A screening for high copy suppressors of the sit4 hal3 synthetically lethal phenotype reveals a role for the yeast Nha1 antiporter in cell cycle regulation.

A screening for multicopy suppressors of the G(1)/S blockage of a conditional sit4 hal3 mutant yielded the NHA1 gene, encoding a Na(+),K(+)/H(+) antiporter, composed of a transmembrane domain and a large carboxyl-terminal tail, which has been related to cation detoxification processes. Expression of either the powerful Saccharomyces cerevisiae Ena1 Na(+)/H(+)-ATPase or the Schizosaccharomyces pombe Sod2 Na(+)/H(+) antiporter, although increasing tolerance to sodium, was unable to mimic the Nha1 function in the cell cycle. Mutation of the conserved Asp residues Asp(266)-Asp(267) selectively abolished Na(+) efflux without modifying K(+) efflux and did not affect the capacity of Nha1 to relieve the G(1) blockage. Mutagenesis analysis revealed that the region near the carboxyl-terminal end of Nha1 comprising residues 800-948 is dispensable for sodium detoxification but necessary for transport of K(+) cations. Therefore, this portion of the protein contains structural elements that selectively modulate Nha1 antiporter functions. This region is also required for Nha1 to function in the cell cycle. However, expression of the closely related Cnh1 antiporter from Candida albicans, which also contains a long carboxyl-terminal extension, although allowing efficient K(+) transport does not relieve cell cycle blockage. This indicates that although the determinants for Nha1-mediated regulation of potassium transport and the cell cycle map very closely in the protein, most probably the function of Nha1 on cell cycle is independent of its ability to extrude potassium cations.

The Saccharomyces cerevisiae gene SIT4 encodes a type 2Arelated Ser/Thr protein phosphatase (1) that is a homologue to Schizosaccharomyces pombe Ppe1 (2), Drosophila PPV (3), and human PP6 phosphatases (4). This phosphatase plays an important role in cell cycle regulation because it is required in late G 1 for progression into S phase (5,6) and controls expression of SWI4, CLN1, and CLN2 in a pathway that is additive to that of CLN3. SIT4 is required for both efficient DNA synthesis and bud emergence (7). Consequently, cells lacking Sit4 are either inviable or exhibit a slow growth phenotype (5).
The cell growth defect of sit4⌬ mutants is largely overcome by high copy expression of the SIS2/HAL3 gene (8). HAL3 encodes a protein with a very acidic carboxyl-terminal region that was also identified by its ability to confer tolerance to high levels of sodium and lithium to yeast cells (9) and that has close relatives in plants (10). Although deletion of HAL3 causes no evident growth defect, the gene is essential in the absence of SIT4 function presumably because cells become fully arrested at the G 1 /S transition (8). Therefore, Sis2/Hal3 appears to be a component of the cell cycle regulatory machinery. Recent work in our laboratory has revealed that Hal3 is a negative regulatory subunit of another Ser/Thr protein phosphatase, Ppz1 (11) and that, in fact, the function of Hal3 in cell cycle regulation is mediated by the Ppz1 phosphatase (12), which plays an opposite role to that of Sit4 in regulating the cell cycle.
The observation that sit4 and hal3 mutations display synthetic lethality prompted us to design a screen to search for novel components of the cell cycle regulatory machinery. To this end we constructed a conditional sit4 hal3 mutant strain and transformed these cells with multicopy genomic libraries to identify genes that allowed growth under restrictive conditions. We describe here the isolation of the NHA1 gene, encoding a 985-residue Na ϩ ,K ϩ /H ϩ antiporter (13). This protein has been previously shown to be involved in maintaining Na ϩ and K ϩ fluxes and buffering cytosolic pH (14,15). We present evidence that the function of Nha1 in the cell cycle requires structural determinants present at its carboxyl-terminal region and that this region is able to modulate the diverse functions of the protein.
Recombinant DNA Techniques, Plasmids, and Gene Disruptions-E.
* This work was supported by Grants PB98-0565-C4-02 and PB98-1036 (to J. A. and J. R., respectively) from Dirección General de Investigación Científica y Técnica, Spain and by "Ajut de Suport als Grups de Recerca de Catalunya" 1999SGR-00100 (to J. A) from the Generalitat de Catalunya. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  coli and S. cerevisiae cells were transformed using standard techniques as described previously (11). Standard recombinant DNA techniques were carried out as described previously (17).
Construction of plasmid pJQ10, which allows overexpression of Ena1 from the PGK1 promoter, can be found in Benito et al. (18). Plasmid pCSCY, carrying a truncated version of NHA1, was described by Bañ uelos et al. (14). Expression of the S. pombe Sod2 antiporter was accomplished as follows. The intronic region of sod2 ϩ was removed by sequential PCR (19) so that artificial BamHI sites were added at the ends of the open reading frame. The amplification fragment was digested with BamHI and cloned into the BglII site of plasmid pDB20LBglII (20) to allow expression from the powerful ADH1 promoter, yielding plasmid pDBSod2. Plasmid VHS5A/Ura consists of a YEplac195 vector (21) containing a XbaI-SphI 4.3-kbp insert starting from a XbaI site located at position Ϫ672 from the ATG codon of the NHA1 open reading frame and includes 187 nucleotides of the YEp13 vector (from the BamHI cloning site to the vector SphI site). Plasmid VHS5A/Leu consists of a YEplac181 vector containing the same insert. Expression of the Candida albicans antiporter Cnh1 in budding yeast was accomplished as follows. A 4-kbp genomic fragment cloned into plasmid pBK-CMV (Stratagene) and comprising the entire CNH1 open reading frame plus about 0.5 and 0.95 kbp of 5Ј-and 3Ј-flanking sequences, respectively (22), was recovered by digestion with SacI and XbaI and cloned into these sites of plasmid YEplac195. This construct was then introduced into strains JC002 and B31.
Disruption of the NHA1 gene in strain JA121 to generate strain JC074 was carried out by replacing a 1.0-kbp SnaBI-HincII fragment with a blunt-ended BglII 3.1-kbp fragment from plasmid YEp13 containing the LEU2 marker. This disruption is identical to that introduced at the NHA1 locus in strain B31 (14).
Construction of Strain JC002 and Screening for Suppressors-To produce strain JC002 the HAL3 promoter was replaced in strain JA110 (sit4⌬), following the short flanking homology gene replacement strategy, by a cassette containing the tetO promoter element, which can be strongly repressed by doxycycline. The substitution cassette was obtained by PCR from plasmid pCM224 (24) using oligonucleotides that included sequences Ϫ300 to Ϫ260 and ϩ1 to ϩ42 of the HAL3 gene. Positive clones were selected by growth on geneticin (G418) plates, and the correct replacement was verified by PCR analysis. As expected, strain JC002 failed to produce macroscopic colonies upon plating on media containing 20 g/ml doxycycline.
For suppressor screening, strain JC002 was transformed with two different genomic libraries constructed in YEp13 and YEp24 and plated on CM synthetic medium (lacking uracil or leucine as needed) in the presence of 20 g/ml doxycycline. Macroscopic colonies observed after 96 h of incubation were considered as positives and recovered. Clones bearing genes SIT4 or HAL3 in the plasmid-borne genomic insert were discarded at this stage by analytical PCR with appropriate oligonucleotides. Plasmids were recovered from the rest of the clones and subjected to restriction mapping with EcoRI to identify identical inserts. The fragment of the yeast genome contained in each independent clone was identified by DNA sequencing using specific oligonucleotides priming to regions nearby the cloning site of the vector, and comparison of the sequence was obtained with the Saccharomyces Genome Database.
Mutagenesis of NHA1-Mutated versions of NHA1 were generated by sequential PCR (19)  with BglII and replacing this region with PCR-amplified DNA fragments containing the indicated deletions. The mutagenized fragment was reintroduced directly into plasmid VHS5A/Ura. Changes to Asn of residues Asp 241 (GAT 3 AAT) and Asp 266 -Asp 267 (GAT 3 AAC, GAC 3 AAC) were introduced in a 0.53-kbp PstI-PstI fragment, and this fragment was used to directly replace the wild type fragment in VHS5A/Ura. Other Techniques-The budding index was determined by microscopic counting, and the DNA content was monitored by flow cytometry essentially as described by Clotet et al. (12). Growth on plates (drop tests) was assessed as described by Posas et al. (25). Tolerance to cations was determined in liquid cultures or plates as described previously (26). Transport of sodium and potassium was measured as described by Bañ uelos et al. (14) and De Nadal et al. (27).

High Copy Expression of the NHA1 Gene Suppress the Lethal
Phenotype of a sit4 hal3 Mutant-Strain JC002 was designed to produce a model for conditional blockage in the cell cycle at the G 1 /S transition and, therefore, to serve as a starting point for the screening depicted in Fig. 1. Consequently, we first tested the terminal phenotype of strain JC002 upon exposure to doxycycline by measuring both the DNA content and the budding index of the cultures. As shown in Fig. 1, after 12 h of growth in the presence of doxycycline, most cells have unbudded, rounded morphology and a haploid content of DNA, indicating that most of them were in G 1 phase. Therefore, strain JC002 was considered a suitable model for our studies.
The transformation of strain JC002 with multicopy genomic yeast libraries yielded about 200 clones able to develop macroscopic colonies within 2-4 days of incubation on synthetic medium in the presence of doxycycline. These clones were denominated VHS (for viable hal3 sit4). Characterization of the molecular nature of the inserts present in the different plasmids revealed a number of genes responsible for suppression that will be described in detail elsewhere. In several cases, inserts contained genes previously related to cell cycle regulation at the G 1 /S transition, such as CLN3 and BCK2. However, the presence of the NHA1 gene in three of those VHS clones drew immediately our attention because, although previously related to cation detoxification, Nha1 had never been directly related to cell cycle regulation. Subcloning of the genomic inserts revealed that the suppressor effect could be indeed attributed to the NHA1 gene (Fig. 2).
Because the G 1 /S blockage of strain JC002 is based on the effect of doxycycline and because NHA1 encodes a permease, we considered the possibility that expression of the antiporter might affect doxycycline transport, giving rise to artifactual effects. To test this possibility, we transformed the diploid strain JC173 (heterozygous for the sit4 hal3 mutations) with the VHS5A/Ura construct, a high copy plasmid bearing the NHA1 gene, and induced sporulation. Tetrad analysis revealed that, when bearing the NHA1 gene, viable sit4 hal3 haploids could be recovered, thus confirming the notion that high levels of Nha1 have a positive effect on cell cycle G 1 /S transition. This idea was reinforced by the observation that the transformation of sit4 cells with the same plasmid partially alleviated the slow growth phenotype of these cells (not shown) and accelerated the recovery from an ␣-factor G 1 arrest as deduced from monitoring the budding index and DNA content of the cultures (Fig. 3). However, in contrast with the sit4 hal3 phenotype, a haploid sit4 nha1 mutant (strain JC094) is viable as established by tetrad analysis of the heterozygous diploid strain JC074 (not shown). Alternative Transporters Involved in Sodium Detoxification Cannot Replace Nha1-NHA1 was initially identified as a gene able, when in high copy number, to increase tolerance to sodium and lithium cations. We decided to test whether overexpression of the Na ϩ -ATPase Ena1, which represents the most important element for sodium and lithium efflux in budding yeast, could mimic the effect of expression of the Nha1 on JC002 cells. Fig. 4A shows that a high level of Ena1 results in a remarkable increase in lithium tolerance (lithium, a highly toxic analog of sodium, was used here to avoid osmotic effects derived from the use of high concentrations of NaCl). In contrast, no growth was observed when JC002 cells were incubated in the presence of doxycycline. Then we constructed a system to overexpress the Na ϩ /H ϩ antiporter Sod2 from the fission yeast S. pombe. This protein is quite similar to the NH 2 -terminal half of Nha1 (over 35% identity with about 58% conserved residues) and has been shown to function in budding yeast (28,29). Fig. 4B shows that the ability of our Sod2 expression system to increase sodium tolerance was even higher than that conferred by high copy NHA1 expression. However, overexpression of Sod2 was completely unable to suppress the conditional sit4 hal3 phenotype of strain JC002, indicating that this alternative Na ϩ /H ϩ antiporter cannot re-place Nha1. High copy expression of the S. cerevisiae KHA1 K ϩ /H ϩ antiporter (30) in JC002 cells also failed to mimic the effect of Nha1 on cell growth (data not shown).
The Carboxyl-terminal Region of Nha1 Is Required for Function in the Cell Cycle-Despite the results described above, the possibility that the role of Nha1 in sodium detoxification might be related to the observed phenotype in the cell cycle still remained. To evaluate this possibility we transformed strain JC002 with plasmid pCSCY, which carries in high copy a version of NHA1 encoding a protein that lacks 97 residues of the carboxyl terminus (14). This gene product has been characterized as being even more active than the wild type protein in cation detoxification (Ref. 14 and this work). However, to our surprise, pCSCY was unable to allow growth of strain JC002 in the presence of doxycycline (Fig. 5B). To confirm this observation we introduced, by site-directed mutagenesis, stop codons at positions 923, 948, and 979 of the NHA1 open reading frame and tested the effect of these constructs in strain JC002 for cell growth and in strain B31 (ena1-4⌬ nha1⌬) for sodium tolerance. As shown in Fig. 5B, all constructs tested conferred a level of sodium tolerance comparable with that of the wild type gene (or even higher). Versions of the antiporter lacking from residues 948 or 979 to the end of the protein allowed growth of strain JC002 under nonpermissive conditions, but the version lacking from residue 923 to the end of the protein did not, indicating that the region between residues 923 and 947 contains elements essential for this function. To establish the limits of the carboxyl-terminal region required for the Nha1 antiporter to play a positive role in the cell cycle, several deletions starting from residue 642 were made. The results presented in Fig. 5C clearly show that forms of Nha1 lacking residues from 642 to 828 cannot produce viable JC002 cells in the presence of doxycycline. When residues 642-799 are deleted, the protein still retains its function, although in this case growth is clearly less vigorous than that observed for the full protein or for a version with deletions from residues 642 to 700. None of the deletions tested resulted in significant changes in the capacity of the antiporter to confer sodium tolerance.
We have also constructed a version of Nha1 carboxyl terminally tagged with a triple hemagglutinin epitope, and we observed that this construct was able to both increase sodium tolerance in B31 cells and allow growth of JC002 cells. However, a version carrying a large NH 2 -terminal deletion (from residues 18 to 345) that removed the first six predicted transmembrane segments but entirely retained the carboxyl-terminal half was not functional at all (data not shown). This amino terminally deleted version was recovered only in the soluble cell fraction. Moreover, we fused the carboxyl-terminal half of Nha1 (from residue 441, right after the last predicted transmembrane domain, to the stop codon) to the end of the open reading frame of Ste2, a seven transmembrane domain phero-mone receptor. This hybrid protein, as the entire Nha1, was recovered in nonsoluble fractions but did not suppress the G 1 /S blockage of JC002 cells. Immunoblot analysis indicated that all these constructs were expressed at similar levels.
The Capacity of Nha1 to Detoxify Sodium Ions Is Not Required for Its Role in the Cell Cycle-The data shown above clearly indicated that the capacity for sodium detoxification was not sufficient to ensure the role of Nha1 in the cell cycle. To further test whether sodium detoxification was a necessary event, we took advantage of previous knowledge on S. pombe Sod2 function. We constructed versions of Nha1 in which the conserved Asp residues Asp 241 and Asp 266 -Asp 267 were mutated to Asn. These mutations were known to reduce or suppress the transporter activity of Sod2 (31). As shown in Fig. 6B, mutation of the conserved Asp 241 did not affect the ability of Nha1 to confer Na ϩ tolerance to B31 cells, whereas the change of both Asp 266 -Asp 267 completely abolished this property. Remarkably, when the same constructs were tested for growth of JC002 in the presence of doxycycline, cells carrying the D241N version grew rather poorly, whereas the Asp 266 -Asp 267 construct grew similarly to those bearing the wild type antiporter (Fig. 6A). All these results clearly indicate that the ability of Nha1 to detoxify sodium cations is not required to fulfill its role in the cell cycle.
It has been recently reported that, in addition to sodium, FIG. 3. Effect of high copy expression of NHA1 on sit4 cells. Wild type (q) or sit4 cells (E) bearing an empty plasmid as well as sit4 cells carrying the VHS5A/Ura plasmid (, YEpNHA1) were arrested in G 1 by incubation with ␣-factor. The pheromone was washed out, and entry into cell cycle was monitored by both measuring the budding index (right) and the DNA content (left).

FIG. 4. Expression of S. cerevisiae
Ena1 Na ؉ -ATPase or S. pombe Sod2 Na ؉ /H ؉ antiporter does not mimic the effect of Nha1 on cell cycle. A, YEp24 plasmid carrying no insert (YEp ), plasmid pJQ10 (ENA1) overexpressing the Na ϩ -ATPase, or the NHA5A/Ura construct were introduced in JC002 cells. Positive clones were tested for growth on CM plates lacking uracil and containing doxycycline (20 g/ml) or 200 mM LiCl after 3 days. B, left, strain JC002 was transformed with an empty YEplac181 plasmid (YEp ), the VHS5A/Leu construct (NHA1), or plasmid pDBSod2 (Sod2), and positive clones were streaked on CM plates lacking leucine in the absence (upper half) or the presence (lower half) of doxycycline (dox). Right, the above-mentioned constructs were introduced into G19 cells (ena1-ena4⌬) and tested for growth in YPD medium adjusted to pH 5.5 in the presence of 0.4 M NaCl. Relative growth was calculated as the ratio between growth in the presence and growth in the absence of added salt and expressed as a percentage. Data are means Ϯ S.E. from five to seven independent clones.
Nha1 can also extrude potassium cations and that the lack of functional Nha1 results in cells unable to grow at high potassium levels (14). We tested the growth of B31 cells bearing these constructs in the presence of high concentrations of KCl (Fig. 6B). As expected, high copy number wild type NHA1 improved growth under this condition. Mutation of Asp 241 significantly reduced this effect. In contrast, the Asp 266 -Asp 267 version behaved virtually as the wild type protein. This was surprising and pointed to the possibility that the Asp 266 -Asp 267 mutation might affect the ability of Nha1 to extrude sodium without modifying potassium transport. To assess this possibility, we measured the transport of these cations in cells carrying the mentioned constructs. Fig. 7 shows that, as previously documented, strain B31 has virtually no sodium or potassium transport and that high copy expression of Nha1 restores the ability of the cell to extrude these cations. Interestingly, expression of the Asp 266 -Asp 267 version of the antiporter does not restore extrusion of sodium, but these cells extrude potassium as efficiently as cells carrying wild type Nha1. In contrast, cells carrying the Asp 241 version display a sodium transport relatively close to that of wild type Nha1 (note the slopes at initial time points), but its ability to extrude potassium cations is rather poor.
The observation that a given mutation in Nha1 might specifically affect sodium transport without altering potassium efflux prompted us to investigate how diverse carboxyl terminally altered versions of Nha1 described in Fig. 5 might affect growth of strain B31 in high potassium. Strain B31 cannot grow on plates containing medium-high concentrations of potassium or rubidium cations. We observed that whereas these cells, which carried in high copy the wild type NHA1 gene, displayed no growth defect at 1.5 M KCl or 1 M RbCl, versions containing a stop codon at position 923 as well as those lacking residues 642-799 and 642-828 grew very poorly under these conditions (data not shown). To more accurately monitor this phenomenon, potassium efflux was measured. As shown in Fig.  8, B31 cells expressing those Nha1 versions unable to allow growth on high potassium displayed a virtually null (⌬923 and ⌬642-828) or severally impaired (⌬642-799) potassium efflux. Remarkably, cells carrying Nha1 versions with defects in potassium transport perfectly corresponded with those unable to allow growth of JC002 cells under nonpermissive conditions (compare Figs. 5 and 8). These observations indicate that the function of Nha1 on cell cycle maps very closely in the Nha1 protein to the ability of the antiporter to regulate transport of potassium ions and proves that changes at the carboxyl-terminal moiety of the protein affect these functions.
The C. albicans Cnh1 Antiporter Mimics Nha1 Function in Potassium Transport but Not in the Cell Cycle-The observation that a specific region of the carboxyl-terminal half of Nha1 was important for both potassium transport and cell cycle regulation prompted us to consider the possibility that both functions could be linked and specifically whether the ability to extrude potassium cations was the reason for Nha1 to function in the cell cycle. To test this possibility we expressed in budding yeast the recently described Na ϩ /H ϩ antiporter Cnh1 from C. albicans that, in contrast with S. pombe Sod2, also contains a long carboxyl-terminal tail that shows a significant level of identity to that of Nha1 (see Ref. 22 and Fig. 9A). It is worth noting that, although expression of the Cnh1 antiporter in budding yeast was shown to increase sodium tolerance (22), it was unknown whether Cnh1 could also transport potassium cations. As shown in Fig. 9B, high copy expression of the C. albicans antiporter allowed S. cerevisiae B31 cells to grow in the presence of 1 M KCl similarly to cells carrying the Nha1 antiporter, and this correlated with a highly efficient efflux of potassium (Fig. 9C) almost identical to that conferred by expression of budding yeast Nha1. Therefore, we show that Cnh1, in addition to Na ϩ , can also extrude potassium cations. However, the heterologous antiporter failed in supporting growth of JC002 cells in the presence of doxycycline (Fig. 9D), indicating that C. albicans Cnh1 contains structural requirements sufficient for mimicking Nha1 function in potassium transport but not in cell cycle regulation. DISCUSSION We describe here the construction of the conditional sit4 hal3 strain JC002, which reproduces a blockage in the G 1 /S transition of the cell cycle, and the use of this strain to screen for genes able to overcome such blockage. In addition to SIT4 and HAL3, this screening yielded several genes known to perform key roles in cell cycle regulation, such as CLN3, BCK2, or SWI4, thus proving its usefulness as a tool for cell cycle studies.
Among the different genes identified, NHA1 showed a rather strong phenotype, allowing a relatively vigorous growth of strain JC002 under restrictive conditions. NHA1 encodes an antiporter that was formerly identified in high copy number as able to increase sodium tolerance in yeast cells (13). More recently it has been documented that this gene product mediates sodium and potassium efflux (14) and that it may be involved in the regulation of intracellular pH (15). Our work demonstrates that Nha1 positively functions in the cell cycle and that this ability is not related to its capacity to extrude sodium cations. For instance, we show that neither the powerful S. cerevisiae Ena1 Na ϩ -ATPase nor a functional fission yeast Sod2 antiporter, both able to actively detoxify sodium cations, can replace Nha1. Moreover, mutation in Nha1 of the conserved residues Asp 266 -Asp 267 fully abolished sodium efflux but did not affect the ability of the antiporter to relieve the cell cycle blockage. An equivalent mutation (Asp 310 -Asp 311 ) has been shown to abolish the ability of the recently reported Cnh1 Na ϩ /H ϩ antiporter from C. albicans to increase sodium tolerance when overexpressed in S. cerevisiae (22). Interestingly, the Asp 266 -Asp 267 mutation generated a version of Nha1 fully able to transport potassium ions, whereas mutation of the conserved Asp 241 residue did not alter sodium tolerance and had little effect on sodium efflux but significantly reduced potassium efflux. As far as we know, this is the first report of mutations in a Na ϩ /H ϩ antiporter that allow discrimination between sodium and potassium cations. It should be noted that although several mutations have been recently described to selectively affect cation uptake in the plant Hkt1 potassium transporter (32,33), this protein belongs to a very different family, which is related to the Trk yeast transporters (for review, see Ref. 34). It has been proposed, on both theoretical and experimental grounds, that a conserved structure in bacterial and yeast antiporters involving the conserved Asp 241 and Asp 266 -Asp 267 residues may serve to coordinate transported cations (31,35). Our results demonstrate that, although this may be the case for sodium, the model does not apply to the transport of potassium, which might require alternative binding structural determinants. This finding is particularly relevant because, although influx of potassium has been relatively well characterized in fungi, the mechanisms for extrusion of this cation have yet to be clarified (34).
The observation that Sod2 cannot play the role of Nha1 in the cell cycle was intriguing because the fission yeast protein is almost 40% identical to the NH 2 -terminal half of Nha1 and the two proteins have in common many structural features. However, Sod2 lacks the carboxyl-terminal extension found in the budding yeast antiporter, and our data clearly show that a region in the vicinity of the carboxyl-terminal end of Nha1 spanning from residues 800 to 948, although dispensable for sodium tolerance, is required for the antiporter to alleviate cell cycle blockage. These findings clearly show that the carboxyl terminus of Nha1 contains structural determinants that modulate specific functions of the antiporter.
On the basis that the mutated versions of Nha1 that cannot extrude potassium cations do not support growth of strain JC002 in the presence of doxycycline, it could be hypothesized that the role of Nha1 in the cell cycle relies on the ability of the antiporter to extrude potassium cations. However, this is most probably not the case as deduced from the following observations. First, a positive effect of potassium extrusion on the cell cycle would be difficult to reconcile with previous experimental evidence that associates increased entry of potassium with acceleration of the cell cycle (36). This scenario, in contrast, would fit with the identification of HAL4/SAT4 and HAL5 in our screening for suppressors of the G 1 /S blockage in strain JC002. 2 These genes encode partially redundant protein kinases that activate the Trk1-Trk2 potassium transporters and, therefore, increase influx of potassium (37). On the other hand, a sit4 mutant does not show an enhanced sensitivity to high potassium (data not shown), suggesting that the cell cycle defect of these cells is not due to excessive accumulation of potassium.
The function of Nha1 as cation/proton antiporter is based in the existence of an electrochemical gradient of protons across the membrane. This is exemplified by the observation that at pH 7.0 high copy expression of Nha1 fails to increase tolerance of B31 cells to sodium or potassium cations (14). Consequently, if the function of Nha1 in the cell cycle would depend on its ability to extrude potassium (or any other cation) by electroneutral exchange with protons, this function would also be largely abolished when pH approaches neutrality. However, we have observed that the ability of Nha1 to allow growth of a sit4 hal3 strain is maintained even at pH 7.0 (data not shown), supporting the notion that the effect on the cell cycle does not depend on the function of Nha1 as cation/proton antiporter. Finally, if potassium efflux due to Nha1 would drive the release of the G 1 /S blockage, expression of an equally efficient extrusion system should reproduce the effect of Nha1 on the cell cycle. In contrast, we illustrate here (Fig. 9) that the C. albicans antiporter, although fully able to restore potassium transport with similar potency to that of Nha1, could not suppress FIG. 9. Expression of the C. albicans Cnh1 antiporter functionally replaces Nha1 on potassium transport but not on suppression of G 1 /S blockage. A, pairwise comparison using the Clustal W program of the carboxyl-terminal regions of the Nha1 and Cnh1 proteins. Black boxes indicate identical amino acids, and open boxes denote conserved changes. Dashes indicate gaps introduced to maximize identities. The region of Nha1 identified as necessary for potassium efflux and cell cycle function is highlighted by a dotted line. B, strain B31 was transformed with YEplac195, YEpNHA1, and YEpCNH1, and its sensitivity to 1 M KCl was tested in YPD medium at pH 5.5. Data are means Ϯ S.E. of six independent clones. C, potassium efflux in cells expressing the C. albicans Cnh1 antiporter. The above-mentioned strains were tested as described in the legend of Fig. 7. q, YEp vector; E, YEpNHA1; , YEpCNH1. Data are means Ϯ S.E. from three independent experiments. D, the indicated constructs were introduced into JC002 cells, and growth was scored in the absence or the presence of doxycycline (dox) after 3 days. the G 1 /S blockage of strain JC002. In conclusion, most probably cell cycle and potassium extrusion regulation are independent functions of Nha1 that map closely within the carboxyl-terminal moiety of the protein.
From our data, the important role of the carboxyl terminus of Nha1 in modulating the functions of the protein is evident. Comparison of the Cnh1 and Nha1 carboxyl-terminal sequences (Fig. 9A) reveals that Cnh1 is about 39 residues shorter than Nha1. The absence of this tail, however, is not the reason that explains why Cnh1 fails to function in the cell cycle because a version of Nha1 lacking almost exactly this region (NHA1⌬948) is still functional. Within residues 800 -948 of Nha1, the region from amino acids 815 to 912 present the highest degree of divergence with the corresponding region of Cnh1 (see Fig. 9A), and, therefore, it is a likely candidate to account for the observed differences. A carboxyl-terminal tail is also found in the related Zsod2 and Zsod22 Na ϩ /H ϩ antiporters from Zygosaccharomyces rouxii (38,39), although in these cases the carboxyl-terminal extension is even shorter than that of Cnh1 and, as a consequence, does not contain sequences that appear to be important in Nha1 for potassium efflux and cell cycle regulation.
Computational analysis indicates that, with high probability, the carboxyl-terminal region of Nha1 is cytosolic (40,41). As it has been postulated and, in some cases, experimentally determined for mammalian Na ϩ /H ϩ antiporters (42), the carboxyl-terminal region might serve as a regulatory region of the protein, for example modulating the function of the transmembrane domain through covalent modification or interaction with other proteins. This would explain the effect of mutations within the carboxyl-terminal moiety in the transport of potassium. In addition, a possibility worth consideration is that, besides the previously recognized antiporter function of Nha1, the carboxyl-terminal segment of the protein might receive inputs from the transmembrane domain and transmit signals to the inside of the cell that would influence cell cycle progression. We wish to stress that a number of related precedents can be found in the literature. For instance, the ammonium permease Mep2 (but not its homologs Mep1 or Mep3) has been shown to participate in the transmission of the signal that induces pseudohyphal growth in yeast under low ammonium conditions. Interestingly, mutagenesis analysis proved that signaling was independent of the ability of Mep2 to perform its previously recognized function as an ammonium permease (43).