The Transcriptional Response of the Yeast Na+-ATPase ENA1 Gene to Alkaline Stress Involves Three Main Signaling Pathways*>

Adaptive response of the yeast Saccharomyces cerevisiae to environmental alkalinization results in remodeling of gene expression. A key target is the gene ENA1, encoding a Na+-ATPase, whose induction by alkaline pH has been shown to involve calcineurin and the Rim101/Nrg1 pathway. Previous functional analysis of the ENA1 promoter revealed a calcineurin-independent pH responsive region (ARR2, 83 nucleotides). We restrict here this response to a small (42 nucleotides) ARR2 5·-region, named MCIR (minimum calcineurin independent response), which contains a MIG element, able to bind Mig1,2 repressors. High pH-induced response driven from this region was largely abolished in snf1 cells and moderately reduced in a rim101 strain. Cells lacking Mig1 or Mig2 repressors had a near wild type response, but the double mutant presented a high level of expression upon alkaline stress. Deletion of NRG1 (but not of NRG2) resulted in increased expression. Induction from the MCIR region was marginal in a quadruple mutant lacking Nrg1,2 and Mig1,2 repressors. In vitro band shift experiments demonstrated binding of Nrg1 to the 5· end of the ARR2 region. Furthermore, we show that Nrg1 binds in vivo around the MCIR region under standard growth conditions, and that binding is largely abolished after high pH stress. Therefore, the calcineurin-independent response of the ENA1 gene is under the regulation of Rim101 (through Nrg1) and Snf1 (through Nrg1 and Mig2). Accordingly, induction by alkaline stress of the entire ENA1 promoter in a snf1 rim101 mutant in the presence of the calcineurin inhibitor FK506 is completely abolished. Thus, the transcriptional response to alkaline stress of the ENA1 gene integrates three different signaling pathways.

Environmental changes force cells to adapt to the new situation to survive. The budding yeast Saccharomyces cerevisiae grows better at acidic pH and for this microorganism, sudden alkalinization of the medium represents a stress situation because it interferes with the proton gradient essential for uptake of many solutes from the medium, such as nutrients or diverse cations (1,2). Growth at neutral or alkaline external pH requires two cation pumps: the vacuolar membrane H ϩ -ATPase, which is composed of multiple subunits and is essential for acidification of the vacuole (see Ref. 3 for a recent review), and the Na ϩ -ATPase encoded by the ENA1/PMR2 gene (4,5). Consequently, null mutants in many components of the vacuolar ATPase complex or strains lacking a functional ENA1 gene are unable to proliferate even under mild alkaline conditions (6,7).
The Ena1 P-type ATPase also represents a major element in the detoxification of sodium and lithium cations and ena1 mutants are largely defective in cation efflux and extremely sensitive to these toxic ions. ENA1 is the first member of a cluster composed by four to five genes encoding very similar proteins (ENA1-ENA5) (4,5,7) and, although early work suggested post-transcriptional regulation of the ATPase (5), compelling evidence indicates that regulation of the Na ϩ -ATPase function in the cell is mainly based on the control of the expression of the gene. ENA1 is hardly expressed under standard growth conditions, but its expression is rapidly induced after exposure to saline, osmotic or alkaline stress. The complex regulation of ENA1 expression after saline stress (8) has been characterized in some detail in the past. The presence of glucose in the medium represses ENA1 expression by a mechanism that involves the Snf1 kinase and the repressor complex Mig1-Ssn6-Tup1, which targets to a MIG element located at position Ϫ534/Ϫ544 in the ENA1 promoter (9,10). In addition to the release from Mig1-mediated repression, induction of ENA1 by high salt involves at least three different pathways. Activation of the HOG pathway in response to high osmolarity results in phosphorylation and activation of the Hog1 mitogen-activated protein kinase, which migrates to the nucleus and phosphorylates the Sko1 repressor (9). Phosphorylation of Sko1 prevents its binding to the cAMP response element present in the promoter at position Ϫ502/Ϫ509. Activation of the ENA1 promoter by high salt is also greatly influenced by the calcineurin pathway (5,8,(11)(12)(13)(14). Acti-* This work was supported in part by Grants BMC2002-04011-C05-04 and BFU2005-06388-C4-04-BMC (to J. A.) and BFU2004-02855-C0202 (to F. M.) from the Ministerio de Educació n y Ciencia, Spain, and Fondo Europeo de Desarrollo Regional. 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 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental Table S1. 1 Currently holds a fellowship from the Spanish Ministry of Education and Science. 2 Recipient of a fellowship from the Generalitat de Catalunya, Spain. 3 Recipient of a fellowship from the Spanish Ministry of Education and Science. 4 Recipient of a fellowship from FICYT (Fundació n para el Fomento en Asturias de la Investigació n Científica Aplicada y la Tecnología). 5 Recipient of "Ajut de Support a les Activitats dels Grups  vation of calcineurin increases ENA1 transcription through dephosphorylation and activation of the Crz1/Tcn1/Hal8 transcription factor (15)(16)(17), which binds to specific DNA sequences (calcineurin-dependent response elements). In the case of the ENA1 promoter, two such elements are present at positions Ϫ813/ Ϫ821 and Ϫ719/Ϫ727, the downstream element being more important for the transcriptional response under saline stress (18). Finally, a role of the TOR pathway in the regulation of ENA1 expression has been documented. In this case, saline stress would inhibit the TOR pathway, so the Gln3 and Gat1 transcription factors would no longer be retained in the cytoplasm and, upon entry into the nucleus, would activate ENA1 by (presumably) binding to the diverse GATA sequences present in the ATPase gene promoter (19). In contrast to the fairly well defined response of ENA1 to high salt, the mechanisms underlying its activation by alkaline stress are not fully understood. Current evidence suggests that high pH stress triggers a burst of cytoplasmic calcium that results in activation of calcineurin and subsequent Crz1-dependent induction of ENA1 expression (20,21). Functional mapping of the promoter revealed that the calcineurin-dependent pH-responsive region could be restricted to a region, named ARR1, which contains the calcineurin-dependent response element at Ϫ719/Ϫ727, whereas a second pH-responsive but calcium-independent region (ARR2) was located downstream ARR1, spanning from position Ϫ490 to Ϫ573 (21). Induction of ENA1 by high pH is also influenced by the Rim101 pathway. Rim101, a transcription factor initially identified as a positive regulator of gene expression in meiosis and sporulation (22), is a homolog of the Aspergillus nidulans PacC pH response regulator (23). S. cerevisiae rim101 mutants display sensitivity to high pH and decreased expression of several genes induced by high pH (24). However, in contrast to what is known in A. nidulans, there is evidence that Rim101 acts in vivo as a repressor that targets to, at least, two previously known transcriptional repressors: Smp1 and Nrg1 (25). It has been proposed that activation of Rim101 would block the expression of Nrg1, thus releasing the ENA1 promoter from the negative control exerted by the repressor. Consistent with this model, alkaline induction of ENA1 is impaired in a rim101 mutant (21,24,25) and this effect is suppressed by deletion of Nrg1 (25).
The ENA1 promoter contains two sequences that were recognized as consensus binding sites for Nrg1 (CCCTC and CCCCT), at positions Ϫ725/Ϫ729 and Ϫ647/Ϫ651, respectively (25). The former is located in ARR1 and, in fact, overlaps with the downstream calcineurin-dependent response element sequence found in this region. The latter is located between ARR1 and ARR2, and is probably not functional because this segment of the promoter is not responsive to high pH (21). In contrast, whereas functional mapping of the ENA1 promoter revealed that the alkaline response from both ARR1 and ARR2 regions was clearly reduced by mutation of RIM101, no evident Nrg1 site was identified in ARR2 (21). Furthermore, there is evidence that blocking the calcineurin-mediated response in a rim101 mutant does not fully abolish the transcriptional response of ENA1 to alkaline stress and that the ARR2 region of this promoter must contain regulatory elements other than those controlled by Rim101 (21). All these evidence definitely indicates that still unknown, Rim101-and calcineurin-independent regulatory components of the ENA1 response to high pH stress must exist. In this work we characterize a third regulatory pathway relevant in this response and provide a working model that integrates our current knowledge in this field.

EXPERIMENTAL PROCEDURES
Yeast Strains and Growth Media-The S. cerevisiae strains used throughout this work are derived from the DBY746 haploid wild type strain and are listed in Table 1. Yeast cultures were grown at 28°C in YPD medium (10 g/liter yeast extract, 20 g/liter peptone, 20 g/liter dextrose) or complete minimal medium. When indicated, synthetic complete medium lacking appropriate supplements was used to maintain selections for plasmids (26).
Recombinant DNA Techniques-Escherichia coli DH5␣ was used as a host for DNA cloning experiments. Bacterial cells were grown at 37°C in LB medium containing, when needed, 50 g/ml ampicillin for plasmid selection. E. coli cells were transformed by standard treatment with calcium chloride (27). S. cerevisiae cells were transformed by a modification of the lithium acetate method (28). Restriction mapping, DNA ligations, and other recombinant DNA techniques were carried out by standard methods (27). Purification of DNA fragments from complex mixtures, including PCRs and restriction nuclease digests was performed by agarose gel electrophoresis. The appropriate bands were recovered from agarose gel slices using the Agarose Gel DNA Extraction Kit (Roche).
Construction of ␤-Galactosidase Reporters-The following reporter plasmids, containing diverse regions of the ENA1 promoter, were used. Construction of pMRK213 and pKC201 was described in Refs. 21 and 10, respectively. Plasmids pMRK505, pMRK525, and pMRK564 were constructed as follows: the corresponding ENA1 promoter region was amplified by standard PCR using the appropriate primers (see supplemental data) and plasmid pKC201, which contains the entire promoter, as a template. The PCR products were digested by restriction enzymes XhoI and SalI and cloned into the XhoI restriction site of pSLF⌬-178K. This plasmid is a CYC1 promoter-lacZ fusion from which the CYC1 UAS elements have been deleted (32). All inserts were verified by sequencing using the BigDye Terminator version 3.1 Cycle Sequencing Kit (Applied Biosystems).
␤-Galactosidase Activity Assays-Yeast cells were grown to saturation in the appropriate drop-out media overnight and then inoculated in YPD at A 660 0.2. Growth was resumed until A 660 of 0.5 to 0.7 and cultures were then distributed into 1.0-ml aliquots and centrifuged for 5 min at 1620 ϫ g. The supernatant was discarded and cultures were resuspended in the appropriate media. Unless otherwise stated, alkaline stress was provoked by resuspending cells in YPD containing 50 mM TAPS adjusted to pH 8.5 by addition of potassium hydroxide (autoclaving decreases the pH of the medium up to pH 8.0 approximately). Non-induced cells were resuspended in YPD medium adjusted to pH 6.3 after autoclaving. Cell were grown for 60 min and ␤-galactosidase activity assayed as described (33). Saline stress was induced by addition of 0.4 M NaCl. Growth was resumed and cells were collected after 60 min.
Real-time Reverse Transcriptase-PCR-Saturated cultures of the appropriate strains were diluted to A 660 of 0.2 in YPD and grown up to A 660 of 0.5-0.7. Then, two 10-ml aliquots of each culture were centrifuged during 5 min at 1620 ϫ g at room temperature and pellets were resuspended in 10 ml of fresh YPD containing 50 mM TAPS buffered at pH 8.0 (alkaline induction) or YPD (no induction), respectively. After 10 min, cells were harvested by centrifugation for 2 min at 1620 ϫ g at 4°C, washed once with cold water, and centrifuged again to obtain cell pellets that were immediately frozen and stored at Ϫ80°C until RNA purification. When inhibition of calcineurin was desired, FK506 (final concentration of 1.5 g/ml) was added to the medium 1 h prior initiation of the alkaline treatment. The drug was also present in final resuspension media during stress.

ENA1 Response to Alkaline Stress
Total RNA was purified using the RiboPure-Yeast kit (Ambion) following the manufacturer's instructions. RNA quality was assessed by denaturing 0.8% agarose gel electrophoresis and RNA quantification was performed by measuring A 260 in a BioPhotometer (Eppendorf). Real time PCR was performed in a SmartCycler (Cepheid) apparatus, using the QuantiTect SYBR Green reverse transcriptase-PCR kit (Qiagen) and 10 ng of total RNA. For ENA1 amplification, oligonucleotides ENA1fw and ENA1rev were used. Control experiments were carried out by amplifying a fragment of the PHO84 gene using oligonucleotides PHO84fw and PHO84rev. In this case, due to the slower induction of this gene after alkaline stress, cell samples were taken after 30 min of induction. Reverse transcription was performed for 30 min at 50°C, followed by incubation at 95°C for 15 min. Finally, 45 PCR cycles (15 s at 94°C, 30 s at 50°C, and 30 s at 72°C) were carried out.
Expression in E. coli of S. cerevisiae Nrg1-For expression in E. coli, the NRG1 open reading frame (approximately 0.7 kbp) was amplified by PCR using oligonucleotides 5Ј-Nrg1 and 3Ј-Nrg1. The amplification fragment was cleaved with BamHI and XhoI restriction enzymes and cloned into the BamHI-XhoI site of pGEX-4T-1 (Amersham Biosciences) to yield plasmid pGEX-4T1-Nrg1. E. coli BL21-Codon Plus (DE3)-RIL cells (Stratagene) were transformed with plasmid pGEX-4T1-Nrg1 and grown overnight at 37°C in LB medium containing 100 g/ml ampicillin and 34 g/ml chloramphenicol. The culture was then diluted 1/10 in the same medium and cultures were grown to an A 660 of 0.6 -0.8. At this point isopropyl ␤-D-thiogalactopyranoside was added to the medium to a final concentration of 1 mM and cells were induced for 3 h at 37°C. Cells were centrifuged at 1620 ϫ g at 4°C, washed twice, and resuspended in ice-cold lysis buffer (20 ml/liter of culture) that contained 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 2 mM dithiothreitol, 10% glycerol, and 0.1% Triton X-100, plus a protease inhibitor mixture (0.5 mM phenylmethylsulfonyl fluoride, 0.5 mM benzamidine, 1 g/ml pepstatin A, 1 g/ml leupeptin). Cells were harvested by sonication (3 ϫ 15 s), centrifuged at 800 ϫ g at 4°C for 5 min, and the supernatant collected. The GST-Nrg1 fusion protein was purified by incubating the bacterial crude lysate with glutathione-Sepharose 4B beads (Amersham Biosciences), allowing gentle shaking for 1 h at 4°C. GST-Nrg1 was eluted with a 10 mM glutathione solution (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM dithiothreitol).
Electrophoresis Mobility Shift Assays-Probes were labeled with 32 P by PCR using [␣-32 P]dCTP and the appropriate primers, and purified using S-200 HR microspin columns (Amersham Biosciences). Various amounts of GST-Nrg1 were incubated with 10 ng of ␣-32 P-labeled probe (ϳ10,000 cpm) and 1 g of poly(dI-dC) for 30 min at 30°C in a reaction volume of 20 l containing 20 mM HEPES (pH 7.6), 1 mM MgCl 2, 60 mM KCl, 12% glycerol, 6 g of bovine serine albumin, 10 M ZnCl 2 , and 1 mM dithiothreitol. Probe 578, which was added as a nonspecific competitor where indicated, was obtained by amplification from genomic DNA of the Ϫ578/Ϫ640 region of the ENA1 promoter, using primers ena1_pr_5Ј_640 and ena1_pr_3Ј_578. The reaction samples were loaded onto 5% non-denaturing polyacrylamide gels and electrophoresed at 150 V for 3 h in a 22.5 mM Tris base, 22.5 mM boric acid, and 0.63 mM disodium EDTA buffer, adjusted to pH 8.0. The gel was dried and subjected to autoradiography.
Chromatin Immunoprecipitation Assay-The NRG1-HA strain was obtained by transformation of the W303-1A wild type strain with a modification module containing as selectable marker the E. coli kan r gene. The resulting strain has the 3Ј end of NRG1 tagged with 3 hemagglutinin epitope (HA) sequences. The C-terminal modification module was obtained by PCR using as template plasmid pFA6a-3HA-kanMX6 (34) and primers containing the NRG1-specific sequences of the forward primer (O5) chosen to end just upstream of the stop codon, preserving the reading frame of the tag, whereas those of the reverse primer (O6) to end just downstream of the stop codon. To test the Nrg1-HA construct we used immunoblot analysis. Twenty to 40 g of proteins from a NRG1-HA yeast extract were separated by SDS-12% polyacrylamide gel electrophoresis and transferred to enhanced chemiluminescence nitrocellulose membranes (Amersham Biosciences) by electroblotting, which were then incubated with an anti-HA antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Horseradish peroxidase-conjugated protein A was used as secondary reactant. The complex was detected by the ECL detection system (Amersham Biosciences).
Chromatin immunoprecipitation assays were performed essentially as described previously (35,36) with the following modifications. Yeasts cells were grown in 100 ml of low pH (pH 6.2) YPD medium at 28°C to an A 600 nm of 1.0. The cells from one-half (50 ml) of the culture were shifted from low to high pH (pH 8.0) YPD medium for 30 min. Each sample was treated with formaldehyde (final concentration 1%) for 60 min at 20°C, and 2.5 ml of 2.5 M glycine was added to stop the cross-linking reaction. Cells were harvested and disrupted by vortexing in the presence of glass beads, and the lysate was sonicated to generate DNA fragments that ranged in size from 300 to 500 bp. To immunoprecipitate HA-tagged proteins, we incubated the extract overnight at 4°C with anti-HA antibodies, and the extract/antibody mixture was incubated for 3-4 h with protein A-Sepharose beads (Amersham Biosciences). Immunoprecipitates were washed 4-fold with 1 ml each of lysis buffer (50 mM HEPES, pH 7.5, 250 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, and 1 g/ml pepstatin A). The DNA was eluted with elution buffer (100 mM sodium bicarbonate and 1% SDS). After reversal of the formaldehyde-induced cross-links, 1/5000 of input DNA and 1/45 of each immunoprecipitated DNA were used as templates for amplification by PCR (35 cycles), using primers O1 (sense), O2 (sense), O3 (sense), and O4 (antisense). Band intensities were quantified using the 1D Image Analysis Software (Kodak Digital Science).
Growth Tests-Sensitivity of yeast cells to high pH was evaluated by growth on liquid cultures performed in 96-well plates. Two hundred fifty-l cultures at an initial A 660 of 0.01 were grown for 14 -20 h at 28°C in YPD containing 50 mM TAPS buffered at the indicated pH values. Growth was monitored in an iEMS Reader MF apparatus (Labsystems) at 620 nm. DECEMBER 1, 2006 • VOLUME 281 • NUMBER 48

RESULTS
Functional Mapping of the Calcineurin Independent Response of the ENA1 Promoter-We previously demonstrated that the region spanning between nucleotides Ϫ490/Ϫ573 of the ENA1 promoter, named ARR2, was able to elicit a transcriptional response that was not affected by blockage of the calcineurin pathway (21). To further characterize this region, we constructed several reporter plasmids containing different portions of the ARR2 region. As can be observed in Fig. 1, full response to alkaline stress was observed in a construct (pMRK505) containing the Ϫ531/Ϫ573 region. However, the region from Ϫ490/Ϫ531 (pMRK564), which roughly represents the 3Ј-half of ARR2,wascompletelyunabletosustain a transcriptional response to alkaline stress. Therefore, the calcineurin independent response of the ENA1 promoter could be mapped to the 43-nucleotide fragment contained in pMRK505. This region was named MCIR (for minimum calcineurin independent response) and, consequently, the pMRK505 reporter plasmid was used for further characterization of the response.
Effect of Lack of the Snf1 Protein Kinase and the Mig1,2 Transcriptional Repressors-Because of the presence of a MIG element in the MCIR region, we considered it necessary to evaluate the possible role of the Mig1 and Mig2 repressors, which are able to bind to the MIG element. As shown in Fig. 2, mutation of Mig1 or Mig2 barely affects the response to high pH, although slightly higher basal levels (that is, in the absence of stress) are observed. In contrast, mutation of both genes results in markedly increased basal levels and enhanced response to alkaline stress. Interestingly, mutation of the Snf1 protein kinase, which is known to inhibit Mig1, results in a dramatic decrease in response, suggesting that this kinase plays an important role in controlling activation of the ENA1 promoter by acting on the ARR2 region. We also observed that the expression from the MCIR region in a strain lacking both Snf1 and Mig1 is identical to that found in the snf1 mutant. In contrast, the response in the snf1 mig2 mutant was almost as strong as in the mig2 strain. The fact that a strain lacking both Mig1 and Mig2 proteins was still able to promote induction from the MCIR after alkaline shock suggests that additional regulatory elements must be present in this region. Interestingly, mutation of SNF1 markedly decreased the very high expression characteristic of the mig1 mig2 strain. This could be explained if one assumes that Snf1 is controlling an additional negative regulator of ENA1 expression located within the ARR2 region.
Influence of the Rim101 Pathway in the Alkaline Induction from the MCIR-We had previously observed that mutation of RIM101 decreased the response from the ARR2 region of the ENA1 promoter (21). Therefore, it was reasonable to assume that the Rim101 pathway could provide the missing regulatory mechanisms mentioned above. To test this, we constructed strains lacking the RIM101 gene, as well as the Nrg1 and Nrg2 repressors that have been shown to be functionally related to Rim101, and tested the expression in these cells from the reporter plasmid pMRK505. As it can be observed in Fig. 3, cells lacking Rim101 presented a decreased response, although the  decrease was not as strong as in the case of snf1 cells. The absence of Nrg1 resulted in increased basal expression and response to high pH, whereas mutation of NRG2 had little effect and the expression level in the double mutant was only slightly higher than in the nrg1 strain. These results were indicative that Nrg1 was playing a repressor role in the alkaline response from the MCIR region. We observed that mutation of NRG1 strongly increases the basal levels and alkaline response from plasmid pMRK505 in the rim101 mutant, whereas mutation of NRG2 has almost no effect. Therefore, lack of Nrg1, but not Nrg2, abolishes the decrease in response from the MCIR to the high pH observed in the rim101 mutant. Because Nrg1 and Nrg2 are also known to be under regulation of the Snf1 kinase, we constructed cells deficient in all possible combinations of these genes and tested expression from pMRK505. As shown in Fig. 3, mutation of NRG1 drastically counteracts the effect of the snf1 mutation, whereas lack of Nrg2 has no effect. Taken together, these results suggested that both Snf1 and Rim101 may act by controlling the repressor effect of Nrg1, being the role of Nrg2 virtually negligible.
The data collected so far suggest that expression from the MCIR region in response to alkaline pH is mostly controlled through the release from the negative effect of the Mig1,2 and Nrg1 repressors. If so, one would expect that cells lacking all four repressors, Nrg1,2 and Mig1,2, would show a very high basal activity from the pMRK505 reporter and, in addition, further increase induced by alkaline shock would be either eliminated or severely impaired. In agreement with this hypothesis, it can be observed (Fig. 4) that in a quadruple mig1,2 nrg1,2 mutant, expression from the MCIR give rises to higher basal levels than in mig1,2 or nrg1,2 cells and that its activity after alkaline stress increases only marginally (1.3-fold, in comparison with 8.5-fold in the wild type strain). It is worth noting that lack of Mig1,2 results in basal expression levels from pMRK505 that almost double those observed in nrg1,2 cells, suggesting that the Mig1,2 repressors exert a stronger repressor control on this promoter region.

Identification of Nrg1
Binding to the ARR2 Region-In contrast with the evidence described above, canonical sequences previously defined for Nrg1 (25,37) were not identifiable within the ARR2 region. Therefore, we considered it necessary to directly test whether or not Nrg1 was able to interact with this region and the possible functional relevance of such an interaction. To this end, we expressed in E. coli a GST-Nrg1 fusion protein and tested its ability to bind in vitro to the ARR2 region. As can be observed in Fig. 5A, the presence of GST-Nrg1 in the mixture is able to shift the electrophoretic mobility of a radioactively labeled ARR2 fragment and this fragment can be competitively displaced by unlabeled ARR2. To further map the interaction region, we generated additional fragments either fully outside (5Ј) the ARR2 region (fragment 578) or overlapping its 5Ј end (fragments 541, 551, and 561) and tested their ability to bind Nrg1. Our results (Fig. 5B) indicate that the 578 fragment does not bind at all Nrg1, but fragment 561 still does. This suggests that the Nrg1 binding region is located at the 5Ј end of ARR2 and can be mapped between nucleotides Ϫ561/Ϫ574.
To test the biological relevance of the binding of Nrg1 to the ARR2 region, we carried out chromatin immunoprecipitation experiments. For this purpose, we replaced in the wild type strain W303-1A the native NRG1 open reading frame with a HA-tagged version and immunoprecipitated, using anti-HA antibodies, fragmented genomic DNA prepared from this strain. As shown in Fig. 6, from the immunoprecipitated DNA we could specifically amplify fragments surrounding the ARR2 region, indicating that binding was occurring in vivo under standard culture conditions. More importantly, when the same experiment was made in cells subjected to alkaline shock for 30 min, the amount of amplified material was substantially reduced. It is worth noting that this does not appear to be a general, nonspecific effect because in a parallel experiment, using HA-tagged Med8 (which binds to the HXK2 gene), binding was not altered in cells exposed to high pH (not shown). These results provide evidence that Nrg1 binds in vivo to the ARR2 region and that this binding is altered as a result of sudden environmental alkalinization.   DECEMBER 1, 2006 • VOLUME 281 • NUMBER 48

Induction of the ENA1 Promoter by Ambient Alkalinization
Involves Three Main Signaling Pathways-The results obtained indicated that the calcineurin-independent activation in response to high pH of the ARR2 region of the ENA1 promoter was controlled through the Snf1 and Rim101 pathways. Therefore, calcineurin, Snf1, and Rim101 would define the main elements governing the transcriptional response of the ENA1 gene to ambient alkalinization. In this case, one would expect that blockage of these three pathways would result in the incapacity of the ENA1 promoter to respond to alkaline stress. To test this hypothesis, we combined the snf1 and rim101 mutations and tested expression from the reporter plasmid pKC201, which contains the entire ENA1 promoter, in the presence of the calcineurin inhibitor FK506. As shown in Fig. 7A, systematic blockage of each signaling component results in additive loss of transcriptional response, whereas loss of all three elements results in a fully insensitive promoter. Therefore, calcineurin, Snf1, and Rim101 define three pathways that account for virtually every positive input to the ENA1 promoter in response to extracellular alkalinization. However, the relevance of the different pathways in control of the ENA1 promoter is not the same and, under this specific experimental condition, a sequence Snf1 Ͼ calcineurin Ͼ Rim101 can be established (Fig. 7A). These results were substantiated by direct analysis of the ENA1 mRNA levels by real-time reverse transcriptase-PCR in different mutants after exposure to alkaline stress (Fig. 7A, inset). As observed using the ␤-galactosidase reporter, inhibition of calcineurin in a strain lacking Snf1 and Rim101 fully blocks the high pH-induced expression of the ENA1 gene. In contrast, induction of the PHO84 gene was unaltered in this strain when compared with wild type cells (not shown).
As mentioned in the Introduction, ENA1 expression can be induced by saline as well as by alkaline stress. We wanted to compare the relative relevance of the pathways studied in this work in response to both types of stress. Cells were exposed to pH 7.9 or 0.4 M NaCl, because these treatments provoke an ENA1 response of similar potency and timing. Under these conditions (Fig. 7B), the induction of ENA1 by saline stress and high pH was similarly decreased in a calcineurin mutant. In contrast, lack of Snf1 had a strong effect on the alkali-induced response but produced only a moderate decrease in expression in cells exposed to NaCl. Mutation of RIM101 barely affected the response to saline stress. We have also compared how blockage of the three signaling pathways affects cell tolerance to high pH. As shown in Fig. 8, chemical inhibition of the calcineurin pathway and mutation of SNF1 decrease similarly the tolerance to high pH, whereas mutation of RIM101 has a lesser effect. In all cases, the phenotype observed was less intense than that produced by the absence of the ATPase gene. However, cells lacking both Snf1 and Rim101 grown in the presence of the calcineurin inhibitor FK506 were clearly more sensitive to high pH than ena1-4 cells.

DISCUSSION
The promoter of the ENA1 ATPase gene can integrate a large number of signals produced in response to changes in the ionic composition of the environment. Previous work has defined in some detail how the calcineurin pathway regulates the expression of ENA1 under alkaline pH stress (20,21) and there is also compelling evidence that activation of the Rim101 pathway can result in increased ENA1 expression by acting on the Nrg1 repressor (21,24,25), although in this case the target elements in the ENA1 promoter were still undefined. It was evident that activation of the calcineurin and Rim101 pathways could not fully explain the response of the ATPase gene under alkaline conditions (21), indicating that additional mechanisms must be involved. In this work we delineate the three main regulatory pathways that would account for the full response of ENA1 under alkaline stress and their possible interactions (Fig. 9). As previously reported, calcineurin would control ENA1 by binding of the calcineurin-activated transcription factor Crz1 to the calcineurin-dependent response element present in the ARR1 region. The calcineurin-independent response would be under the control of two additional pathways. One of them would involve the MIG binding site at nucleotides Ϫ534/Ϫ544, previously defined as important for response to glucose starvation (9). Our results suggest that both Mig1 and Mig2 act on the ARR2 region, as expression from the MCIR is stronger in the mig1 mig2 strain than in each single mutant, but that probably only Mig2 is under the control of Snf1 (Fig. 2). This would explain the early observation by Alepuz and co-workers (10) that a snf1 mutant displays decreased ENA1 expression at pH 8.5 when compared with a wild type strain. The Snf1 kinase is a key regulator of the yeast response to decreased glucose availability (38) and the activation of this kinase in response to alkaline stress suggests that exposure to high pH could mimic a situation of glucose starvation. This notion would be supported by the observation that severe alkaline stress induces the expression of many genes that are also induced by low glucose availability (20). The third pathway would be defined by Rim101. Our data support the previous observation that the impaired expression of ENA1 in a rim101 mutant is overridden by deletion of Nrg1 (25). Interestingly, we observe that Nrg2 does not seem to play a role in repression of ENA1 under high pH stress. This could be considered unexpected, because Nrg1 and Nrg2 are closely related proteins and a recent genomic survey revealed that both proteins exert a similar control over most Nrg-regulated promoters (39).
Our results provide the basis to understand how Nrg1 regulates ENA1 expression. Previous results indicated that alkaline-induced expression from both ARR1 and ARR2 regions were sensitive to the absence of Rim101 (21). It has been suggested that Nrg1 could bind to two sequences (CCCCT and CCCTC) that occur in the ENA1 promoter at positions Ϫ650 and Ϫ725, respectively (25). However, whereas the latter may well account for the decreased response from the ARR1 region in a rim101 strain, the former occurs in a region between ARR1 and ARR2 that has been shown to be insensitive to high pH (21). In addition, this location would not explain the presence of a Rim101-sensitive component within the ARR2 region (21). In this work we present evidence that the Nrg1 binding element must be located at the 5Ј-end of ARR2, comprising nucleotides Ϫ561/Ϫ573. Interestingly, this short region does contain, in the non-coding strand, a AGACCCT sequence that is somewhat different from the initially proposed sequence for recognition by Nrg1 (37) but that closely matches the consensus sequence ggAC-CCT, identified in very recent studies as a very likely Nrg1 binding element (40,41). Therefore, we propose that this sequence is responsible for the control exerted by the Rim101 pathway on the ARR2 region of ENA1. The fact that we detect in vivo binding of Nrg1 to the ARR2 region and that this binding is substantially decreased upon exposure of the cells to high pH stress provides further support to our hypothesis and suggest that Nrg1 binding to this region has physiological relevance. containing the native (Nrg1) or the HA-tagged version of Nrg1 (Nrg1-HA) were grown on YPD medium and switched at the indicated pH values for 30 min before cells were collected. Fragmented DNA was obtained and immunoprecipitated as described under "Experimental Procedures" and amplified by PCR using the indicated combination of oligonucleotides. Lanes 10 and 11 show the amplification of a region of the HXK2 promoter with oligos A (ACTACGAGTTTTCTGAACCTCC) and B (TAATTTCGTGGATCTCGCAATC). This region contains an Rgt1 binding site but is devoid of Nrg1 binding sequences. The amplification fragments were resolved by agarose electrophoresis. Migration of standard markers is indicated on the right. B, the intensity of the amplification DNA band from the input sample (empty bars) and the chromatin immunoprecipitate (filled bars) obtained for each condition was integrated and the ratio calculated when possible (numbers in parentheses). Data represent the mean Ϯ S.D. of three experiments in duplicate. DECEMBER 1, 2006 • VOLUME 281 • NUMBER 48

JOURNAL OF BIOLOGICAL CHEMISTRY 36639
We observe that mutation of NRG1 largely counteracts the decrease in response from the MCIR of both the rim101 and the snf1 mutants, suggesting that the ability of Nrg1 to interact and repress the ENA1 promoter would be under the dual regulation of Rim101 and Snf1, as depicted in Fig. 9. Therefore, Snf1 would control two different repressors, Mig2 and Nrg1. This would fit with the observation that derepression under alkaline stress is somewhat stronger in a rim101 nrg1 mutant than in a snf1 nrg1 strain and would explain why the absence of Snf1 causes a stronger effect on ENA1 expression than that of Rim101 (Figs. 3 and 7A). The possibility of Snf1 controlling the activity of Nrg1 on a given promoter is not surprising, as previous evidence indicates that Nrg1 interacts with the Snf1 kinase (42) and has a role in regulating diverse glucose-repressed genes (37,42,43).
On the basis of previously reported data and the experiments presented here it is clear that the mechanisms for controlling ENA1 expression in response to saline and alkaline stress are somewhat overlapping but not identical. Whereas the calcineurin pathway plays a relevant role in both saline and alkaline stress, mapping of the ARR2 region clearly shows that the cAMP response element responsible for derepression of the ENA1 promoter in response to osmotic stress (9,44) is irrelevant under alkaline stress. This result allows explaining the previous observations that a hog1 mutant has unaltered ENA1 transcriptional response (21) and does not display increased sensitivity under high pH conditions. Our observation that blocking calcineurin, Snf1, and Rim101 pathways fully abolishes the response from the ENA1 promoter suggests that, in contrast to what has been described for saline stress (19), the TOR pathway and the Gln3/Gat1 GATA transcription factors are not relevant for ENA1 regulation under alkaline stress. This is remarkable, because Ure2, Gln3, and Gat1 have been found to be required for normal high pH tolerance (45)(46)(47). Therefore, the pH-related phenotypes of these mutant strains cannot be explained through a role for the corresponding gene products in the control of ENA1 expression.
The Ena1 ATPase plays an important role in the adaptation to high pH conditions, as evidenced by the intense pH-sensitive phenotype of a strain defective in the ATPase function. However, we observe (Fig. 8) that simultaneous blockage of the three main pathways governing ENA1 expression under alkaline stress results in a phenotype substantially stronger than that observed for the ATPase-deficient strain. This clearly indicates that one or more of these pathways play role(s) relevant for high pH tolerance that are independent of the ATPase function. Although this investigation is out of the scope of the present work, possible alternatives can be envisaged. For instance, it is suggestive that the calcineurin pathway was found to be relevant in the maintenance of cell integrity in front of cell wall damaging conditions (48,49), whereas exposure to high pH has FIGURE 7. Activation of three signaling pathways, calcineurin, Snf1, and Rim101, accounts for full response of the ENA1 promoter to alkaline stress. A, the indicated strains were transformed with the pKC201 plasmid that contains the entire promoter of ENA1 and cells were subjected to alkaline stress in the presence or absence of FK506 (1.5 g ml Ϫ1 ). Empty bars denote ␤-galactosidase activity under nonstressing conditions (pH 6.3) and filled bars under stressing conditions (pH 8.0). Data are mean Ϯ S.E. from at least six independent clones. The inset shows the induction of the ENA1 gene evaluated by real-time PCR in the indicated strains or conditions. The expression level of the non-induced (pH 6.3, empty bars) wild type strain was considered as the unit. Filled bars denote cells subjected to alkaline stress (pH 8.0) for 10 min. Data represent the mean of two independent determinations. B, the indicated strains were transformed with the pKC201 plasmid and subjected to alkaline (pH 7.9, filled bars) or saline stress (0.4 M NaCl, hatched bars) for 60 min before cells were collected and ␤-galactosidase activity measured. Data represent the mean Ϯ S.E. of six experiments. FIGURE 8. Effect of blockage of calcineurin, Snf1, and Rim101 pathways on high pH tolerance. The indicated strains were inoculated at an initial A 660 of 0.01 and in YPD medium buffered at the indicated pH values. FK506 was used at 1.5 g ml Ϫ1 . Growth is represented as the percentage of cell density at a given pH relative to cells growing at pH 6.0. Data correspond to the mean Ϯ S.E. of three experiments.
been recently proposed as a circumstance able to induce cell wall damage (20). 7