INTRODUCTION

Living cells actively maintain low cytoplasmic sodium ion concentrations against large, inwardly directed Na⁺ gradients. In animal cells, the intracellular Na⁺/K⁺ ratio is largely dependent on the plasma membrane Na⁺/K⁺-ATPase, a P-type ion pump, that drives Na⁺ ions out of the cell in exchange for K⁺ ions (1). The resultant Na⁺ gradient serves as the primary energy source for the transport of other ions and metabolites via an array of secondary, Na⁺-coupled carriers. In lieu of sodium, plants and fungi utilize H⁺-coupled circuits, driven by the plasma membrane H⁺-ATPase, PMA1, also a member of the P-ATPase family (2, 3). Mechanisms for dealing with toxic concentrations of Na⁺ have only recently begun to be elucidated at a molecular level, but are of increasing urgency as soil salinity rises and poses a significant threat to agricultural production worldwide (4, 5). Because of a basic similarity in ion transport processes, the use of yeast as a model system to identify genes involved in halotolerance is of particular applicability to higher plants. The recent availability of the complete genome sequence from Saccharomyces cerevisiae has brought to light unexpected homologs of both prokaryotic and eukaryotic Na⁺ transporters, providing a convenient starting point for dissecting the individual contributions of these transporters toward sodium tolerance.

Multiple transport pathways, all of which are under complex regulation, appear to mediate cellular Na⁺ homeostasis in S. cerevisiae. One route of Na⁺ entry is thought to be the K⁺ transporters, TRK1 and TRK2 (6-10). Under conditions of Na⁺ stress or K⁺ starvation, the principal cation carrier is TRK1, which has been proposed to limit Na⁺ entry by increasing K⁺/Na⁺ discrimination (9, 11). Activation of calcineurin, the Ca²⁺- and calmodulin-dependent protein phosphatase, is required for the transition in response to Na⁺ stress (12). The primary pathway for Na⁺ extrusion from the cell is through the P-type ion pumps encoded by the PMR2/ENA locus, consisting of an unusual tandem array of nearly identical genes (PMR2A-E; Refs. 13-15). Expression of PMR2A is induced by high pH and Na⁺ stress (16) and is modulated by a host of factors, including calcineurin (17, 18). Three distinct genes encoding putative Na⁺/H⁺ antiporters have been identified by the genome sequencing project. One of these, NHA1, was recently cloned by selection for increased NaCl tolerance from a multicopy genomic library (19). Although activity and localization of the NHA1 protein has not yet been established, it is homologous (40% identity) to a plasma membrane Na⁺/H⁺ exchanger, encoded by the sod2 gene in the fission yeast Schizosaccharomyces pombe, which mediates Na⁺ extrusion at acidic to neutral pH (20). Disruption of NHA1 confers significant Na⁺ sensitivity in a strain lacking the PMR2 locus, but only a weak phenotype in a wild-type background. A protein with homology to putative Na⁺/H⁺ exchangers from Enterococcus hirae and Lactococcus lactis, as well as to a putative K⁺/H⁺ exchanger (KefC) from Escherichia coli (22-24), is encoded by the yeast gene YJL094c (21). A third yeast gene, YDR456w, encodes a protein sharing significant homology (~30% identity) with the amiloride-sensitive Na⁺/H⁺ exchangers (NHE1-4) in animal cells (25). The latter play physiologically vital roles in the regulation of intracellular pH, in Na⁺ concentration, and in cell volume control. Activation by a variety of growth factors, hormones, and cytoplasmic acidosis leads to H⁺ extrusion from the plasma membrane, in exchange for an influx of Na⁺ ions (reviewed in Refs. 26 and 27).

In this work, we report that mutations in the plasma membrane H⁺-ATPase, PMA1, confer Na⁺ tolerance in yeast. We show that the mutant cells sequester Na⁺ in a slowly exchanging pool, and that this sequestration is mediated via the yeast homolog of the mammalian amiloride-sensitive Na⁺/H⁺ exchanger, which we term NHX1.

EXPERIMENTAL PROCEDURES

All strains used in this study are isogenic to W303 (ade2-1 can1-100 his3-11, 15 leu2-3, 112 trp1-1 ura3-1) (28). Spontaneous Na⁺-tolerant suppressors of calcineurin deficiency were isolated independently in two strains described previously (29), K473-2 (MATa pmc1::LEU2 cnb1-2) and K482-3 (MAT pmc1::TRP1 cnb1-3), by selection on YPD agar medium supplemented with 1.0 M NaCl. After 3 days at 30 °C, a single Na⁺-tolerant colony was picked from each patch, purified, and subjected to complementation testing. Of 41 independent suppressors, 35 were recessive and defined two complementation groups of 33 and 2 alleles, respectively. The remaining six suppressors were dominant or semidominant; after crossing with the parent of the opposite mating type, four alleles were tightly linked to pmc1::TRP1 and one allele was tightly linked to pmc1::LEU2 (zero recombinants in 26 complete tetrads). The PMC1 gene encodes a vacuolar Ca²⁺ pump, which plays no detectable role in Na⁺ tolerance and is adjacent to the PMA1 gene. Complementation tests with well characterized alleles of pma1 (30) showed all five suppressors linked to PMC1 are alleles of PMA1. Strains K638 (PMA1) and K804 (pma1-4) are MAT derivatives of W303 carrying pmc1::TRP1 and pmr2::HIS3 (pmr2A-E) null alleles. YR89 (pmr2A-E) was a gift from Hans Rudolph (University of Stuttgart, Germany).

YPD medium contained 2% glucose, 1% yeast extract, and 2% peptone (all from Difco), and was adjusted to pH 7 with sodium phosphate, where indicated. APG is a synthetic minimal medium containing 10 mM arginine, 8 mM phosphoric acid, 2% glucose, 2 mM MgSO₄, 1 mM KCl, 0.2 mM CaCl₂, and trace minerals and vitamins, at pH 6.7, as described (11). Where indicated, NaCl was added, or the pH was adjusted to 5.0 by addition of acetic acid. Growth assays were performed by inoculating 1 ml of APG medium in a multiwell plate with 2-5 µl of a saturated seed culture. Growth was monitored by measuring absorbance at 600 nm after culturing for 48 h at 30 °C.

The chromosomal pma1 alleles were rescued by digestion of genomic DNA with PstI, to release a ~23-kilobase pair fragment containing the disrupted pmc1 gene (29) and the adjacent pma1 gene. After treatment with DNA ligase and transformation into E. coli, Amp^R plasmids were recovered, and a 4.8-kilobase pair HindIII fragment containing intact pma1 was cloned into the yeast centromeric vector YCplac111 (31). Plasmids were transformed into strain YR89 (pmr2A-E) using the lithium acetate procedure (32). A series of YCplac111-based plasmids containing various fragments of the pma1-4 allele, subcloned into wild-type PMA1, were generated and also introduced into strain YR89. By comparison of the Na⁺-tolerant phenotype conferred by the plasmids, the mutation in the pma1-4 allele was localized to a 615-base pair fragment between BstEII and EcoRI, and identified by DNA sequencing. A single base change, G  A, resulting in the substitution Glu³⁶⁷  Lys, was found.

The NHX1 gene was cloned by amplification of genomic DNA from K638 using the polymerase chain reaction and the following primers: sense primer 5-CGCCATTGTGTATCCATTTATGC-3 at 598 from the initiating codon ATG, and the antisense primer 5-CTCACCAATTATACGAGTAG-3 at +350 from the terminating codon TAG. The amplified product was digested with HindIII and NheI, and the resultant 2662-base pair fragment cloned into the HindIII and XbaI sites of pBluescript II KS (Stratagene). A null allele (nhx1::URA3) was constructed by replacement of the 1286-base pair EcoRV-SpeI fragment of NHX1 with a 1-kilobase pair SmaI-HindIII fragment of URA3, after treatment of DNA fragments with Klenow enzyme to create blunt ends.

Yeast strains were grown in APG medium, supplemented where indicated with 20 mM NaCl, to a density of 0.9-1.4 OD₆₀₀ units/ml. Cells were harvested by centrifugation, washed once in APG medium ± 20 mM NaCl (depending on growth conditions), and resuspended in the same medium at a density of 4-8 × 10⁸ cells/ml. Uptake was initiated by diluting the cells with an equal volume of APG medium containing NaCl at a final concentration of 20 mM and ²²NaCl (NEN Life Science Products) at 13-30 µCi/ml. Samples were incubated at 23 °C, and at the indicated times aliquots were withdrawn and the reaction quenched by the addition of ice-cold AP medium containing 75 mM potassium gluconate. The reaction mixture was rapidly filtered through 0.45-µm HAWP membranes (Millipore) and washed twice with 6 ml of quench buffer. Radioactivity retained on the filters was measured by liquid scintillation counting. In assays of ²²Na efflux, cells were loaded with ²²NaCl exactly as above. After incubation for 45 min, cells were collected by centrifugation, washed twice in ice-cold Buffer A (10 mM Tris-HCl, pH 6.0, 2 mM MgCl₂, 1 mM KCl, 1% glucose, 0.6 M sorbitol), and resuspended in the same buffer at room temperature. At the indicated times, aliquots were withdrawn, filtered, and processed as described above. Steady-state labeling of cells with ²²Na was performed in 1 ml of APG medium containing 1-4 µCi of ²²NaCl and varying concentrations of nonradioactive NaCl. Cultures were incubated at 23 °C for 96 h, after which the cells were collected by rapid filtration. Filters were washed twice with 6 ml of APG medium, and radioactivity assessed by liquid scintillation counting. The optical density of a duplicate, nonradioactive culture was measured to quantitate growth.

After labeling with ²²Na as described above, cells were collected by centrifugation, washed twice with ice-cold Buffer A (see above), and resuspended at 1 × 10⁸ cells/ml in Buffer A, essentially as described by Anraku and co-workers (33). One half of the suspension was permeabilized by addition of CuSO₄ to a final concentration of 500 µM, while the other half received an equal volume of water (22.5 µl). After incubation at 23 °C for the indicated times, aliquots were withdrawn, filtered, and processed as described above. To assay ²²Na influx in permeabilized cells, cultures (2 × 10⁸ cells/ml) were preincubated with 500 µM CuSO₄ in Buffer A for 1 h as above. Uptake was initiated by diluting cells with an equal volume of Buffer A supplemented to give a final concentration of 20 mM NaCl, 9 µCi/ml ²²Na, 2 mM Tris-ATP, and 500 µM CuSO₄. At the indicated times, aliquots were withdrawn for filtration as above.

RESULTS

Calcineurin promotes Na⁺ tolerance by increasing expression of the PMR2A/ENA1 gene (17, 18) and by converting the K⁺ transport system to a high affinity, Na⁺ discriminatory state (12). Loss of calcineurin function, as in cnb1 mutants, therefore causes sensitivity to high salt conditions. To learn more about salt tolerance in yeast, we isolated spontaneous Na⁺-resistant suppressors of the cnb1 mutants by selection on YPD medium supplemented with 1.0 M NaCl. Five dominant mutations were mapped to the PMA1 locus as described under "Experimental Procedures." All five mutations in PMA1 conferred different degrees of Na⁺ tolerance (Fig. 1A), and Li⁺ tolerance (data not shown), relative to the parent strains. All five Na⁺-tolerant mutations in PMA1 also conferred sensitivity to H⁺ (low pH) and were recessive to wild type for this phenotype. Several pma1 alleles isolated in genetic screens not involving high salt (30) were also found to be recessive for H⁺ sensitivity and dominant or semidominant for Na⁺ tolerance when calcineurin is inactivated (data not shown). These results suggest the Na⁺-tolerant alleles of PMA1 are loss-of-function mutations with regard to H⁺ pumping and thus are referred to as recessive pma1 alleles.

pmc1 cnb1

pma1-18

pma1-3

pma1-4

pma1-30

pma1-4

pma1-4

pmc1 vcx1

pma1-4

pmr2A-E

pmr2A-E

Because of its strong Na⁺-tolerant phenotype, pma1-4 was chosen for further analysis. The mutant allele was rescued from the chromosome, subcloned, and partially sequenced (see "Experimental Procedures") to reveal the nature of the mutation: a single amino acid substitution, Glu³⁶⁷  Lys, within the highly conserved phosphorylation domain of the ATPase (Fig. 1 B). ATPase activity measured in total membrane preparations, was reduced by 45%, relative to the PMA1 control (data not shown). Mutations in the immediate vicinity of this residue (Ser³⁶⁸) have been reported to cause a similar reduction in ATPase activity, H⁺ pumping, and defects in membrane potential (34, 35). Thus, it was unlikely that the pma1-4 mutant had gained a novel ability to pump Na⁺ ions, despite its dominance over PMA1 in Na⁺ tolerance. The Na⁺-tolerant phenotype of the pma1-4 mutation was also found to be independent of the entire PMR2 locus, in the absence or presence of calcineurin function (Fig. 1C). This ruled out the possibility that PMR2 expression or function was improved by the pma1 mutant. Finally, it should be noted that the effects of pma1-4 on Na⁺ tolerance were found to be completely independent of the vacuolar Ca²⁺ transporters encoded by PMC1 and VCX1 (data not shown), mutations in which were included in some experiments for technical reasons. These results suggest that pma1-4 promotes Na⁺ tolerance through processes independent of calcineurin and the PMR2-encoded Na⁺ efflux pumps.

The pma1-4 mutation Limits Na⁺ Influx after Exposure to Na⁺ in the Growth Medium

Because the plasma membrane H⁺-ATPase generates the driving force for active transport, we tested the hypothesis that reduced influx of Na⁺ in the pma1-4 mutant contributes to Na⁺ tolerance. Fig. 2A shows the time course of ²²Na uptake by the pma1-4 mutant (K804) and the isogenic PMA1 strain (K638), after culture in APG medium lacking NaCl (see "Experimental Procedures"). Previous work has shown that under these conditions, uptake of monovalent cations (K⁺, Rb⁺) is in the low affinity mode and is unaffected by dissipation of the H⁺ gradient in response to protonophore addition or ATP depletion (11). Consistent with these observations, initial rates of ²²Na uptake were similar in both mutant and control, although at longer time points, uptake in the mutant consistently exceeded that of the PMA1 control cells (Fig. 2A). A shift in the cation uptake system to a high affinity, K⁺-selective mode has been reported to occur within 4 h of K⁺ depletion from the growth medium; under these conditions, transport is sensitive to protonophore addition and ATP depletion and is dependent on the TRK1 transporter (11). Induction of the K⁺-selective mode has also been reported to occur after exposure to Na⁺ or Li⁺ in the growth medium (9). Fig. 2B shows that, after growth in APG medium containing 20 mM NaCl, ²²Na uptake in the control strain K638 was drastically reduced, consistent with increased discrimination against Na⁺ ions. Furthermore, uptake was reduced by an additional 2-fold in the pma1-4 mutant, relative to control. The results demonstrate that, after salt adaptation, the pma1-4 mutation increases Na⁺ tolerance by limiting Na⁺ influx, possibly through modulation of the TRK1 transporter.

pma1-4 pmr2A-E pmc1)

PMA1 pmr2A-E pmc1)

Na⁺ Tolerance Is pH-dependent and Correlates with Intracellular Na⁺ Sequestration in the pma1-4 Mutant

External pH had a striking effect on Na⁺ tolerance; at neutral pH, the pma1-4 mutant (K804) grew to higher cell densities (10-fold relative to control, K638) in response to increasing concentrations of NaCl (Fig. 3A). However, Na⁺ tolerance increased with decreasing pH and was similar in both pma1-4 mutant and control strains at pH 5 (Fig. 3B). To assess intracellular Na⁺ levels, cells were grown to saturation in APG medium containing ²²NaCl (0.1 µM) and varying concentrations of unlabeled NaCl (see "Experimental Procedures"). Surprisingly, accumulation of Na⁺ in the sodium-tolerant mutant exceeded that of control cells, by 2-5-fold, in media containing tracer ²²Na alone (Fig. 3, C and D, inset). As external NaCl concentrations increased, intracellular Na⁺ in the mutant reached stable levels at both pH values tested (Fig. 3, C and D) and correlated with a similar stability in cell density. Overall, a 7-fold increase in external NaCl concentrations elicited only a 2-fold increase in intracellular Na⁺ in the mutant. In PMA1 control cells, a similar response was observed at pH 5; in contrast, at neutral pH, intracellular Na⁺ rose by 12-fold over the same range and was accompanied by a drastic decline in cell density. Thus, the ability to limit intracellular Na⁺ levels correlates with sodium-tolerant growth.

To determine the intracellular distribution of sodium, plasma membranes were selectively permeabilized by exposure to Cu²⁺ ions, as described by Anraku and co-workers (33). At least 50% of intracellular Na⁺ in the pma1-4 mutant appeared to be retained in a Cu²⁺-resistant pool under all conditions tested (Fig. 3, E and F). By comparison, in the PMA1 control, approximately 90% of intracellular Na⁺ was released after Cu²⁺ treatment of cells grown in media containing low NaCl. Based on a previous demonstration (33) that the Cu²⁺ permeablization technique allows the specific extraction of cytosolic pools of ions, amino acids, and other metabolites, our results strongly suggest a cytoplasmic localization of sodium in the PMA1 control cells. After exposure to higher concentrations of NaCl, particularly at acidic pH, Na⁺ sequestration in the PMA1 control increased to levels comparable to the pma1 mutant. A range of control experiments, including assays of ⁴⁵Ca²⁺ uptake, release of nucleotides, and turbidity changes in the cell suspension (33), were used to confirm that both mutant and control cells were equally permeabilized by the treatment with Cu²⁺ ions (data not shown). Therefore, these results demonstrate that the pma1-4 mutation enhances the ability to sequester Na⁺ in an intracellular pool, and that this sequestration may contribute significantly to halotolerance.

A Novel Na⁺/H⁺ Exchanger, NHX1, Contributes to the Na⁺-tolerant Phenotype of pma1-4

Based on previous observations that reduced H⁺ pumping in pma1 mutants leads to reductions in cytoplasmic pH (36), we hypothesized that cytosolic pH may be a controlling factor in sodium tolerance. Because cytoplasmic acidosis has been shown to activate the amiloride-sensitive Na⁺/H⁺ exchanger in mammalian cells (37), it seemed likely that activation of a similar exchanger might mediate Na⁺ tolerance in yeast. Systematic sequencing of the yeast genome has uncovered a gene present on chromosome IV (YDR456w), encoding a protein of 633 amino acids that we have named NHX1 on the basis of homology with the family of Na⁺/H⁺ exchangers (NHE1-4) found in mammals and other vertebrates, as well as in the nematode Caenorhabditis elegans, and the euryhaline crab Carcinus maenas (Fig. 4). We analyzed the effect of targeted disruptions of this gene in the following isogenic set of yeast strains: K601 (W303- derivative; see "Experimental Procedures"), K638 (pmc1 pmr2 PMA1), and K804 (pmc1 pmr2 pma1-4). In each of these genetic backgrounds, disruption of NHX1 led to a significant decrease in Na⁺ tolerance at pH 4 or 5 (Fig. 5, B, D, and F), substantiating its prominent role in mediating salt tolerance at acid pH. At neutral pH, however, the disruption had little or no effect on sensitivity to Na⁺ in the PMA1 strains (Fig. 5, A and C). In striking contrast, disruption of NHX1 effectively nullified the Na⁺-tolerant phenotype of the pma1-4 mutant at neutral pH (Fig. 5E). The results demonstrate that Na⁺ tolerance in the pma1-4 mutant is largely mediated by the putative Na⁺/H⁺ exchanger, NHX1.

nhx1

pmc1 pmr2 PMA1

pmc1 pmr2 pma1-4

To determine if sodium sequestration in the pma1-4 mutant was mediated by NHX1, we followed the time course of ²²Na efflux from Cu²⁺-treated cells, relative to untreated cells (Fig. 6A). In the PMA1 control strain, addition of Cu²⁺ elicited an immediate efflux of ²²Na, with less than 10% remaining in 20 min. By contrast, efflux of ²²Na was slow and biphasic in the pma1-4 mutant, with approximately 50% remaining in the cells after treatment with Cu²⁺ for 2 h. Disruption of NHX1 in the pma1-4 mutant effectively nullified the ability to sequester sodium, resulting in almost complete release of intracellular ²²Na upon permeabilization of the plasma membrane. The data also indicated that, in the absence of Cu²⁺ treatment, ²²Na efflux was low or absent in all three strains (Fig. 6A). Thus, there was no evidence that NHX1 was involved in a Na⁺ efflux mechanism in the pma1 mutant cells; instead, the data suggest that the PMR2/ENA pumps are the major route for Na⁺ efflux.

pma1-4

pma1-4 nhx1

Uptake of ²²Na was also monitored in Cu²⁺-permeabilized cells (Fig. 6B). The results demonstrate a substantial enhancement of Na⁺ uptake into Cu²⁺-resistant compartment(s) in the pma1-4 mutant, relative to the PMA1 control, that is completely abolished by the disruption in NHX1. Thus, a putative Na⁺/H⁺ exchanger is required for sequestration of Na⁺ within an intracellular, Cu²⁺-resistant pool, possibly the vacuole. One possibility is that mislocalization of the mutant pma1 protein to an intracellular compartment may stimulate Na⁺/H⁺ exchange by contribution of a proton motive force, although a plasma membrane localization for pma1-4 has been confirmed by immunofluorescence (data not shown). We suggest that cytoplasmic acidosis, caused by the pma1-4 mutation or by growth in low pH conditions, increases expression or function of NHX1 to promote Na⁺ tolerance.

DISCUSSION

Mutations in the plasma membrane H⁺-ATPase confer sensitivity to weak acids and low extracellular pH, reflecting a reduced ability to pump protons from cells (39). The resultant defect in the electrochemical H⁺ gradient is also believed to be responsible for the observed resistance of pma1 mutants to cytotoxic drugs, such as hygromycin B and Dio-9, that are dependent on the membrane potential for entry into the cell (36, 39, 40). There are conflicting reports in the literature on the role of the plasma membrane H⁺-ATPase in salt tolerance, possibly because complex stress and osmotolerance mechanisms are invoked at the very high ionic conditions used in those studies. On the one hand, some pma1 mutants have been reported to show decreased tolerance to high ionic and osmotic conditions (YPD medium supplemented with 0.75 or 1.5 M KCl or NaCl, or 2.5 M glycerol; Ref. 39); in contrast, other mutants apparently show increased tolerance to very high NaCl concentrations (2.5 M), high ethanol concentrations (12.5%), and heat shock (25-48 °C) (41). No molecular insights are available on the role of PMA1 in either situation.

In the present work, we have attempted to focus on Na⁺ ion toxicity by using low to moderate NaCl concentrations (20-200 mM) and avoiding conditions of osmotic shock and stress response, to identify novel Na⁺ transport pathways. We report that mutations in pma1 suppress the Na⁺-sensitive phenotype characteristic of a calcineurin defect. Two transport processes have previously been identified as downstream targets of calcineurin; increased sodium efflux by PMR2 (17, 18) and decreased sodium uptake via TRK1 utilize activated calcineurin in response to Na⁺ stress (12). Our studies demonstrate that pma1 mutations affect a pathway different from those already described. Thus, the Na⁺-tolerant phenotype of pma1 mutations does not require the presence of the PMR2 locus and is independent of calcineurin. Furthermore, our data suggest that the role of calcineurin in Na⁺ tolerance is almost entirely mediated by PMR2, since addition of the calcineurin inhibitor FK506 did not significantly alter the Na⁺ tolerance profile in pmr2 strains differing only in the pma1 mutation (Fig. 1). All of the Na⁺-tolerant pma1 mutants we found were also sensitive to acidotic conditions, as has been found for other pma1 alleles. Although H⁺ tolerance was restored in PMA1/pma1 diploids, some degree of Na⁺ tolerance was retained, suggesting that the two phenotypes may not entirely overlap.

Mechanism of Na⁺ Tolerance in the pma1-4 Mutant

To begin to elucidate the molecular mechanism for Na⁺ tolerance in pma1 mutants, we focused on a defined allele, pma1-4, in which the mutation (Glu³⁶⁷  Lys) was localized to the conserved phosphorylation domain. ATP hydrolysis was decreased by 45% in this mutant and was presumably accompanied by a concomitant reduction in H⁺ pumping, as has been reported for other mutations within this region (Ser³⁶⁸  Phe; Refs. 34 and 35). A priori, we considered three general mechanisms likely to reduce cytosolic sodium and confer sodium tolerance: reduced Na⁺ influx, increased Na⁺ efflux, and intracellular sequestration of Na⁺.

If Na⁺ influx is coupled to the electrochemical H⁺ gradient, then a reduction in the activity of the H⁺-ATPase would be expected to cause a similar reduction in the rate of Na⁺ entry. However, the analysis of Na⁺ uptake is complicated by the reported change in affinity of the TRK system for K⁺ after Na⁺ stress, resulting in increased discrimination against Na⁺, and limiting Na⁺ uptake (9, 11). We therefore monitored ²²Na uptake in cells cultured in the absence and presence of NaCl (20 mM). After prolonged growth in NaCl-containing media, the initial rate in the mutant was 2-fold lower than control (Fig. 2B), consistent with a similar decrease in ATPase activity. Under these conditions, Na⁺ uptake is likely coupled to H⁺ symport, presumably via the TRK1 transporter. We conclude that the Na⁺ tolerance of pma1 mutations is due in part to an ability to limit Na⁺ entry. We were not able to detect any increase in rates of ²²Na efflux from cells of the pma1 mutant; instead, in the absence of the plasma membrane Na⁺ pumps (PMR2/ENA), ²²Na efflux was virtually abolished (Fig. 6). Therefore, it was unlikely that a Na⁺ efflux mechanism was activated by the pma1 mutation.

Interestingly, in cells cultured in Na⁺-free media, there was little or no difference in the initial rate of ²²Na uptake (Fig. 2A) in the pma1-4 mutant relative to the PMA1 control, suggesting that Na⁺ influx under these conditions may be largely downhill, and unaffected by a reduction in the proton motive force. Rather, there was a reproducibly large enhancement of uptake at prolonged times in the mutant. This increased uptake is consistent with data from steady state labeling of cells with ²²Na (Fig. 3, C and D), showing increased intracellular Na⁺ in the mutant at low to moderate extracellular NaCl concentrations. The elevated levels of Na⁺ observed in the pma1 mutants suggested that enhanced compartmentalization of Na⁺ may provide a mechanism for salt tolerance. Using Cu²⁺ ions as a means to selectively permeabilize the plasma membrane, we were able to provide clear evidence for intracellular sequestration of Na⁺ (50% of total) in the pma1-4 mutant. Because of the striking correlation between Na⁺ sequestration and Na⁺-tolerant growth in the pma1 mutant, we propose that the ability to compartmentalize Na⁺ is an important mechanism for salt tolerance.

The pma1 mutations reported in this study effectively enhance Na⁺ tolerance at neutral to alkaline pH. Inasmuch as reductions in plasma membrane H⁺-ATPase activity have been linked to cytoplasmic acidosis (36), we predicted the existence of a Na⁺/H⁺ exchanger that would be activated by a decrease in cytoplasmic pH. There is precedence for low pH-induced activation of Na⁺/H⁺ exchangers from bacteria (38) and mammals (37), via defined H⁺ sensor sites on a cytosolic loop of the polypeptide. We show in this work that deletion of the NHX1 gene, encoding a homolog of mammalian plasma membrane Na⁺/H⁺ exchangers (NHE1-4), reverses the Na⁺-tolerant phenotype of the pma1 mutation. Furthermore, we show that NHX1 plays a prominent role in halotolerance at acidic pH. Because acidification of the extracellular medium with a weak acid, acetate, can lead to acidification of the cytosol and vacuole, the experimental observations of Fig. 5 are compatible with NHX1 residing in the plasma membrane or some intracellular membrane. However, we predict that NHX1 will likely localize to an intracellular acidic organelle, such as the vacuole, and contribute to Na⁺ sequestration. Indeed, deletion of NHX1 led to a loss in the ability to sequester Na⁺, concomitant with a loss in Na⁺ tolerance. In our model (Fig. 7), a reduction in plasma membrane H⁺-ATPase activity contributes to Na⁺ tolerance by (i) reducing the driving force for Na⁺ uptake and (ii) enhancing the activity of a novel intracellular Na⁺/H⁺ exchanger, via a decrease in cytosolic pH. It is noteworthy that increased sodium sequestration was also observed in PMA1 control cells at higher salt concentrations, particularly upon extracellular acidification. This would suggest that NHX1 is normally induced by low pH and/or high Na⁺, but is constitutively activated in the pma1 mutant. In experiments currently under way in our laboratory,¹ the NHX1 gene has been cloned, tagged with the hemagglutinin epitope, and shown to complement the Na⁺-sensitive phenotype of nhx1 yeast strains, allowing for further studies on the mechanism and localization of this novel exchanger in the near future.

The results described in this report are consistent with a vacuolar localization of NHX1, although other sites cannot be ruled out at this time. A well documented response to salinity stress employed by salt-tolerant plants is the effective vacuolar compartmentalization of sodium via an amiloride-sensitive Na⁺/H⁺ exchanger located on the tonoplast membrane (reviewed in Refs. 5 and 42). For example, in the salt-tolerant alga, Dunaliella parva, cells grown in medium with 0.4 M NaCl contained 0.3 M Na⁺ in the vacuole while maintaining 0.03 M Na⁺ in the cytoplasm (43). The activity of vacuolar Na⁺/H⁺ antiport has been shown to increase in response to increasing NaCl concentrations in the growth medium (44, 45), suggesting that it is involved in conferring salt resistance. However, despite biochemical characterization of vacuolar Na⁺/H⁺ antiport proteins, their molecular identity remains elusive. Our studies showing that intracellular compartmentalization of Na⁺ in yeast is mediated by a novel Na⁺/H⁺ exchanger should be particularly useful in elucidating salt tolerance mechanisms in higher plants. In yeast, the role of the vacuole as storage organelles and for cytosolic ion regulation has been generally established; however evidence for internal compartmentation of Na⁺ is scarce. In one report (46), yeast vacuoles were shown to accumulate Li⁺ to an estimated concentration of 0.24 M, relative to a cytosolic concentration of 0.06 M, after incubation in 200 mM LiCl. A similar phenomenon has been reported in Na⁺-grown yeast (47). In addition to detoxification of the cytosol, salt accumulation in the vacuole may play an important role in osmoregulation; in keeping with this proposal, mutants with reduced vacuoles rapidly lose viability in response to osmotic shock (49). Salt in the vacuole has been proposed to serve as a "cheap osmoticum" (5), and future studies will be aimed at manipulating Na⁺ transport systems, such as NHX1, to study osmotic and salt stress at a cellular level.