding GAL4-BD–NHX1Ct was introduced into yeast strain AH109 (MATa, trp1-901, leu2-3, 112, ura3-52, his3-200, gal4Δ, gal80Δ, LYS2::GAL1UAS-GAL1TATA-HIS3, GAL2UAS-GAL2TATA-ADE2, ura3::MEL1UAS-MEL1TATA-lacZ; Clontech) and used for screening with the S. cerevisiae-S288C (Log) cDNA library (Nojima cDNA Library Collection No.35, kindly provided by Dr Hiroshi Nojima, Osaka University, Japan) cloned into the GAL4-AD fusion expression vector pAD3. Independent Leu+ Trp+ transformants (3.4×106) were screened on SD medium lacking adenine, histidine, leucine and tryptophan for selection. A total of 94 positive clones were isolated and then retested for interaction on selection plates. The cDNA inserts from a plasmid obtained from the positive clones were amplified by PCR, and the PCR products were directly subjected to DNA sequencing.
Expression and purification of GST and MBP fusion proteins
E. coli BL21(DE3) cells transformed with an expression plasmid were induced to express the fusion proteins with 0.4 mM IPTG (isopropyl β-D-thiogalactopyranoside) at 30 °C for 4 h. The cells were harvested, suspended in PBS buffer (7.81 mM Na2HPO4, 1.47 mM KH2PO4, 137 mM NaCl, and 2.68 mM KCl) containing 1 mM PMSF, 1 μg/ml aprotinin, 1 μg/ml leupeptin and 1 μg/ml pepstatin, and then disrupted with a French pressure cell press. Triton X-100 was added to the cell lysates at a final concentration of 0.5%, and then the lysate was centrifuged at 12000 g for 30 min. The resulting supernatants were incubated with glutathione–Sepharose (GE Healthcare) or amylose–agarose (New England Biolabs) resins for 4 h at 4 °C. The resins were washed with PBS buffer, and bound proteins were eluted with PBS buffer containing 20 mM glutathione or 10 mM maltose.
In vitro pull-down assay
MBP–MTH1 protein (0.5 μg) or 100 μg of lysate from yeast cells expressing Mth1p–GFP was incubated at room temperature (25 °C) for 1 h with 1.0 μg of GST, GST–NHX1Ct (full length), GST–NHX1Ct (489–560), GST–NHX1Ct (561–688) or GST–NHX1Ct (524–604) proteins immobilized to glutathione–Sepharose. After extensive washes with PBS buffer, bead-bound fractions were resolved by SDS/PAGE (10% gels), and then GST, MBP or GFP proteins were detected by immunoblotting with anti-GST, anti-MBP or anti-GFP (GF200; Nacalai Tesque) antibodies respectively.
Immunoblotting experiments and antibodies
Protein samples to be analysed were subjected to SDS/PAGE (10% gels). The separated proteins were transferred on to a hydrophobic PVDF membrane (Millipore). The membranes were incubated with 10% (w/v) skimmed milk powder in PBST (PBS with Tween 20; 7.81 mM Na2HPO4, 1.47 mM KH2PO4, 137 mM NaCl, 2.68 mM KCl and 0.1% Tween 20) and then treated with primary antibodies. After washes with PBST, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies. Immunoreactive bands were visualized by means of the enhanced chemiluminesence method (Amersham Pharmacia Biotech). The intensity of immunoreactive bands was quantified using Image J. The results shown are the means±S.D. for at least three independent experiments. The rabbit polyclonal anti-GST antibody was prepared by immunizing a rabbit with the bacterially expressed GST followed by affinity purification. The mouse anti-MBP antibody was purchased from Sigma–Aldrich. Rabbit anti-GFP serum and mouse anti-Vph1p antibodies were purchased from Invitrogen Molecular Probes. Horseradish peroxidase-conjugated secondary antibodies against rabbit and mouse IgG were purchased from Vector Laboratories.
Yeast cells expressing Nhx1p–GFP or Mth1p–GFP were grown in SCD or SCGal medium at 30 °C to the exponential phase of growth and observed under a fluorescence microscope (BX51; Olympus) equipped with a NIBA filter. Images were recorded using an ORCA-ER1394 digital camera (Hamamatsu Photonics).
Cell growth assay
Yeast cells were grown in YPAD or SCD medium at 30 °C to the expotential phase of growth and then diluted serially as indicated. The cells were spotted on to SD or SGal plates (made with 20 mM Mes/Tris, pH 5.5) supplemented with hygromycin. The plates were incubated for 3–6 days at 30 °C. For the growth assay in acidic pH medium, SD or SGal plates adjusted to pH 2.5 with phosphoric acid were used. Cell growth was monitored spectrophotometrically (D600) and the relative growth rate was expressed as a percentage of the rate of growth in the absence of hygromycin or in culture medium adjusted to pH 5.5. The results are means±S.D. for at least three independent experiments.
Other procedures and materials
DNA manipulations were performed in accordance with procedures published previously . Protein levels were measured by the Bradford assay as described previously . Restriction enzymes, Thermococcus kodakaraensis DNA polymerase and T4 DNA ligase were purchased from Toyobo and Takara. Oligomer primers used in the present study were synthesized by Invitrogen. Other materials were of the highest commercially available grade.
Mth1p interacts with the hydrophilic tail of Nhx1p
To survey binding partners of Nhx1p, we performed a yeast two-hybrid screening with the hydrophilic tail of Nhx1p [residues 489–633; NHX1Ct (full length)] as bait. We obtained 514 positive clones from 3.4×106 independent clones in an S. cerevisiae cDNA library and then sequenced the plasmid DNA carried by 94 randomly selected positive clones. Two of the sequenced clones carried a plasmid with a gene encoding MTH1 . These plasmids complemented the Ade− His− phenotype of the yeast mutant cells in the presence of a plasmid carrying GAL4-BD-fused NHX1Ct (full length) (Figure 1).
Screening of NHX1Ct-binding partners by a yeast two-hybrid assay
We analysed further the interaction of Nhx1p and Mth1p using a yeast two-hybrid system and identified a specific region within the C-terminal tail of Nhx1p required for interaction with Mth1p (Figure 2). For this, we expressed fragments containing the N-terminal (residues 489–560), central (residues 524–604), or C-terminal (residues 561–633) region of NHX1Ct (Figure 2A), together with GAL4-AD–MTH1. Yeast cells expressing each fragment alone or co-expressing the N-terminal region with GAL4-AD–MTH1 did not grow on the selection plates during incubation for 6 days (Figure 2B, numbers 4, 5, 6 and 8), whereas the cells co-expressing the central or C-terminal regions with GAL4-AD–MTH1 showed obvious cell growth on the selection plates after 6 days of incubation (Figure 2B, numbers 7 and 9). From the growth rate of yeast cells on the selection plates, the C-terminal region was estimated to interact with Mth1p more weakly than the central region, and the N-terminal region interacted very little.
Identification of Mth1p-binding region within the Nhx1p C-terminus by a yeast two-hybrid assay
We next confirmed the interaction between recombinant GST–NHX1Ct, a fusion of GST and the hydrophilic tail of Nhx1p, and MBP–MTH1 produced in E. coli cells and purified, by means of an in vitro pull-down assay. GST–NHX1Ct clearly bound to MBP–MTH1, whereas GST alone did not (Figure 3A). This result indicates that the hydrophilic tail of Nhx1p directly interacts with Mth1p. The GST–NHX1Ct (489–633) also interacted with Mth1p–GFP expressed in yeast cells (Figure 3B). Moreover, Mth1–GFP specifically bound to GST–NHX1Ct (524–604), but not efficiently to GST–NHX1Ct (489–560) and GST–NHX1Ct (561–633) peptide, confirming the results of the two-hybrid assay. These results strongly suggested that the central region of Nhx1p hydrophilic domain (residues 489–560) is responsible for binding Mth1p.
In vitro binding between the Nhx1p C-terminus and Mth1p
Mth1p is a negative regulator of Nhx1p activity
Glucose induces expression of HXTs (hexose transporters) [40,41]. Mth1p localizes mainly in the nucleus in the absence of glucose, but a small fraction of total Mth1p is also present in the plasma membranes due to interactions with the plasma membrane glucose sensor Snf3p/Rgt2p [41,42]. Mth1p can shuttle between nucleus and cytoplasm. However, in culture medium containing glucose, Mth1p associated with glucose sensor in the plasma membrane is ubiquitylated in response to a glucose signal and rapidly degraded . Thus the protein level of Mth1p is low in the presence of glucose [41,43]. Removal of glucose prevents this degradation of Mth1p and permits its translocation into the nucleus. In the nucleus, Mth1p bound to the repressor protein Rgt1p suppresses the expression of the HXTs [31,41,44,45]. These findings indicate that Mth1p is a negative transcriptional regulator of an extracellular glucose-sensing mechanism. As reported previously , we found that a Mth1p–GFP fusion, tagged with GFP chromosomally at its C-terminus, was present in both the nucleus and cytoplasm in the cells cultured in medium containing 2% (w/v) galactose (results not shown) and was detected at an expected size by immunoblotting using an anti-GFP antibody (Figure 3B, input). In contrast, a only very weak Mth1p–GFP signal was detected in the cells cultured in medium containing 2% (w/v) glucose by both of microscopic observation (results not shown) and immunoblotting (Figure 3B, control). Thus the majority distribution of Mth1p in yeast cells was not identical to that of Nhx1p (i.e. at the endosomal membranes), implying that the interaction between these two proteins takes place for small fractions of Nhx1p and Mth1p, and is transient.
To clarify the physiological significance of the binding of Mth1p to Nhx1p, we analysed the effect of MTH1-knockout on Nhx1p-activity-mediated resistance to hygromycin. Deletion of NHX1 retarded cell growth on agar plates containing hygromycin and 2% (w/v) galactose (Figure 4A). This increase in sensitivity to hygromycin upon NHX1-knockout was also observed on glucose/agar plates (Figure 4B). On agar plates containing 2% (w/v) galactose, growth of cells with a disrupted MTH1 gene (mth1Δ) was substantially more resistant to hygromycin than wild-type cells (Figure 4A). However, mth1Δ cells showed a resistance to hygromycin similar to that of the wild-type cells on agar plates containing 2% (w/v) glucose (Figure 4B), consistent with Mth1p being degraded in the wild-type cells on glucose plates. In a liquid culture assay, deletion of the MTH1 gene caused higher resistant to hygromycin than that of the wild-type cells in galactose-containing medium, consistent with the results of the agar-plate assay (Figure 4C). These results suggest that deletion of MTH1 elevated Nhx1p activity, suggesting that Mth1p is a negative regulator of Nhx1p. Cells with a double deletion of NHX1 and MTH1 grew at a rate similar to that of nhx1Δ cells in the presence of hygromycin (Figures 4A–4C), also supporting the notion that Mth1p is a negative regulator of Nhx1p.
Growth of MTH1-knockout cells is more resistant to hygromycin than that of wild-type cells in medium containing galactose
S. cerevisiae utilizes a variety of sugars as carbon and energy sources. Whereas glucose enters the glycolysis pathway directly, galactose, glycerol and ethanol are converted into intermediate metabolites before entering the glycolysis pathway. Outside of the cells yeast hydrolyse raffinose (a trisaccharide composed of glucose, galactose and fructose) by glycosidases, such as invertase and melibiase, secreted from cells in order to obtain glucose [46,47]. Therefore we analysed the effect of other carbon sources, such as raffinose or glycerol/ethanol, on the increase in resistance to hygromycin seen upon MTH1 knockout. Cell growth of mth1Δ cells showed that the resistance to hygromycin was similar to that of the wild-type cells on agar plates containing 2% (w/v) raffinose (Figure 4D), whereas mth1Δ cells showed more resistance to hygromycin than that of the wild-type cells on plates containing 3% glycerol and 2% (v/v) ethanol (Figure 4E). These findings strongly suggest that regulation of Nhx1p activity by Mth1p depends on the extracellular glucose level.
Another known phenotype of nhx1Δ cells is that they grow more slowly than the wild-type cell in acidic medium, possibly due to a defect in the regulation of vacuolar or cytoplasmic pH . To confirm a functional interaction between Nhx1p and Mth1p, we also analysed the role of Mth1p in cells growth at pH 2.5. Indeed, nhx1Δ cells grew more slowly than wild-type cells on agar plates and liquid cultures at pH 2.5, whereas mth1Δ and wild-type cells showed similar growth on culture medium adjusted to pH 5.5 (Figure 5). When yeast cells were grown in medium containing galactose as the sole carbon source, mth1Δ cells showed a slightly higher growth rate on agar plates adjusted to pH 2.5 than the wild-type cells (Figure 5A). In contrast, nhx1Δ mth1Δ cells showed a growth defect similar to that of the nhx1Δ cells (Figure 5A). As expected, deletion of MTH1 as a control had no effect on cell growth in glucose medium at acidic pH (Figure 5B). The results of a liquid-culture assay were essentially the same as those with agar plates (Figure 5C). These results also strongly suggest that the interaction with Mth1p regulates Nhx1p activity in response to extracellular carbon sources.
MTH1-knockout cells show higher cell growth at low pH than wild-type cells in medium containing galactose
Next, in order to test whether the increasing resistance to hygromycin observed in mth1Δ cells is due to up-regulation of Nhx1p activity, we examined how cells expressing mutant Nhx1p, lacking the binding region for Mth1p, grow in medium containing hygromycin. To do this, we constructed a series of Nhx1p mutants with various deletions from the C-terminus of the hydrophilic region (Figure 6A). All Nhx1p mutants were expressed and localized at the endosomes, like the wild-type Nhx1p (Figure 6B). Expression of the wild-type Nhx1p in the nhx1Δ cells restored the hygromycin-sensitivity of nhx1Δ cells to the wild-type level on both galactose- (Figure 6C) and glucose- (Figure 6D) containing agar plates. As shown in Figure 6(C), the Nhx1p mutant with a deletion of the C-terminal 40 amino acids (Δ604) showed slightly lower resistance to hygromycin than the wild-type Nhx1p on agar plates containing 2% (w/v) galactose. Moreover, nhx1Δ cells expressing Nhx1p mutants with a deletion of the C-terminal 80 and 120 amino acids (Δ560 and Δ523 respectively), regions that include the Mth1p-binding region, showed higher resistance to hygromycin than the Δ604 mutant, whereas a mutant Nhx1p with a complete deletion of the hydrophilic region of Nhx1p (Δ488) caused severe retardation of cell growth in the presence of hygromycin (Figure 6C). In addition, on glucose/agar plates, deletion of the C-terminal 40 amino acids (Δ604) caused a slight decrease of cell growth in the presence of hygromycin (Figure 6D). However, the Δ560 and Δ523 Nhx1p mutants showed essentially the same level of cell growth in the presence of hygromycin as the Δ604 mutant (Figure 6D). These results suggest that the increasing resistance to hygromycin on galactose plates is due to impairment of the interaction between Nhx1p and Mth1p. Moreover, these results also indicate that residues within the regions 489–523 and 605–633 play an important role in Nhx1p activity, whereas residues 524–604, which include the Mth1p-binding region, are not involved in the regulation of Nhx1p activity on glucose/agar plates.
Disruption of the Nhx1p–Mth1p interaction increases resistance to hygromycin
Mth1p does not affect the expression level of Nhx1p
To analyse the down-regulation of Nhx1p activity by Mth1p, we tested whether deletion or overexpression of MTH1 affects the expression level of Nhx1p. When these strains were grown in medium containing 2% (w/v) galactose (Figure 7A) or 2% (w/v) glucose (Figure 7B), Nhx1p–GFP fusions in both mth1Δ cells and MTH1-overexpressing cells were expressed at the same level as in wild-type cells. These results suggest that the apparent activation of Nhx1p activity by the loss of MTH1 is not due to an increase in the expression of Nhx1p in cells.
Mth1p does not affect the levels of Nhx1p within yeast cells
In the present study we found that the yeast endosomal NHE Nhx1p interacts with Mth1p, a transcriptional regulator mediating a signal of extracellular carbon sources, especially in the presence or absence of glucose. Mth1p negatively regulates Nhx1p activity when yeast cells are grown in medium containing galactose or glycerol/ethanol, but not glucose or raffinose. This is the first study showing that Nhx1p activity is regulated by interaction with a transcriptional regulator in response to changes in extracellular carbon sources. Nhx1p activity contributes to the regulation of organellar pH. Therefore we propose that the regulation of Nhx1p function by Mth1p is required to maintain the pH of organellar lumina regardless of changes in extracellular carbon source.
Yeast V-ATPase activity is tightly controlled by the extracellular glucose level [23–25]. In medium containing 2% (w/v) glucose, most Vo subunits of V-ATPase are assembled with V1 subunits, but after 5 min of glucose depletion, only 20% of Vo subunits are bound to V1 subunits [23–25]. The dissociated V1 subunit shows very little ATPase activity, and the Vo subunit does not appear to form an open proton pore [48,49]. Thus this dissociation between the V1 and Vo subunits of V-ATPase impairs the coupling of ATP hydrolysis and proton transport, leading to alkalization of the organellar lumen [48–51]. In this situation, it would be reasonable to prevent alkalization in the organelle lumen by decreasing Nhx1p activity through the actions of Mth1p, as we found in the present study. Regulation of V-ATPase activity by glucose has also been observed in mammalian cells . Therefore mammalian organelle-type NHEs may also be controlled in response to extracellular glucose, as yeast Nhx1p is. However, because Mth1p is not conserved in mammalian cells, mammalian organelle-type NHEs may not be regulated in exactly the same way as yeast Nhx1p.
We have shown that Mth1p does not affect the stability of Nhx1p in cells (Figure 7). Deletion or overexpression of MTH1 also did not cause mislocalization of Nhx1p–GFP fusions within cells grown in medium containing either 2% (w/v) glucose or galactose (results not shown). These findings indicate that the apparent activation of Nhx1p upon the loss of MTH1 (Figures 4 and 5) is not due to an increase in the expression level or change of the intracellular localization of Nhx1p in cells. Moreover, overexpression of MTH1 in wild-type (NHX1+) cells decreased the resistance to hygromycin, although no increase in sensitivity to hygromycin was observed in nhx1Δ cells (results not shown). Accordingly, we propose that the interaction with Mth1p inhibits the antiporter activity of Nhx1p, and that Nhx1p is constitutively activated in mth1Δ cells. Although we measured the vacuolar pH of the wild-type and mth1Δ cells grown in medium containing glucose or galactose, by using the pH-sensitive fluorescent dye BCECF [2′,7′-bis-(2-carboxyethyl)-5(6)-carboxyfluorescein] trapped in yeast vacuoles as described previously , we did not detected a significant difference in the vacuolar pHs between the wild-type and mth1Δ cells (results not shown). This result suggested that the difference in vacuolar pH between the wild-type and mth1Δ cells is very small or transient. We also investigated the effect of MTH1-deletion on CPY (carboxypeptidase Y) secretion, defects in which are a known phenotype of nhx1Δ cells . The mth1Δ cells showed the same level of CPY secretion as the wild-type cells in medium containing glucose or galactose (results not shown). This result also suggests the change of vacuolar (or endosomal) pH by MTH1 deletion is very small or transient.
It has been shown previously that yeast Nhx1p also interacts with Gyp6p, a GTPase-activating protein for the yeast Rab family member Ypt6p . Gyp6p also interacts with the C-terminal region (residues 607–633) of the Nhx1p C-terminus, but in a different area to the Mth1p-binding site, so Gyp6p may not compete with the binding of Mth1p to Nhx1p. The ability of Nhx1p to bind Gyp6p and Mth1p suggests that there are at least two independent regulatory mechanisms. Δ488 or Δ604 Nhx1p mutants showed retardation of cell growth in the presence of hygromycin relative to that of the Δ523 mutant and wild-type Nhx1p respectively (Figure 6). Therefore we conclude that the N-terminal and C-terminal halves (residues within 489–523 and 605–633 respectively) of the Nhx1p C-terminal hydrophilic region also play important roles in Nhx1p activity regardless of the extracellular carbon source. Both Δ488 and Δ604 Nhx1p mutants were stably expressed and localized at dot-like structures, similar to the wild-type Nhx1p (Figure 6B), suggesting that these regions are involved in the antiporter activity of Nhx1p, not in endosomal targeting of Nhx1p. Although the exact roles of these regions in Nhx1p activity remains unknown, these results might indicate that some Nhx1p-enhancing factors besides Mth1p and Gyp6p interact with the Nhx1p C-terminus.
In mammalian cells, CHP1 (calcineurin homologous protein 1) interacts with an more N-terminal membrane-proximal region in the hydrophilic tail of NHE1 . This binding plays an essential role in stabilizing the functional NHE1 molecule with the antiporter activity [29,30]. However, because CHP1 does not bind to the Nhx1p orthologue NHE6 , an alternative molecule, similar to CHP1, might be involved in Nhx1p function. The N-terminal half of the C-terminal hydrophilic region of the yeast plasma membrane NHE Nha1p is also essential for ion transport  and is capable of binding Cos3p, which enhances the antiporter activity . These findings together with the present results suggest that the N-terminal half or membrane-proximal region of the C-terminal hydrophilic region of NHEs is generally important for their function and regulation.
We thank Dr Hiroshi Nojima (Osaka University, Japan) for providing the yeast cDNA library.
calcineurin homologous protein 1
green fluorescence protein
PBS with Tween 20
receptor for activated C-kinase 1
Keiji Mitusi conducted most of the experiments and wrote the original draft of the manuscript. Masafumi Matsushita provided suggestions and discussion for the completion of the manuscript. Hiroshi Kanazawa suggested the original direction of the research, edited the manuscript prior to submission and supported the project financially.
This work was supported by a Grant-in-Aid from the Japanese Ministry of Education, Science, Sports, Technology, and Culture [grant number 21370055].