Functional study of the Saccharomyces cerevisiae Nha1p C-terminus


  • Olga Kinclová,

    1. Department of Membrane Transport, Institute of Physiology CzAcadSci, 14220 Prague 4, Czech Republic.
    2. Laboratory of Microbiology and Genetics, UPRES A-7010-CNRS, Université Louis Pasteur, 67083 Strasbourg, France.
    Search for more papers by this author
  • José Ramos,

    1. Department of Microbiology, Escuela Técnica Superior de Ingenieros Agrónomos y Montes, 14080 Córdoba, Spain.
    Search for more papers by this author
  • Serge Potier,

    1. Laboratory of Microbiology and Genetics, UPRES A-7010-CNRS, Université Louis Pasteur, 67083 Strasbourg, France.
    Search for more papers by this author
  • Hana Sychrová

    Corresponding author
    1. Department of Membrane Transport, Institute of Physiology CzAcadSci, 14220 Prague 4, Czech Republic.
    Search for more papers by this author


Saccharomyces cerevisiae cells possess an alkali metal cation antiporter encoded by the NHA1 gene. Nha1p is unique in the family of yeast Na+/H+ antiporters on account of its broad substrate specificity (Na+, Li+, K+) and its long C-terminus (56% of the whole protein). In order to study the role of the C-terminus in Nha1p function, we constructed a series of 13 truncated NHA1 versions ranging from the complete one (2958 nucleotides, 985 amino acids) down to the shortest version (1416 nucleotides, 472 amino acids), with only 41 amino acid residues after the last putative transmembrane domain. Truncated NHA1 versions were expressed in an S. cerevisiae alkali metal cation-sensitive strain (B31; ena1–4Δ nha1Δ). We found that the entire Nha1p C-terminus domain is not necessary for either the proper localization of the antiporter in the plasma membrane or the transport of all four substrates (we identified rubidium as the fourth Nha1p substrate). Partial truncation of the C-terminus of about 70 terminal amino acids improves the tolerance of cells to Na+, Li+ and Rb+ compared with cells expressing the complete Nha1p. The presence of the neighbouring part of the C-terminus (amino acids 883–928), rich in aspartate and glutamate residues, is necessary for the maintenance of maximum Nha1p activity towards sodium and lithium. In the case of potassium, the participation of the long C-terminus in the regulation of intracellular potassium content is demonstrated. We also present evidence that the Nha1p C-terminus is involved in the cell response to sudden changes in environmental osmolarity.


Maintenance of intracellular ion homeostasis is one of the crucial requisites for the survival of any cell. Intracellular cations and anions participate in many physiological functions, including the adaptation of cells to changes in the environment. In yeast cells, potassium is the major cytoplasmic cation involved, among other things, in the regulation of cell volume and intracellular pH. Thus, yeast cells normally spend energy to accumulate large amounts of K+ (Rodríguez-Navarro, 2000). On the other hand, a high internal concentration of sodium cations is toxic, and the Na+ surplus must be efficiently eliminated from cells. In order to maintain an optimum cytoplasmic concentration of all cations, yeasts, similar to other types of cells, possess many transport systems that mediate the cation efflux or influx with different substrate specificities and diverse mechanisms.

In the model yeast Saccharomyces cerevisiae, the high potassium content corresponds to a steady state between simultaneous influx and efflux across the plasma membrane (Ortega and Rodríguez-Navarro, 1985). K+ is supposed to be taken up very efficiently by K+/H+ symporters and probably also via inward-rectifying channels (Rodríguez-Navarro, 2000), and possibly also by a K+/H+ antiporter (Lapathitis and Kotyk, 1998). The efflux is mediated by at least three types of transport systems: K+/H+ antiporters (Ramírez et al., 1996; Bañuelos et al., 1998), an outward-rectifying channel (Zhou et al., 1995) and an ATPase (Haro et al., 1991; Bañuelos et al., 1998).

The internal content of Na+ in S. cerevisiae cells reflects its extracellular concentration. It is almost negligible in cells growing on standard media containing very low amounts of NaCl, but it can become very high (and toxic) when NaCl is added (salt stress). Sodium probably enters the cell as a low-affinity substrate of several systems transporting cations, mainly those involved in potassium uptake (Rodríguez-Navarro, 2000). To maintain the necessary low cytosolic Na+ concentration upon salt stress, sodium must be rapidly expelled from cells or sequestered inside the cells. The main system extruding Na+ (and also Li+) ions from S. cerevisiae cells is a plasma membrane P-type ATPase encoded by a tandem array of nearly identical genes (ENA1–4/PMR2A–E;Haro et al., 1991; Wieland et al., 1995). In addition to the Na+-ATPase, two Na+/H+ antiporters, Nhx1p and Nha1p, participate in the resistance of S. cerevisiae cells to high external concentrations of salts. The endosomal/prevacuolar Na+/H+ exchanger Nhx1p mediates intracellular Na+ sequestration (Nass et al., 1997; Nass and Rao, 1998) and contributes to cell resistance to a hyperosmotic sorbitol shock (Nass and Rao, 1999). From its sequence, Nhx1 antiporter belongs to a subgroup of intracellular Na+/H+ exchangers related to the mammalian NHE family. The second univalent cation/proton antiporter, Nha1p (Prior et al., 1996), contributes to cell cation homeostasis by mediating the efflux of Na+, Li+ and K+ (Bañuelos et al., 1998). The overexpression of the ENA1 or the NHA1 gene in an ena1–4Δ nha1Δ strain showed complementary action of both systems in the maintenance of intracellular steady-state concentrations of K+ and Na+. The Nha1 antiporter is responsible for cell growth on high concentrations of KCl and NaCl at acidic external pH values, and the Ena1 ATPase is necessary at higher pH values (Bañuelos et al., 1998). In contrast to the complementary action at the protein activity level, the expression of both systems is different. ENA1 expression is highly regulated, and it can be induced by external Na+, Li+ or high pHout values (Garciadeblas et al., 1993). It is also modulated by several components of the regulatory network, including calcineurin (Nakamura et al., 1993; Mendoza et al., 1994) or the HOG pathway (Marquez and Serrano, 1996). NHA1 expression, on the other hand, seems not to be inducible by salts, pH changes or osmotic shocks, and NHA1 transcription is constitutive and very low (Bañuelos et al., 1998).

The NHA1 gene product is a protein of 985 amino acids that belongs to the ever-increasing family of specific fungal Na+/H+ exchangers. Until now, four other fungal genes encoding homologous antiporters from three different yeast species have been cloned and characterized. The Nha1p homologue responsible for Na+ and Li+ tolerance in the fission yeast Schizosaccharomyces pombe is encoded by the sod2 gene (Jia et al., 1992). In highly osmotolerant Zygosaccharomyces rouxii, two almost identical genes, ZSOD2 and ZSOD22, were found (Watanabe et al., 1995; Iwaki et al., 1998). More recently, the CNH1 gene was isolated from Candida albicans, and its product was characterized as a functional Na+/H+ antiporter (Soong et al., 2000). All five yeast antiporters contain a very short hydrophilic N-terminus, probably 12 highly conserved hydrophobic transmembrane domains and a less conserved hydrophilic C-terminus. The S. cerevisiae Nha1p is the longest member of the family; its extremely long C-terminal domain consists of 554 amino acid residues, i.e. 56.2% of the whole protein (985 amino acids). All data available up to now suggest specific functions for the Nha1p C-terminus. (i) In contrast to homologous yeast antiporters of S. pombe, Z. rouxii and C. albicans, which are supposed to transport only sodium and lithium, Nha1p also mediates the efflux of potassium (Bañuelos et al., 1998), so the long C-terminus could be involved in particular substrate specificity. (ii) The originally cloned NHA1 was truncated of about 300 nucleotides at its 3′ end (Prior et al., 1996), and the sodium tolerance of cells harbouring this truncated version (Nha1p shortened by 97 amino acids at its C-terminus) was higher compared with cells with the complete Nha1p (Bañuelos et al., 1998). Thus, the C-terminus could be important for the regulation of transporter activity, similar to that observed for many membrane transporters, e.g. yeast amino acid permeases Gap1p (Stanbrough and Magasanik, 1995) and Bap2p (Grauslund et al., 1995) or yeast and plant plasma membrane H+-ATPases (Portillo et al., 1989; Palmgren et al., 1991). Moreover, the long C-terminus could also be involved in some signalling similar to the C-termini of yeast glucose sensors Snf3p and Rgt2p (Özcan et al., 1998; Lafuente et al., 2000) or the Ssy1 protein, which participates in sensing extracellular amino acids via its long N-terminus (Klasson et al., 1999). (iii) As the presence of Nha1p in cells was found to influence intracellular pH (Sychrováet al., 1999), the Nha1p C-terminus could participate in the regulation of cytoplasmic pH, perhaps in a similar way to the C-terminal domain of the human NHE1 Na+/H+ antiporter (Wakabayashi et al., 1992).

To understand the role of the C-terminus in Nha1p function, we constructed a series of step-by-step truncated NHA1 versions, ranging from the complete gene (2955 nucleotides, 985 amino acids) down to the shortest version (1416 nucleotides, 472 amino acids) with only 41 amino acid residues after the last putative transmembrane domain. Truncated NHA1 versions were expressed in an S. cerevisiae alkali metal cation-sensitive strain (B31; ena1–4Δ nha1Δ), and the localization of the encoded proteins, together with their activity and substrate specificity, was tested. Using green fluorescent protein (GFP) tagging, we could localize both the complete ‘native’ Nha1p and its shortest version lacking most of the C-terminus in the plasma membrane. By testing cation tolerance and cation efflux activity, we demonstrated that the Nha1p C-terminus is not important for substrate specificity but is involved in discrimination among substrates and influences efflux activity. We also present evidence for participation of the Nha1p C-terminus in the cell response to hyperosmotic shock.


Construction of the NHA1 truncated versions

Figure 1 shows a proposed model for the Nha1p secondary structure with designated sites of truncations constructed for this study. The originally cloned NHA1 was truncated of about 300 nucleotides, and its corresponding protein (888 amino acids) seemed to have a higher transport activity than the complete Nha1p (985 amino acids) (Bañuelos et al., 1998). First, a series of plasmids containing the original complete gene and six NHA1 versions truncated by steps of 300 nucleotides was constructed to determine whether longer truncations would result in an even greater increase in activity. The length of the encoded proteins was 985, 883, 774, 680, 568 and 472 amino acids respectively (Fig. 1). Among the initial set of plasmids with truncated NHA1 genes, we also included two versions encoding proteins 945 and 915 amino acids long, respectively, i.e. ending between the originally cloned truncated antiporter (888 amino acids) and the complete native protein (985 amino acids). All truncated genes, together with the complete one, were cloned behind the NHA1 own promoter in the multicopy YEp352 vector, and the resulting plasmids were transformed into the S. cerevisiae B31 strain (ena1–4Δ nha1Δ), which is very sensitive to alkali metal cations because of the lack of the plasma membrane Na+/H+ antiporter and Na+-ATPases.

Figure 1.

Proposed topological model of the Nha1p. Grey residues with numbers correspond to the last amino acid residues of constructed truncated Nha1p versions.

Plasma membrane localization of Nha1p

The Nha1p was localized by Western blot analysis to the plasma membrane of S. cerevisiae (Bañuelos et al., 1998). In order to verify the localization of truncated Nha1p, the shortest truncated and the complete genes were fused downstream with the sequence encoding GFP. The attachment of GFP to protein C-termini did not change the protein activity; the expression of NHA1–GFP fused versions in S. cerevisiae B31 resulted in the same level of salt tolerance as in cells transformed with plasmids containing untagged NHA1 genes (see below). As judged by fluorescence (Fig. 2B) and confocal (Fig. 2C) microscopies, the complete Nha1p appeared to have a peripheral localization that was in agreement with the presumed plasma membrane occurrence. In the case of the Nha1p version lacking almost the whole C-terminus (protein length 472 amino acids), the same peripheral localization was observed (Fig. 3), demonstrating that the C-terminus truncations did not affect the Nha1p localization within the cell.

Figure 2.

Nomarski (A), fluorescence (B) and confocal (C) micrographs of B31 cells expressing the complete Nha1 antiporter tagged with GFP.

Figure 3.

Nomarski (A) and fluorescence (B) micrographs of B31 cells expressing the truncated (472 amino acids) Nha1 antiporter tagged with GFP.

The role of the C-terminus in Nha1p substrate specificity

The role of the long C-terminus in the alkali metal cation tolerance was studied in drop tests on plates containing media with increasing amounts of Na+, Li+ or K+ and adjusted to different pH values. The level of tolerance of B31 cells harbouring different NHA1 constructs depended on the external pH, being highest at acidic pH, moderate at pH around 6.0 and negligible at pH > 7.0 (e.g. cells harbouring the complete Nha1p could grow on 1100 mM NaCl at pHout 3.5, on 800 mM NaCl at pHout 5.5 and, at pHout 7.0, they tolerated only 130 mM NaCl). This observation corresponds with the nature of the cation/H+ antiporter function, as the higher external pH reduces the (electrochemical) proton gradient that serves as a driving force for cation efflux.

The presence of any of the truncated versions in cells provided a high potassium tolerance, up to 1600 mM KCl in the external medium at pH 5.5, whereas cells lacking any Nha1p could tolerate only about 600 mM KCl in the medium (Fig. 4, black bars). This observation indicated that the extremely long C-terminus is not involved in Nha1p potassium substrate specificity.

Figure 4.

Na+, Li+ and K+ tolerance of B31 cells harbouring truncated Nha1p versions at pHout 5.5. Tolerance of B31 cells transformed with plasmids containing truncated NHA1 genes to alkali metal cations was tested on plates containing increasing amounts of NaCl (0–1000 mM; grey bars), KCl (0–1800 mM; black bars) and LiCl (0–50 mM; dotted bars) at pH 5.5. Columns show maximum concentrations of salts tolerated by cells with particular truncated Nha1p. Numbers below the columns indicate the protein length (amino acids). C, control B31 cells containing an empty vector.

On the other hand, the C-terminus contains regions important for sodium and lithium tolerances. In the case of Li+, C-terminus truncation of about 70 amino acids improved the tolerance of cells significantly (up to 35 mM LiCl at pH 5.5) in comparison with cells expressing the complete Nha1p (20 mM LiCl) (Fig. 4, dotted bars). With longer truncations (Nha1p length 883–472 amino acids), the cell tolerance to lithium decreased, and the presence of the two shortest Nha1p versions (472 and 568 amino acids long respectively) rendered cells almost as sensitive to Li+ cations as control cells (Fig. 4). A somewhat similar effect of the C-terminus truncations was observed for sodium tolerance. Cells with Nha1 proteins truncated at the C-terminus of about 70–100 amino acids (antiporter length 915 and 883 amino acids respectively) could grow on 900 mM NaCl at pH 5.5, whereas cells harbouring the complete Nha1p tolerated only 800 mM NaCl in the medium (Fig. 4, grey bars). In contrast to lithium, large truncations did not greatly affect the sodium tolerance. Cells containing the shortest Nha1p truncated versions tolerated more than twice as high amounts of external NaCl as control cells lacking any Nha1p (Fig. 4).

Although rubidium is usually used as a potassium analogue in the kinetic characterization of uptake systems (Ramos and Rodríguez-Navarro, 1986), it has not been tested so far as a substrate for Nha1p. Growth of cells harbouring different Nha1p constructs was estimated on plates supplemented with RbCl, and Nha1p was found to be responsible for cell tolerance to a high external concentration of Rb+ cations. With regard to differences in rubidium tolerance levels among truncated Nha1p versions, similar results as shown for sodium in Fig. 4 were obtained. Cells containing 915-amino-acid-long Nha1p (truncation of 70 amino acids) were more tolerant to RbCl than cells with entire Nha1p (1500 versus 1400 mM RbCl respectively), whereas cells harbouring Nha1p with the longest deletion tolerated about 1200 mM RbCl, and the control cells lacking any Nha1p could grow only if the external concentration of Rb+ cations was lower than 800 mM. Drop test experiments showed that different Nha1p truncated versions provide cells with diverse tolerances to Rb+ but not to K+ thus, the Nha1p is probably able to distinguish between these two cations, and the C-terminus is involved in this substrate discrimination. From the above experiments, we can conclude that the Nha1p is a transport system with broad substrate specificity for alkali metal cations, and its C-terminus is involved in substrate discrimination.

Importance of the Nha1p region between amino acids 920–928 for Li+ tolerance

The test of cell tolerance to alkali metal cations revealed a significant increase in Li+ tolerance if the Nha1p was shortened from 945 amino acids to 915 amino acids (Fig. 4). This part of the protein, rich in aspartate and glutamate residues, contains a putative α-helix structure (Fig. 1). To elucidate which amino acid residues in this region are indispensable to attain the highest activity, we constructed a second series of truncated Nha1 proteins ending after amino acid residues 940, 936, 928, 923 and 920. When the Li+ tolerance of cells harbouring Nha1 proteins with lengths ranging from 915 to 945 amino acids was compared, the highest tolerance was observed only for cells with Nha1p ending between amino acids 928 and 920 (Fig. 5). Cells with shorter or longer Nha1p were more Li+ sensitive. Results obtained in a similar test with Na+ cations indicated that the maximum sodium tolerance is also reached when the Nha1p is truncated down to 928 amino acids, but the maximum tolerance plateau was longer, ending with a truncation down to 883 amino acids (data not shown).

Figure 5.

Phenotype of cells with Nha1 proteins ending in the region 915–945 amino acids. Drop test shows the tolerance of B31 cells transformed with plasmids encoding seven truncated versions of Nha1p to 30 mM LiCl.

Alkali metal cation efflux mediated by truncated Nha1 antiporters

To verify that the tolerance of cells observed in drop tests resulted from the transport activity of truncated Nha1 proteins, sodium, lithium and potassium effluxes were measured in B31 cells harbouring different NHA1 constructs. For Na+ and Li+ efflux measurements, cells were preincubated in YNB media adjusted to pH 7.0 and supplemented with 100 mM NaCl or 50 mM LiCl for 60 min. At pHout 7.0, the Nha1 antiporter is almost inactive because of the absence of a proton gradient across the plasma membrane; thus, B31 cells expressing different Nha1p versions were preloaded with approximately the same amounts of Na+ (Li+) as control cells lacking the Nha1 efflux system. Preloaded cells were resuspended in a Na+(Li+)-free incubation buffer of pH 5.5, containing 10 mM KCl to prevent Na+ (Li+) reuptake (the uptake of all alkali metal cations is mediated via a Trk1–Trk2 transport system that has the highest affinity towards potassium; thus, the excess of K+ in the external medium saturates the Trk transport capacity, and the other substrates are not transported; Rodríguez-Navarro, 2000). The loss of cations from cells was followed for 40 min. In the control strain lacking Nha1p, the intracellular concentrations of sodium or lithium did not change during the experiment (Fig. 6A and B), so the observed effluxes of cations were mediated exclusively by Nha1 proteins. In accordance with phenotypes observed on plates, the Nha1p truncated at the C-terminus of about 100 amino acids mediated a higher efflux of sodium and lithium in comparison with the complete Nha1p (Fig. 6A and B). The Na+ efflux rate from cells with the shortest Nha1p (472 amino acids) was low, corresponding to the higher sensitivity of cells with this Nha1p to extracellular sodium (Figs 4 and 6A). Surprisingly, a low but significant efflux of Li+ was observed in cells with the shortest Nha1p (472 amino acids), although these cells did not show an increased Li+ tolerance in drop test experiments (Figs 4 and 6B).

Figure 6.

Na+ (A), Li+ (B) and K+ (C) loss from B31 cells at pHout 5.5. Cells transformed with YEp352 as a control (◆), pNHA1-985 (●), pNHA1-883 (▴) and pNHA1-472 (▪) were preincubated with 100 mM NaCl (or 50 mM LiCl) at pH 7.0 for 60 min, transferred to the incubation buffer at pH 5.5 and the internal content of Na+ or Li+ was followed for 40 min. The loss of K+ was followed for 45 min (no preloading necessary).

As S. cerevisiae cells, like all other organisms, maintain a stable potassium content independent of external potassium concentration in growth media (Ramos et al., 1985), no K+ preloading was necessary. Cells were transferred directly to the K+-free incubation buffer of pH 5.5 supplemented with 10 mM RbCl to avoid potassium reuptake, and the K+ loss from cells was followed for 45 min. The results obtained (Fig. 6C) were consistent with the drop test observations. All three tested Nha1p versions (the complete one and the truncated versions of 883 and 472 amino acids) mediated the efflux of potassium with similar initial rates (Fig. 6C). However, in cells expressing Nha1p without a long C-terminus (472 amino acids), the intracellular K+ concentration at the end of the experiment was higher compared with cells expressing longer Nha1 proteins (Fig. 6C). When the K+ efflux was followed for 120 min, the difference between the complete and the shortest versions increased. In cells with the shortest Nha1p, about 35% of the intracellular potassium was lost during the first 50 min. Then the efflux almost stopped, and the intracellular concentration was stabilized (60% of the initial amount) for the next 70 min. The efflux of K+ from cells harbouring the complete Nha1p was continuous, leaving only about 40% of the initial K+ amount in cells after 120 min.

Interestingly, we did not detect any significant loss of potassium cations from control cells lacking Nha1p, so the other systems supposed to mediate the export of K+ (Kha1 antiporter and Tok1 channel) are probably not active under our conditions.

Cell response to alkalinization of intracellular pH

It has already been shown that Nha1p mediates immediate high efflux of potassium upon alkalinization of intracellular pH and, hence, could contribute to the buffering of cytosolic pH using the outward gradient of K+ to drive in protons (Bañuelos et al., 1998). We wanted to know whether the Nha1p C-terminus was involved in this process. B31 cells without Nha1p or containing the complete (985 amino acids) and the shortest (472 amino acids) Nha1p versions were resuspended in the incubation buffer at pH 8.0 supplemented with 20 mM RbCl (to prevent K+ reuptake), and the efflux of potassium was followed for 120 min (Fig. 7A). Although both versions of Nha1p mediated an efficient efflux of potassium at pH 5.5 (Fig. 6C), only a slight loss of K+ was found in cells harbouring the complete Nha1p during the first 20 min of the experiment at pH 8.0, and no significant efflux was observed at this pH in control cells or in cells harbouring the Nha1p without the C-terminus (Fig. 7A). It is worth mentioning that this experiment was carried out in the absence of external K+, so the proton-motive force (absent at pHout 8.0) was not required to drive the process. Nevertheless, potassium did not leave cells following its concentration gradient. The same experiment was performed in the buffer at pH 8.0 supplemented with NH4Cl (20 mM) to increase the intracellular pH. The alkalinization of cytosolic pH resulted in a rapid loss of potassium from cells with both Nha1p versions (Fig. 7B). The initial rates of efflux were similar but, in the case of cells containing the shortest version of Nha1p, the intracellular concentration of potassium was stabilized at a level about 50% higher compared with cells with the complete Nha1p (Fig. 7B). In these conditions, a slow loss of potassium was also observed in control cells, indicating the activity of another system involved in K+ efflux. In the experiments described above, the influence on cell survival of changes in external and internal pH or potassium depletion was also estimated. Aliquots of cells were collected at the beginning and end of the experiments, diluted and spread in parallel on YNB plates without KCl or supplemented with 100 mM KCl (to exclude the lethal effect of insufficient K+ in media for potassium-depleted cells). When the number of colonies was estimated after 2 days, we concluded that the higher extracellular pH (pHout 8.0 versus 5.5), the alkalinization of the cytosolic pH (pHout 8.0 versus pHout 8.0 + NH4Cl) or the partial potassium depletion (pHout 8.0 + NH4Cl, cells with Nha1p versus control cells) did not cause higher mortality of cells (data not shown).

Figure 7.

K+ loss from B31 cells at pHout 8.0. Cells transformed with YEp352 as a control (◆), pNHA1-985 (●) and pNHA1-472 (▪) were grown in YNB media, transferred to the incubation buffer at pH 8.0 without (A) or supplemented with 20 mM NH4Cl (B). Changes in the internal K+ content were followed for 120 min.

Results of alkali metal cation efflux measurements in cells expressing different truncated versions of Nha1p corresponded to phenotypes observed on plates, confirming that the role of Nha1p in cell tolerance to alkali metal cations is mediated via the Nha1p efflux activity. The long Nha1p C-terminus is important for the rate of sodium and lithium efflux, and it influences the intracellular potassium content.

High-osmolarity stress and the Nha1p C-terminus

Potassium is one of the most abundant osmotic components in cells, and it is also supposed to be involved in cell response to osmotic shock (Rodríguez-Navarro, 2000). As the experiments described above showed that elimination of the Nha1p C-terminus influenced intracellular potassium levels, we studied the possible involvement of the Nha1p C-terminus in the recovery of cells from an osmotic shock caused by a solute different from salts (sorbitol).

The effect of a hyperosmotic shock on the survival of cells harbouring either the complete Nha1p or the shortest truncated version was estimated both in liquid media and on plates. Growing cells were transferred to the fresh YNB media either without (control) or with sorbitol (35% final concentration). Then the growth of cultures was followed turbidimetrically at 600 nm for 24 h. Under non-stress conditions, cultures continued to grow in a constant exponential manner and at the same rate (data not shown). In the media with sorbitol, the optical density (OD) of the cell culture containing complete Nha1p decreased significantly during the first 6 h, and then the cells started to grow again (Fig. 8A, first 15 h of experiment). In contrast, the OD of the culture with cells harbouring the short Nha1p almost did not change during this time period, and slow growth was restored only after about 10 h (Fig. 8A). But once the cells harbouring the truncated Nha1p adapted and again entered the exponential phase of growth, the rate of growth was the same as that of cells with the entire Nha1p (data not shown). To estimate the number of cells surviving the sorbitol shock, aliquots of cell cultures were withdrawn, diluted and spread on YP plates. As shown in Fig. 8B, the survival of cells expressing Nha1p without the C-terminus was notably affected compared with cells with the complete Nha1p. More than 65% of cells with the short Nha1p were dead after 3 h in the presence of sorbitol, whereas only 40% of cells harbouring the complete Nha1p did not survive the high osmolarity of the medium (Fig. 8B).

Figure 8.

Effect of high osmotic shock on B31 cells. Cells transformed with pNHA1-985 (●, black bars) and pNHA1-472 (▪, dotted bars) were grown in YNB media to the early exponential phase and transferred to fresh YNB media supplemented with sorbitol (final concentration 35%).

A. After transfer to a sorbitol-containing medium, changes in cell culture OD at 600 nm were followed for 15 h.

B. Percentage of cells surviving the osmotic shock. Aliquots of cell cultures were diluted and plated on YP plates after different exposures to sorbitol stress (0–10 h). The survival of cells was scored after 2 days. C, control corresponding to the number of cells in the culture without sorbitol.

If the ability of cells to grow and form colonies on plates containing high amounts of sorbitol was tested, a similar negative effect of the C-terminus truncation was observed (Table 1). The mortality of cells with Nha1p only 472 amino acids long was about 50% and 90% on plates containing 30% and 35% of sorbitol, respectively, whereas the number of dead cells containing the complete Nha1p was about 40% and 80% under the same conditions. These results indicated that the long Nha1p C-terminus could play a role in the response of cells to changes in environmental osmolarity.

Table 1. Survival of cells after osmotic shock on YNB plates supplemented with sorbitol.
S. cerevisiae B31
(ena1–4Δ nha1Δ)
Surviving cells (%)
30% sorbitol35% sorbitol


In the present work, we analysed the properties of the Nha1 antiporter and its truncated variants after expression in S. cerevisiae cells deficient in alkali metal cation efflux transporters. The Nha1p is highly homologous to other yeast Na+/H+ antiporters with regard its sequence but not to its substrate specificity. Although the N-terminus, transmembrane domains and connecting loops share more than 60% identity with Z. rouxii and C. albicans homologues (Prior et al., 1996; Soong et al., 2000), the Nha1p C-terminus is much longer compared with other members of the family. Concerning its substrate specificity, the Nha1p is the only yeast antiporter transporting potassium and rubidium besides sodium and lithium. Also in the families of Na+/H+ exchangers from other organisms, such as bacteria, plants and mammalian cells, none has so far been shown to transport other alkali metal cations efficiently besides sodium and lithium. The role of most of the known Na+/H+ antiporters is supposed to be an effective elimination of excess toxic cations from cells at the expense of the gradient of protons across the plasma membrane or, as in the case of some mammalian NHE antiporters, regulation of the intracellular pH using the gradient of sodium (Wakabayashi et al., 1997).

We found that the entire Nha1p C-terminus domain is dispensable (i) for proper folding and secretion of the protein to the plasma membrane; (ii) for alkali metal cation transport activity; and (iii) for substrate specificity. Even the Nha1p with the C-terminus shortened by more than 90% (only 41 of 556 amino acids left) mediates the efflux of all four substrates, and its presence in cells mostly results in increased tolerance to salts. Only in the case of the smallest substrate, lithium, are cells harbouring the shortest Nha1p sensitive to external Li+, although Nha1p-mediated lithium efflux can be measured in these cells. Our results showed clearly that the unique substrate specificity of Nha1p is not connected to specific regions in the C-terminus, so the protein domains necessary for recognition, binding and transport of K+ and Rb+ must be located somewhere within the conserved part of the protein.

On the other hand, we found a C-terminus domain that ‘inhibits’ the maximum Nha1p activity. If the last 70 amino acids are cut off, the Nha1p shows a higher activity towards three of its substrates. The increase in activity (upon gradual truncations in the 920- to 940-amino-acid region) is continuous rather than abrupt; thus, it is probably the steric conformation of the C-terminal structure that modulates the transport activity. It is possible that the last autoinhibiting domain of the C-terminus (amino acids 945–985) can be post-translationally modified or perhaps even cleaved to counteract its negative effect. One of the well-known possibilities of transporter activation is the phosphorylation of its C-terminus, e.g. yeast H+-ATPase (Eraso and Portillo, 1994) or the mammalian NHE1 antiporter (Wakabayashi et al., 1997), and there are potential phosphorylation sites in the last autoinhibiting domain of Nha1p. The preceding domain (amino acids 883–928) important for maximum transport activity of three substrates (Li+, Na+, Rb+) involves a putative α-helix and several possible phosphorylation sites. Three of them (amino acids 894–897, 895–898 and 918–921) could be recognized by the casein kinase II but, although the yeast strain deficient in this kinase activity is hypersensitive to alkali metal cations, it does not show an altered cation efflux (Nadal et al., 1999), indicating that phosphorylation of the C-terminus by casein kinase II is probably not involved in Nha1p activation.

In the case of the fourth substrate of Nha1p, potassium, gradual truncations had no activating or inhibiting effect, and we could demonstrate that the C-terminus is not important for the initial rate of transport or for cell tolerance towards high external concentrations of KCl. In contrast, the activity of Nha1p without its C-terminus retains higher potassium content in cells. It has been observed that the lowest limit of intracellular content of K+, below which cells do not lose more potassium under disadvantageous conditions, is about 230 nmol mg−1 (Ramos and Rodríguez-Navarro, 1986), and our data showed the same amount of intracellular potassium at which the efflux mediated by the complete Nha1p stopped upon intracellular alkalinization (Fig. 7B). However, when the Nha1p C-terminus was absent, the level of intracellular K+ content stayed higher under both normal (pHout 5.5; Fig. 6C) and stressed conditions (pHout 8.0 + NH4Cl; Fig. 7B). As the observed initial rate of K+ efflux upon the alkalinization of cytosolic pH was the same in cells with both complete or truncated Nha1p (Fig. 7B), it seems probable that either the activity of truncated Nha1p is regulated differently or not all of the intracellular K+ is available as substrate for the truncated Nha1p.

A continuous potassium efflux is very important for yeast cells, and a K+/H+ exchange at the plasma membrane level was proposed to serve as a safety valve contributing to the buffering of cytosolic pH (Ortega and Rodríguez-Navarro, 1985; Bañuelos et al., 1998). This is probably the main function explaining the observation of negligible K+ efflux at pHout 8.0 and a rapid massive efflux detected when the internal pH increased simultaneously (Fig. 7A and B). Nevertheless, the Nha1p-mediated efflux of potassium upon alkalinization of the cytosol is not the main protection system, as the number of control cells without Nha1p surviving the alkalinization of cytosolic pH was the same as that of cells harbouring both versions of Nha1p.

The higher content of potassium retained in cells upon efflux mediated by truncated Nha1p raised the question whether these cells were not more resistant to a hyperosmotic shock. After an osmotic shock, first a rapidly operating rescue system is necessary for cell survival. It also includes the mobilization of intracellular, mainly vacuolar, stocks of ions and water. This rescue mechanism is supposed to be constitutively expressed in cells and should be independent of the adaptation mechanisms that prepare the cell for growth resumption (Hohmann, 1997). Our results showed that the lack of Nha1p C-terminus rendered exponentially growing cells more sensitive to the osmotic shock than cells with the complete Nha1p. Moreover, cells with a truncated Nha1 version took a longer time to adapt to sorbitol shock, and culture growth resumed at a slower rate (Fig. 8A). Also, the number of cells surviving the osmotic shock was higher when the complete Nha1p was expressed (Fig. 8B and Table 1). On the other hand, (i) cells with both (truncated and complete) versions of Nha1p grew in a similar manner in the absence of osmotic stress and also after a stress adaptation period; (ii) some preliminary results indicated that the same amount of glycerol was produced by both types of cells growing in hyperosmotic media (data not shown); and (iii) the osmosensitivity of mutant strains defective in the adaptation response, hog1Δ (Norbeck et al., 1996) and imp2Δ (Masson and Ramotar, 1998), cannot be suppressed by Nha1p overexpression (H. Sychrová, unpublished results). All these results indicate that Nha1p and its long C-terminus play a role in the immediate response of cells to hyperosmotic shock as part of the rapid rescue mechanism. In this way, Nha1p could have a similar role at the level of the cell plasma membrane of exponentially growing cells to the Nhx1 Na+/H+ exchanger in the endosomal/prevacuolar compartments of post-diauxic/stationary phase cells (Nass and Rao, 1999). Nhx1p is believed to become activated in response to cell shrinkage upon osmotic stress, and its presence in cells is important for the recovery of sudden exposure to hyperosmotic media (Nass and Rao, 1999). Further studies will be necessary to elucidate the nature and molecular mechanism of the role of the Nha1 antiporter and its C-terminus in the osmosensing and immediate response to hyperosmotic shock.

It seems likely that Nha1p has a dual function in cells, ensuring the low intracellular level of toxic cations and regulating the intracellular content of potassium that is connected with the maintenance of cell volume and intracellular pH. We can conclude that, in Nha1p, (i) the long C-terminus is dispensable for broad substrate specificity and basal transport activity, although it could play a role in the antiporter activation when it is necessary to eliminate toxic alkali metal cations from cells; and (ii) the long C-terminus is involved in the regulation of intracellular potassium content and in the immediate response of cells to hyperosmotic shock.

Altogether, our results show that, in addition to its detoxicating function, Nha1p with its long C-terminus is also involved in some regulatory functions that modify the performance of cells under different conditions.

Experimental procedures

Strains, media and growth conditions

The S. cerevisiae B31 strain (ena1Δ::HIS3::ena4Δnha1Δ::LEU2;Bañuelos et al., 1998) was used in this work for complementation and phenotype characterization. The S. cerevisiae wild-type FL100 haploid strain genomic DNA [prepared as described by Hoffman and Winston (1987)] was used as a template for polymerase chain reactions (PCRs). Yeast cells were grown aerobically at 30°C in standard YP or YNB media supplemented with 2% glucose and appropriate supplements when necessary.

DNA manipulations

DNA manipulations, transformation of yeast and Escherichia coli, and growth of bacterial cultures were performed according to standard protocols (Sambrook et al., 1989).

DNA sequencing and sequence analysis

DNA sequencing was carried out using the ThermoSequenase radiolabelled terminator cycle sequencing kit (Amersham Life Science). The DNA inserts were sequenced on both strands directly from YEp352 using, first, universal primers complementary to the vector polylinker, followed by primers corresponding to the internal sequence of the NHA1 gene. lasergene99 (DNASTAR) was used for DNA and protein sequence analyses.

Plasmid constructions

To construct C-terminus-deleted versions of the Nha1p, first, a sequence corresponding to SmaI–STOP–SphI was added by PCR (Eppendorf Mastercycler 5330; Pfu DNA polymerase, Stratagene) to the 3′ end of the complete and originally cloned truncated NHA1 genes (Table 2, oligonucleotides XSS1 and XSS2; for restriction map, see Bañuelos et al., 1998). Both NHA1 genes with their own promoter region (615 nucleotides) were cloned as XbaI–SphI fragments into the YEp352 (Hill et al., 1986) in which the SmaI site of the polylinker was destroyed (XmaI cut and filling). The resulting plasmids, pNHA1-985 and pNHA1-883, were used for subsequent constructions. Genomic DNA was used as a template for 11 PCR amplifications each time with the primer CLA1 (hybridizing upstream of the unique ClaI site in the NHA1 open reading frame) and a downstream primer corresponding to the desired truncation and introducing the SmaI site in frame to the 3′ end (Table 2). Amplified DNA fragments, ranging from 0.75 kb for oligonucleotide S472 to 2.18 kb for oligonucleotide S945, were digested with restriction enzymes ClaI and SmaI and used to replace a 2.30 kb ClaI–SmaI fragment of the pNHA1-985. The resulting plasmids harbouring truncated NHA1 gene versions are listed in Table 2; their numbers correspond to the length (amino acids) of the encoded Nha1p.

Table 2. Oligonucleotides and corresponding created plasmids.
  • a

    . Numbers correspond to the length (amino acids) of resulting truncated Nha1p.

  • SphI–STOP–SmaI sequence double underlined; SmaI site underlined.


Plasmids pNHA1-985GFP and pNHA1-472GFP were constructed by cloning the 4.88 kb and 3.35 kb SacI–SmaI fragments from pNHA1-985 and pNHA1-472, respectively, in frame into the corresponding sites in the pGRU1 (a gift from B. Daignan-Fornier).

Microscopy analysis

Cells from the exponential phase of growth in YNB were viewed with the Leica DMRXA microscope using the fluorescein isothiocyanate (FITC) filter. For confocal microscopy, native cells were observed in the Bio-Rad MRC600 confocal laser-scanning microscope with an oil immersion objective (60×, NA = 1.4).

Salt tolerance determination

The maximum concentrations of LiCl, NaCl, KCl and RbCl tolerated by yeast cells were estimated on solid YNB media supplemented with salts by spotting 3 µl of serial 10-fold dilutions of saturated cultures. To adjust the pH of the media to 3.5, tartaric acid was added after autoclaving. Media with pH 5.5 and 7.0 were supplemented with either 20 mM MES or MOPS, respectively, and the required pH was adjusted with NaOH. Growth was recorded for 5 days.

Cation efflux experiments

For cation loss measurements, cells were grown in YNB media to OD600≈ 0.3, harvested and washed. To measure the efflux of Li+ and Na+, cells were resuspended in the fresh YNB medium adjusted to pH 7.0 with NH4OH and supplemented with 50 mM LiCl or 100 mM NaCl. After 60 min of preloading, the cells were harvested, quickly washed and resuspended in the incubation buffer consisting of 20 mM MES adjusted to pH 5.5 with Ca(OH)2 or 20 mM HEPES adjusted to pH 8.0 with Ca(OH)2. In both cases, buffers contained 0.1 mM MgCl2, 2% glucose and were supplemented with KCl, RbCl or NH4Cl as indicated in the text. For K+ efflux measurements, no preloading was necessary. Cells were resuspended directly in the incubation buffers. Samples of cells were withdrawn at various time intervals, filtered through Millipore membrane filters, washed rapidly with a 20 mM MgCl2 solution, acid extracted and analysed by atomic absorption spectrophotometry as described previously (Rodríguez-Navarro and Ramos, 1984; Haro et al., 1991). Each experiment was repeated at least three times, and a representative experiment is shown.

Osmotic stress experiments

Osmotic shock experiments were carried out in liquid media containing 35% sorbitol. Cells were grown in YNB media to the early exponential phase (OD600≈ 0.2), and 15 ml of the growing cell culture was added to 35 ml of YNB media without (control) or with 50% sorbitol. The OD600 was monitored for 24 h. To estimate the numbers of cells surviving the shock, culture samples were diluted in appropriate YNB media, spread on YP plates, and the colonies were counted after incubation at 30°C for 2 days.

To test the survival of cells after the osmotic shock on solid media, appropriately diluted cell cultures were spread on YNB plates supplemented as indicated in the text.

Osmotic stress experiments were carried out in triplicate; the deviation among parallel results was always < 5%.


We thank Dr M. A. Bañuelos for helpful comments and stimulating discussions, Professor A. Kotyk for critical reading of the manuscript, Dr F. Calero for help with cation efflux experiments, Dr B. Daignan-Fornier for the pGRU1 plasmid, Dr L. Kubínová for assistance with confocal microscopy, and Drs A. Balguerie, M. Bonneu and Professor M. Aigle for help with fluorescence microscopy. This work was supported by grants from the Grant Agency of the Academy of Sciences of the Czech Republic (A5011005) and from the Dirección General de Investigación Científica y Técnica, Spain (PB98-1036).