A Plant Ca2+ Pump, ACA2, Relieves Salt Hypersensitivity in Yeast

Stress responses in both plants and yeast utilize calcium-mediated signaling. A yeast strain, K616, which lacks Ca2+ pumps, requires micromolar Ca2+ for growth. In medium containing 100 μm Ca2+, K616 can withstand osmotic stress (750 mm sorbitol) and ionic stress (300 mm KCl) but not hypersodic stress (300 mm NaCl). Heterologous expression of the endoplasmic reticulum-located Arabidopsis thaliana Ca2+-ATPase, ACA2, permits K616 to grow under NaCl stress even in Ca2+-depleted medium. All stresses tested generated transient elevation of cytosolic Ca2+ in wild type yeast, K601, whereas NaCl alone induced prolonged elevation of cytosolic Ca2+ in K616. Both the Ca2+ transient and survival of cultures subjected to NaCl stress was similar for the ACA2 transformant and K601. However, whereas K601 maintained low cytosolic Na+ predominantly by pumping it out across the plasma membrane, the transformant sequestered Na+ in internal organelles. This sequestration requires the presence of an endomembrane Na+/H+-antiporter, NHX1, which does not play a significant role in salt tolerance of wild type yeast except at acidic pH. Transcript levels of the plasma membrane Na+-ATPase, ENA1, were strongly induced only in K601, whereas NHX1 was strongly induced in both K601 and the ACA2 transformant. The calmodulin kinase inhibitor KN62 significantly reduced the salt tolerance of the ACA2 transformant and the transcriptional induction of NHX1. Thus, the heterologous expression of a plant endomembrane Ca2+ pump results in the rapid depletion of cytosolic Ca2+ and the activation of an alternate mechanism for surviving saline stress.

Eukaryotic cells regulate a variety of cellular processes, including responses to abiotic stresses using calcium-mediated processes. Soil salinity adversely affects plant yields worldwide (1), salt-tolerant cultivars being reported to utilize signal transduction systems involving calcium (2)(3)(4). NaCl-induced [Ca 2ϩ ] cyt transients have been observed in plants (2,5), and the model plant Arabidopsis utilizes a calcineurin-like Ca 2ϩ -sens-ing protein in the salt overly sensitive (SOS) 2 stress signaling pathway (6). Other calcium-sensing proteins, such as calmodulin (CaM) (7) and Ca 2ϩ -dependent protein kinase may also be recruited in regulating Na ϩ homeostasis under saline stress (8,9).
Cells have evolved mechanisms for maintaining low resting levels of Ca 2ϩ in their cytoplasm (10). These involve regulated entry and active removal of Ca 2ϩ , the latter being mediated by high affinity Ca 2ϩ -ATPases and low affinity H ϩ /Ca 2ϩ antiporters (11). The sensitivity and response of the cell to various stresses, including salinity, is dependent on its ability to adequately sequester and use Ca 2ϩ from internal stores (12,13). Ca 2ϩ is stored in the cell wall and organelles with the vacuole serving as the main Ca 2ϩ sequestration site in plant cells, whereas the ER has also been suggested to play an important role in regulating Ca 2ϩ homeostasis (14,15).
Plants, like animals, produce a range of distinct [Ca 2ϩ ] cyt "signatures," such as spikes, waves, and oscillations (16 -18), to different stimuli, which couple to a correspondingly wide array of responses (10,40,19). Sculpting the Ca 2ϩ signature may be expected to require modulation of both Ca 2ϩ entry into the cytosol and of transporters affecting reuptake/efflux. Several lines of evidence, such as the increase in transcript levels of plant endoplasmic reticulum (ER) Ca 2ϩ -ATPases in tomato (20) and tobacco (21), suggest the involvement of Ca 2ϩ pumps in salt tolerance. The presence of multiple isoforms amenable to a range of regulation (11) also supports the hypothesis that Ca 2ϩ -ATPases may play a more direct role in signal transduction (22,23). However, this hypothesis has not thus far been tested.
Higher plants and the yeast Saccharomyces cerevisiae share several conserved elements in the cellular mechanisms of Na ϩ homeostasis (24). Multiple transport pathways mediate cellular Na ϩ homeostasis in S. cerevisiae. Na ϩ is believed to enter the yeast cell via the K ϩ transporters TRK1 and TRK2 (25). The primary route of Na ϩ extrusion across the plasma membrane is through the P-type ion pump ENA1 (26). Na ϩ /H ϩ -antiporters in the plasma membrane (NHA1), the late endosomal/prevacuolar membrane (NHX1), and the vacuolar VNX1 (27), a monovalant cation/H ϩ antiporter, could also serve to deplete [Na ϩ ] cyt . NHA1 has been shown to extrude Na ϩ into acidic medium (28). No significant role in tolerance responses has been demonstrated for NHX1 at neutral pH, but it contributes to endosomal/vacuolar Na ϩ sequestration when stressed in acidic medium (29,30). These multiple transport pathways regulating yeast Na ϩ homeostasis are under complex regulation; ENA1 is transcriptionally activated by several distinct pathways (26,31), of which the HOG1 pathway has also been proposed to enhance activity of the plasma membrane-located NHA1 (32).
A substantive understanding of the mechanisms of Ca 2ϩ homeostasis makes yeast an ideal system to study how Ca 2ϩ pumps could influence Na ϩ homeostasis and salt tolerance. Under normal conditions, extracellular Ca 2ϩ enters the yeast cytosol through an as yet unidentified transporter (33). The Ca 2ϩ -ATPase PMR1 pumps [Ca 2ϩ ] cyt into the ER and Golgi, whereas the high affinity Ca 2ϩ -ATPase, PMC1, and low affinity Ca 2ϩ /H ϩ exchanger, VCX1, serve to sequester it in the vacuole (33). Hypertonic shock, induced by extracellular NaCl, induces release of vacuolar Ca 2ϩ into the cytosol through the tonoplastlocated Ca 2ϩ channel YVC1 (34) and possibly influx of external Ca 2ϩ via the CCH1/MID1 Ca 2ϩ channel on the plasma membrane (35). Expression and function of PMC1, PMR1, and VCX1 are regulated by calcineurin (36), a highly conserved protein phosphatase that is activated by Ca 2ϩ /CaM. In contrast to plants, there are no Ca 2ϩ -ATPases or exchangers detected on the yeast plasma membrane (33) and consequently little or no active efflux of Ca 2ϩ across the plasma membrane.
The ER-located ACA2 and vacuolar ACA4 are Arabidopsis Ca 2ϩ -ATPases that have previously been expressed in yeast (37,38), and the latter has been shown to ameliorate the salt hypersensitivity of the host strain (38). This investigation was carried out to determine whether an ER-located plant Ca 2ϩ pump could facilitate a salt tolerance response and, if so, to understand the underlying cellular mechanism involved. Here we demonstrate that the ER-located ACA2 relieves the hypersensitivity of the salt-sensitive triple mutant yeast strain K616, in which genes encoding the Ca 2ϩ pumps (PMC1, PMR1, and the regulatory subunit of calcineurin, CNB1) have been deleted (39). This strain requires 10 Ϫ4 M Ca 2ϩ to grow and millimolar Ca 2ϩ to survive hypertonic stress. Heterologous expression of ACA2 in this strain alleviates the calcium requirement in both cases. ACA2-expressing cells sequester Na ϩ into internal stores rather than extrude it across the plasma membrane, as does wild type yeast. Using the aequorin reporter system, we demonstrate that ACA2 alters the [Ca 2ϩ ] cyt transient induced by NaCl stress in K616. We propose that this action of the pump contributes to the triggering of an alternative pathway, which activates the endomembrane Na ϩ /H ϩ -antiporter, NHX1, resulting in enhanced salt tolerance.

EXPERIMENTAL PROCEDURES
Yeast Strains and Growth Media-S. cerevisiae wild type strain K601 (MATa, leu2, his3, ade2, trp1, and ura3) and mutant strain K616 (MATa pmr1::HIS3 pmc1::TRP1 cnb1::LEU2, ura3) were grown in complete synthetic medium (SC) and SC minus Trp, His, and Leu, respectively, throughout this study. For the K616 strain, the medium was supplemented with 10 mM CaCl 2 . SC medium consisted of 1.7 g/liter yeast nitrogen base without amino acids, 5 g/liter (NH 4 ) 2 SO 4 , 1.3 g/liter drop-out mix without Ura, Trp, His, and Leu (these amino acids were added from separate stock solutions when required), and 2% (w/w) dextrose as a carbon source. Medium pH was adjusted to 6 (buffered with 5 mM MES) prior to the addition of agar and sterilization. Buffering by MES was found to be efficient, since autoclaving did not introduce significant change in medium pH. In all experiments under hyperosmotic stress, a Ca-EGTA buffer was used to maintain the free Ca 2ϩ of SC medium at 100 M, unless otherwise mentioned.
Yeast Transformation-For transformations with pYX112 constructs, K616 was transformed by the lithium acetate/polyethylene glycol method (40), and transformants were selected for their ability to grow in the absence of uracil on plates containing SC medium minus Trp, Leu, His, and Ura. For complementation studies, Ura-positive colonies were streaked on complete SC plates containing 10 mM EGTA and incubated for 3 days at 30°C.
For transformations with yeast vector pKC147/AEQ, K601 and K616 were transformed and selected for Ura autotrophy as mentioned above. A Ura-positive K616 colony was then subjected to a second transformation with pYX112-ACA2. Transformants were then selected based on their ability to complement on low calcium medium by spreading on SC minus Trp, Leu, His, and Ura, containing 5 mM EGTA. Controls for this selection included a transformation with the same volume of water and a vector alone, both of which showed no growth on the EGTA selection plate. Double transformants were analyzed for the expression of apoaequorin by monitoring luminescence before and after lysis with 5% Triton X-100 and 2 M CaCl 2 and by colony PCR to confirm the presence of the ACA2 gene. They were also tested for salt tolerance in SC-medium containing 400 mM NaCl.
NHX1 Knock-out Mutants in K616 Background-A forward primer (5Ј-AAACGTGATAGCAAGGAACTG-3Ј) that hybridizes 200 bp upstream of the start codon of NHX1 and a reverse primer (5Ј-GGCGTTGAGTAAGAGAGAATG-3Ј) that hybridizes 80 bp downstream of the NHX1 stop codon were used to amplify the NHX1::KanMX cassette from the appropriate deletion strain in the background of the yeast strain, BY4742 (a gift from Dr. R. Rao). The 1.82-kb PCR product thus obtained was gel-eluted and used to transform K616. This strategy allows specific gene disruption by homologous recombination and chromosomal integration. Recombinants were selected on SC medium containing 1 mg/ml G418. A positive transformant was then subjected to a second round of transformation with pYX112-ACA2 and selected on SC minus Ura, Trp, Leu, and His plates containing 1 mg/ml G418. Genomic DNA was isolated from these colonies and analyzed for the presence of ACA2 and the KAN-NHX1 sequences by PCR using specific primers followed by DNA sequencing.
Monitoring Expression of ENA1 and NHX1 by RT-PCR-Total RNA was extracted from yeast cells exposed to 400 mM NaCl for a range of time periods (0 -50 h). The RNA was first treated with RNase-free DNase (Promega, Madison, WI) to eliminate contaminating DNA. 500 ng of RNA each was then subject to a reverse transcriptase reaction using 100 units of Moloney murine leukemia virus reverse transcriptase (Invitrogen) for 1 h at 37°C in buffer containing 50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgCl 2 , 10 mM dithiothreitol, and 0.5 mM each of dNTPs. The cDNA thus obtained was subjected to a 30-cycle PCR reaction using standard conditions with primers specific for ENA1 (forward primer, 5Ј-ATG GGC GAA GGA ACT ACT AA-3Ј; reverse primer, 5Ј-CAC ACA AAT GGC ATA GAT AGC-3Ј) amplifying a 960-bp product and for NHX1 (forward primer, 5Ј-ATG CTA TCC AAG GTA TTG CTG-3Ј; reverse primer, 5Ј-AAC CGA CGA AAA TGT TGC-3Ј) amplifying a 759-bp product. The constitutively expressed ACT1 (actin) was amplified under identical PCR conditions using specific primers (forward primer, 5Ј-GAG GTT GCT GCT TTG GTT AT-3Ј; reverse primer, 5Ј-GCG GTT TGC ATT TCT TGT-3Ј) amplifying a 680-bp product.
Stress Conditions-For drop tests, midlog yeast cultures were washed twice with water, and A 600 was adjusted to 0.5 with sterile double-distilled water. Cells were serially diluted with water to obtain 10-, 10 2 -, and 10 3 -fold dilutions. Five microliters of each dilution was then spotted on SC medium buffered at 20 nM, 100 M, or 5 mM Ca 2ϩ and supplemented with 300 mM KCl, 400/800 mM NaCl, or 750 mM sorbitol concentrations. Growth was recorded after a 3-day incubation at 30°C.
For growth curves, equal cell inocula were added into liquid SC medium buffered for Ca 2ϩ at 100 M with or without 400 mM NaCl. In some experiments, medium was also supplemented with amiloride (500 M). Cells were incubated on an incubator shaker at 30°C. Change in A 600 was monitored over a period of 2 days. Alternatively, sensitivity curves were generated using log phase cells with or without preincubation with inhibitors such as 5 M bafilomycin, N-(p-amylcinnamoyl) anthranilic acid (ACA) ( Estimation of Total Cellular Na ϩ and Ca 2ϩ -Yeast strains were grown to an A 600 of 0.6 in SC medium buffered at 100 M Ca 2ϩ . Cells were either taken as such or stressed with 400 mM NaCl, and 2-ml aliquots were collected over a range of time points extending up to 18 h. Cells were collected by centrifugation, washed four times in ice-cold Buffer B (10 mM Tris-HCl, pH 6.0, 2 mM MgCl 2 , 1% glucose, 0.6 M sorbitol), and dried at 50°C for 4 days. Dried cells were acid-digested in a 4:1 diacid mixture of perchlorate, and nitrate and Na ϩ levels were estimated by flame photometry using a Systronics Flame Photometer 128 (Ahmedabad, India) (41). In a few experiments, growth was carried out at Ca 2ϩ concentrations of 100 M and 5 mM. At an A 600 of 0.6, a set of cells grown in 100 M Ca 2ϩ was washed three times with SC medium and adapted to 20 nM Ca 2ϩ for 1 h before processing these cells, as well as those at 100 M and 5 mM for total Ca 2ϩ estimation using a Zeeman atomic absorption spectrophotometer (model Z-6100; Hitachi).
Permeabilization of Plasma Membranes with Cu 2ϩ Ions-Yeast strains were grown in 300 ml of SC medium, to A 600 of 0.6 and stressed for 3 h with 400 mM NaCl. Cells were harvested by centrifugation, washed twice in ice-cold Buffer B, and resuspended in the same buffer at room temperature at a density of 1 ϫ 10 8 cells/ml. After withdrawing 2-ml aliquots for total Na ϩ estimation, the remaining cell suspension was subjected to plasma membrane permeabilization with Cu 2ϩ as described by Anraku and co-workers (42). CuSO 4 was added to a final concentration of 500 M and incubated at 23°C. At the indicated times, aliquots were withdrawn and washed twice with buffer B, and cells were dried for acid digestion and flame photometry.
Aequorin Luminescence Measurements and [Ca 2ϩ ] cyt Quantification-Aequorin luminescence was determined using a luminometer (TD-20/20; Turner Designs, Sunnyvale, CA) following procedures used to quantify [Ca 2ϩ ] cyt in yeast (19). Cells were transformed with the 2 plasmid pKC147/AEQ (provided by Dr. K. Cunningham) containing the APOAE-QUORIN gene. Transformants, grown to an A 600 of ϳ0.4 -0.6, were harvested by centrifugation, resuspended in fresh medium (SC-Ura ϩ 1 mM EGTA) at an A 600 of 10, and loaded with 25 g/ml coelenterazine (Molecular Probes, Inc., Eugene, OR) for 3 h in the dark at room temperature. Cells were then collected by centrifugation and resuspended in the same volume of fresh medium buffered at 100 M Ca 2ϩ . After a preincubation of 1 h, the cells were diluted with the same medium to an A 600 of 0.5, and a cellular luminescence base line was determined by 2-3 min of recordings at 10-s intervals.
Hypertonic/ionic shock was administered to yeast cells by the addition of a range of NaCl (100 -1200 mM), 300 mM KCl, 50 mM CaCl 2 , or 750 mM sorbitol. When required, loaded cells were preincubated for 20 min with 5 M bafilomycin A1 or 30 M ACA dissolved in Me 2 SO or 10 mM LaCl 3 dissolved in water prior to the acquisition of the basal luminescence reading and the subsequent osmotic/ionic shock. Luminescence from aequorin that remained in cells at the end of an experiment was determined by diluting the cell suspension 1:1 with a solution containing 10% Triton X-100 and 4 M CaCl 2 . [Ca 2ϩ ] cyt was calculated using Equation 1, where L represents the luminescence intensity at any time point, and L max is the integrated luminescence intensity (34). L values did not exceed ϳ1% of the maximal light emission capacity in all three strains. Figures are representative of a minimum of three experiments in each case.

RESULTS
Complementation by Full-length ACA2-K616 transformed with pYX112-ACA2 was selected by Ura autotrophy. K616 is a mutant for the endogenous Ca 2ϩ pumps PMC1 and PMR1, a potentially lethal combination but for a third knock-out mutation of CNB1. In the absence of CNB1, the Ca 2ϩ /H ϩ -antiporter (VCX1) located in the vacuolar membrane is active and suffices to populate intracellular calcium stores, provided K616 is grown in adequately high Ca 2ϩ . Ura-selected colonies expressing ACA2 exhibited growth on complementation plates comprising SC medium depleted of calcium with 10 mM EGTA, unlike K616 and K616 transformed with the vector alone (Fig.  1A). The ACA2 transformant strain will be referred to as K616-ACA2 henceforth for simplicity, whereas the strain transformed with vector alone will be referred to as K616-V. K616, in turn, is derived from the strain K601, which will be treated as wild type for this study.
Saline and Osmotic Sensitivity of Yeast Strains-At very low Ca 2ϩ of 20 nM, K616-V grew very poorly on SC medium (Fig.  1B) and not at all under 400 mM NaCl stress. Interestingly, at a moderate free Ca 2ϩ level of 100 M, K616-V showed effective growth under unstressed conditions but failed to grow under salt stress (Fig. 1B). With 5 mM Ca 2ϩ in the medium, however, K616-V was able to grow effectively under unstressed conditions, as well as mount a reasonable salt tolerance response when stressed with NaCl (Fig. 1B). Thus, K616-V requires two levels of media Ca 2ϩ : a lower concentration (ϳ100 M) for growth and millimolar calcium for reasonable salt tolerance. K616 and K616-V behaved identically in Ca 2ϩ -depleted medium (Fig. 1A) and under hyperosmotic stress (supplemental Fig. 1D). The expression of ACA2 in K616 eliminated both requirements, transformants growing well and exhibiting salt tolerance even at 20 nM Ca 2ϩ (Fig. 1B). NaCl titration carried out at 100 M medium Ca 2ϩ showed that K616 did not grow in 300 mM NaCl (Fig. 1C). Sensitivity to NaCl was detected even at 100 mM, where a significant lag was observed in the growth curve (Fig. 1C). K601 and K616-ACA2, on the other hand, grew well in medium containing up to 1.2 M NaCl (growth curves not shown).
To evaluate whether other hyperosmotic stresses affect these strains similarly, they were subjected to saline and osmotic stress comprising 300 mM KCl or 400/800 mM NaCl or 750 mM sorbitol in SC medium containing 100 M free Ca 2ϩ . Although both K601 and K616-ACA2 grew well on plates containing 400 or 800 mM NaCl, K616-V did not grow on these plates (Fig. 1D). However, all three strains grew reasonably well on plates containing either 750 mM sorbitol or 300 mM KCl, indicating that purely osmotic stresses and ionic stresses up to 300 mM were tolerated by all of them (Fig. 1D). Curing the K616-ACA2 strain of pYX112-ACA2 generated colonies that did not grow on calcium-depleted media and were as sensitive to NaCl stress as was K616 (supplemental Fig. 1, A-D). ATP-induced Ca 2ϩ pumping and vanadatesensitive-ATPase activities were detected in ER-enriched vesicles prepared from K616-ACA2 but not in those from K616, indicating the expression of a functional Ca 2ϩ -ATPase in K616-ACA2 (supplemental Fig. 1, E-G). An N-terminal, autoinhibitory domain in ACA2 (43) does not appear to prevent activity of the full-length construct used here.
ACA2 Changes Ca 2ϩ Homeostasis in K616-To test the degree to which internal calcium stores get filled, cellular Ca 2ϩ of yeast preincubated with external Ca 2ϩ of 20 nM, 100 M, or 5 mM was estimated by atomic absorption spectrophotometry. Total cellular calcium is tightly controlled in wild type, K601, and K616-ACA2, remaining below 0.2 mg/g, dry weight, under all of the Ca 2ϩ concentrations tested ( Fig. 2A). K616 showed Ͼ8-fold higher total calcium than K601 when subjected to 20 nM external Ca 2ϩ ( Fig.  2A). However, at both 100 M and 5 mM calcium, cellular Ca 2ϩ levels were comparable among all of the yeast strains tested, suggesting that internal stores of Ca 2ϩ are adequately filled in all. We focused on 100 M Ca 2ϩ , since it is the concentration that is adequate for growth of K616 but insufficient for mounting a salt tolerance response.
At 100 M Ca 2ϩ , the sensitivity of K616 to hypersodic media could be a consequence of altered signaling due to an aberrant [Ca 2ϩ ] cyt transient in response to saline stress. To test this hypothesis, we used aequorin to monitor [Ca 2ϩ ] cyt . The addition of 400 mM NaCl to the yeast suspension resulted in rapid increases in aequorin luminescence in all three strains (Fig. 2B). In the case of K601 and K616-ACA2, the rise in [Ca 2ϩ ] cyt was limited to a peak value between 0.6 and 0.8 M, from which maximum it decayed to basal levels in 2-3 min (Fig. 2B). In K616, on the other hand, [Ca 2ϩ ] cyt rose to over 1.2 M over the course of 1 min. Thereafter, it declined very slowly, not reaching basal values even after 20 min (Fig. 2B). Effect of Bafilomycin A1 on the NaCl-induced Ca 2ϩ Transient-Treatment with bafilomycin A1, an inhibitor of the vacuolar H ϩ -ATPase, severely perturbed the NaCl-induced [Ca 2ϩ ] cyt transient only in K616 (Fig. 2C). The vacuolar H ϩ -ATPase acidified the vacuole, providing the driving force for Ca 2ϩ sequestration through VCX1.
Changes in the Amplitude and Duration of the [Ca 2ϩ ] cyt Rise with Increasing NaCl Stress-Cells were stressed with varying concentrations of NaCl ranging from 100 to 800 mM. K601 and K616-ACA2 responded in all cases with small transient rises in [Ca 2ϩ ] cyt (data not shown) and were also able to grow well in media containing up to 1200 mM NaCl (growth on 800 mM NaCl shown in Fig. 1D). In K616, however, a transient rise was observed in response to 100 and 200 mM NaCl as opposed to prolonged elevation in response to stresses of 300 mM and greater (Fig. 2D) 1D and supplemental Fig. 2). K616 was reasonably tolerant to sorbitol, KCl, and CaCl 2 stresses (Fig. 1D; data for 50 mM CaCl 2 not shown) and also mounted a transient [Ca 2ϩ ] cyt elevation for these stresses, reaching base-line values in 5-7 min (supplemental Fig. 2). In contrast, NaCl stress resulted in a large and prolonged [Ca 2ϩ ] cyt rise in this strain with an estimated t1 ⁄ 2 of ϳ11 min. [Ca 2ϩ ] cyt did not reach base line over the 20-min duration of the experiment (Fig. 2B).
NaCl Stress Mobilizes Ca 2ϩ from an Intracellular Store-Pretreatment of the three strains with 10 mM LaCl 3 , 5 mM NiSO 4 , or MgCl 2 , all of which have been shown to block yeast plasma membrane Ca 2ϩ channels (35,44), did not show any significant effect on the [Ca 2ϩ ] cyt rise with NaCl stress (supplemental Fig.  3, A and B), suggesting that most, if not all, of the Ca 2ϩ that enters the cytosol is released from internal stores. Chelating external Ca 2ϩ with 5 mM EGTA for 60 min prior to the NaCl insult also had no effect on the induced Ca 2ϩ transient (supplemental Fig. 3B, inset), confirming that the [Ca 2ϩ ] cyt rise is not due to influx from the medium. Of the stresses used, only the transient induced by Ca 2ϩ stress (50 mM CaCl 2 ) was reduced by La 3ϩ (supplemental Fig. 3C), demonstrating that the [Ca 2ϩ ] cyt in the other cases (data for KCl and sorbitol not shown) was mobilized from internal sources.
ACA2 Changes Na ϩ Homeostasis in K616-Cells were grown to midlog phase and stressed for up to 18 h with 400 mM NaCl. Total cellular Na ϩ in K601 did not increase significantly over the first 3 h of stress and increased gradually thereafter until the 18 h time point (Fig. 3A). In contrast, both K616-ACA2 and K616 showed significant increases in intracellular Na ϩ after 3 h and still greater increase after 6 and 18 h (Fig. 3A). K616 survived poorly after 18 h of NaCl stress, with close to 60% of the cells dead (data not shown). Consequently, data at the 3 h time point, where all three strains were viable, have been taken for comparative analysis. K616-ACA2 accumulated about twice as much Na ϩ as the wild type strain K601 after 3 h (Fig. 3A).
Treatment with CuSO 4 permeabilizes the plasma membranes to small solutes but does not permit release from Cu 2ϩresistant intracellular compartments (42), which are termed the  FEBRUARY 8, 2008 • VOLUME 283 • NUMBER 6 JOURNAL OF BIOLOGICAL CHEMISTRY 3501 "nonexchangeable pool." This assay has been used earlier to estimate the sequestration of Na ϩ and K ϩ into late endosomal/ vacuolar compartments of yeast (29,45). Only ϳ25% of total cellular Na ϩ was released from cells stressed for 3 h in the case of K616-ACA2, whereas ϳ75% was released from both K601 and K616, indicating that K616-ACA2 accumulates a larger fraction of Na ϩ in internal compartments (Fig. 3B).

ER Ca 2؉ -ATPase Sculpts Signaling Ca 2؉ Transients
Requirement of NHX1 for Survival of K616-ACA2-The sequestration of Na ϩ into a nonexchangeable pool suggests active transport into intracellular organelles. The yeast prevacuolar Na ϩ /H ϩ -antiporter NHX1 has an amiloride-binding domain, and its activity is reduced by amiloride (46). Amiloride affected the growth of K616-ACA2 in the presence of 400 mM NaCl (Fig.  3C) but not that of K601 (data not shown). Amiloride affected neither strain in the absence of salt, indicating that transporters critical for the tolerance response in the transformant are sensitive to amiloride, whereas those in K601 are not. Further, treatment with amiloride reduced the fraction of total cellular Na ϩ that was not released from K616-ACA2 upon Cu 2ϩ -induced permeabilization of the plasma membrane (Fig. 3B), indicative of a corresponding decline in Na ϩ sequestration in intracellular organelles.
Bafilomycin A1, an inhibitor of the vacuolar H ϩ -ATPase, affected growth of K616-ACA2 when subjected to 18 h of exposure to 400 mM NaCl, without any major effect on K601 growth (Fig. 3D). Survival thus requires the building up of pH gradients across internal membranes, consistent with the use of a transporter coupled to H ϩ movement.
Inserting a kanamycin cassette into the NHX1 gene (NHX1::KanMX) generates yeast strains that are nulls for the transporter. NHX1 knockout (⌬nhx1) was generated in the background of K616, which we named as KNX161, to more directly assess the contribution of this transporter to the alleviation of salt hypersensitivity. The knock-out was selected on G418 and analyzed by sequencing, which confirmed a disruption of NHX1 (Ydr456w) in chromosome IV. Under control conditions, all of the strains tested grew well. Under both Ca 2ϩ -depleted and NaCl-stressed conditions, K616 and KNX616 failed to grow (Fig. 3E). KNX161-ACA2, on the other hand, grew well in Ca 2ϩ -depleted medium (Fig. 3E), but in the presence of 400 mM NaCl, its growth was severely compromised (Fig. 3, E and  F), indicating that the activity of the transporter is critical to the survival of this strain in high salt over the 48 h time scale of this experiment. Cu 2ϩ induced release of close to 75% of total cellular Na ϩ from KNX616-ACA2 (Fig. 3B), indicating that the sequestration seen in K616-ACA2 is dependent on NHX1.
Transcriptional Regulation of ENA1 and NHX1-RT-PCR experiments on RNA extracted from the four strains revealed that ENA1 transcripts increased upon exposure to NaCl in all strains (Fig. 4). The largest increase was seen in K601, with significantly smaller increases seen in the other strains. ENA1 transcripts appeared within minutes of NaCl exposure in K601, increased severalfold over 90 min, and were then sustained over several h. In contrast, a mild transient increase in transcript levels over the control was observed within minutes in K616-ACA2 and KNX616-ACA2, followed by a second modest and delayed induction after several h of salt stress; K616 also showed some induction of ENA1 with NaCl stress.
Transcript levels of NHX1 increased severalfold over the basal level with time of exposure to NaCl in both K601 and K616-ACA2, both the levels and the temporal pattern of induction being comparable in the two cases (Fig. 4). Transcriptional induction of NHX1 over control levels was not observed in K616. No RT-PCR product was observed in KNX616-ACA2, in keeping with a knock-out phenotype. Levels of ACTIN transcripts were unaffected by stress and served as a loading control (Fig. 4).
Identifying Candidate Signaling Intermediates That Regulate Salt Tolerance of K616-ACA2-ACA is reported to both inhibit phospholipase A2 and block TRP channels (47). ACA was preincubated with log phase cells of K601 and K616-ACA2 at sublethal doses (30 and 40 M), followed by an 18-h exposure to a range of NaCl (0 -1000 mM). Inhibitor alone, at 30 M, had no significant effect, and at 40 M, it had little effect on either strain (Fig. 5A). Significant reduction in cell viability of both K601 and K616-ACA2 was observed with increasing NaCl stress, with essentially no survival above 800 mM NaCl (Fig. 5A). Pretreatment with ACA reduced the NaCl-induced [Ca 2ϩ ] cyt transient (supplemental Fig. 3D) and severely affected transcriptional induction of ENA1 and NHX1 in both K601 and K616-ACA2 (Fig. 5, B and C).
In an effort to block a downstream intermediate in the pathway, we used KN62, an inhibitor of CaM kinases, which had no effect on salt tolerance at 20 M and was found to be lethal at 100 M. However, at a sublethal dose of 50 M, the inhibitor had a significant deleterious effect on salt tolerance of K616-ACA2, whereas no perturbation was observed in K601 (Fig. 5D). KN62 thus appears to affect a step in the salt tolerance pathway unique to K616-ACA2. Indeed, RT-PCR analysis showed that this inhibitor blocks the transcriptional induction of both ENA1 and NHX1 in K616-ACA2 (Fig. 5E).

DISCUSSION
Plants show intra-and interspecific variations in their tolerance to salinity (48,49). Adaptive responses to maintain low [Na ϩ ] cyt , especially in photosynthesizing tissues, are complex and occur at both the whole plant and cellular levels (3,41). When exposed to salinity, Arabidopsis uses a pathway comprising  a Ca 2ϩ -binding protein SOS3, which activates the kinase SOS2 that, in turn, phosphorylates the plasma membrane Na ϩ /H ϩ antiporter SOS1, thereby activating it. Mutations in these genes result in an SOS phenotype (6). Salinity has also been shown to induce transient elevation of Ca 2ϩ in the cytosol of plant cells (5,23) and to up-regulate Ca 2ϩ -dependent protein kinases in rice (8,9). Thus, Ca 2ϩ -mediated signaling plays a pivotal role in salt tolerance of plants.
The role of endomembrane Ca 2ϩ -ATPases would appear to be limited to cocking the gun, as it were, by filling the intracellular stores from which Ca 2ϩ can be released in response to specific stimuli. The realization that Ca-ATPases in plants are potentially amenable to post-translational modification has led to the proposal that they may regulate the characteristics of the stimuli-induced Ca 2ϩ transient, thereby dictating the pathway activated (22).
This study explores the relevance of an ER-located Ca 2ϩ -ATPase from Arabidopsis (ACA2) in modulating stress responses. The study has been carried out in yeast, which, like plants, has a calcium-mediated salinity stress response mechanism and is an organism where Ca 2ϩ and Na ϩ homeostasis mechanisms are relatively well understood (24,33). ACA2 has been expressed in a triple mutant yeast strain, K616, which is deficient in Ca 2ϩ pumps and the regulatory subunit of the highly conserved protein phosphatase, CNB1. Calcineurin-induced expression of the plasma membrane Na ϩ -ATPase, ENA1 (26), may be expected to be compromised in this strain, which is seen to be hypersensitive to salt stress. The strain does not grow in low Ca 2ϩ below 100 M (Fig. 1B).
At extremely low external Ca 2ϩ , K616 hyperaccumulates Ca 2ϩ ( Fig. 2A), as previously reported for strains lacking PMR1 (50). This accumulation and possible mislocalization may underlie poor survival of the strain under these conditions. Total Ca 2ϩ is comparable in all three strains studied when suspended in medium containing 100 M or more of Ca 2ϩ ( Fig.  2A), demonstrating that the only remaining Ca 2ϩ transporter, the vacuolar Ca 2ϩ /H ϩ exchanger VCX1, is capable of both filling internal stores and of regulating [Ca 2ϩ ] cyt under optimal growth conditions. When subjected to saline stress, however, K616 is unable to grow in medium containing 100 M Ca 2ϩ , whereas K616-ACA2, with an Arabidopsis ER-Ca 2ϩ -ATPase, grows as well as does K601 (Fig. 1, B and D). It would thus appear that endomembrane Ca 2ϩ -ATPases serve roles beyond the cocking of the gun (i.e. filling Ca 2ϩ stores).
Survival of K616-ACA2 in 400 mM NaCl is comparable with that of K601 (Fig. 1D). On the other hand, total cellular Na ϩ is three times as high in K616-ACA2 as in K601 (Fig. 3A). Wild type yeast maintain low [Na ϩ ] cyt by pumping Na ϩ out across the plasma membrane using the ENA1-4 family of Na ϩ -ATPases (26). Although transcript levels of ENA1 are somewhat higher under salt stress in both strains, the increase in K601 is much larger than in the transformant (Fig. 4). Moreover, the significantly higher levels of total cellular Na ϩ in K616-ACA2 compared with K601 suggest that efflux may not be the primary means of regulating [Na ϩ ] cyt in this strain.
If one assumes that survival requires stringent control of free [Na ϩ ] cyt , as is suggested by studies monitoring [Na ϩ ] cyt in rice (2, 3) and other systems (51), then it follows that the Na ϩ accu-mulated by K616-ACA2 (Fig. 3A) is either sequestered in internal organelles or accompanied by suitable compatible solutes. We have taken recourse to an assay that has been used to estimate Na ϩ in endomembrane compartments of yeast several times over the past decade (29,45). Cu 2ϩ -induced release of cytosolic solutes releases up to 75% of Na ϩ in K616 and K601 but only 25% from K616-ACA2 (Fig. 3B), indicating that ϳ75% of cellular Na ϩ is sequestered in endomembrane compartments. This sequestration requires the presence of the prevacuolar Na ϩ /H ϩ -antiporter, NHX1, as seen by the significantly decreased fraction of nonexchangeable Na ϩ upon blocking NHX1 with amiloride or knocking it out (Fig. 3B). The limited sequestration (25%) observed in K601 and in KNX616-ACA2 (Fig. 3B) could possibly be ascribed to VNX1, a low affinity vacuolar Na ϩ /H ϩ and K ϩ /H ϩ antiporter (27). The ⌬nhx1 strain (KNX616-ACA2) exhibits hypersensitivity to saline stress (Fig. 3, E and F). Knocking out NHX1 could prove lethal due to its role in vesicular trafficking and sorting of proteins destined to the vacuole (52). However, Stevens and co-workers (53) have demonstrated that a knock-out of NHX1 (VPS44) in yeast had no effect on growth under control conditions or under ionic stress. On the other hand, the deleterious effect of bafilomycin on the salt tolerance of K616-ACA2 (Fig. 3D) indicates the involvement of an H ϩ gradient-driven transporter in the tolerance mechanism.
The endosomal compartments where NHX1 is located are small but have been shown to be adequate for enhancing salt tolerance under conditions of acidic pH, as demonstrated by Rao and co-workers (29). In addition, a more recent study indicates that although ENA1 to -4 are the main contributors to Na ϩ homeostasis in S. cerevisiae, there is some contribution of NHX1 and VNX1 mediating Na ϩ transport into the yeast endosomal/vacuole, respectively, even at neutral pH (27).
To address the question of how a plant ER-located Ca-ATPase could bring about changes in adaptive Na ϩ homeostasis in K616, we looked at [Ca 2ϩ ] cyt profiles using aequorin as a reporter. The 3:1 stoichiometry of the functional complex and probable nonuniform expression mean that absolute concentrations estimated using the commonly used expressions for normalization should be used cautiously. The time courses of the transients monitored are, however, likely to be reliable.
[Ca 2ϩ ] cyt rises sharply in all three strains in response to stress. It decays to basal levels in a few minutes in all instances when a successful stress response is mounted, irrespective of the stress (supplemental Fig. 2 and Fig. 2, B and D), whereas [Ca 2ϩ ] cyt remains above 400 nM for an extended period in all instances when cells fail to survive the stress (Fig. 2, B and D).
Rapid refilling of internal stores could promote survival in one or more of several ways. (a) Stores are now made available for subsequent release. However, the fact that all stresses stimulated robust release of Ca 2ϩ in K616 suggests that initial filling is not a limiting factor. Although only a single transient increase in [Ca 2ϩ ] cyt is observed over the time course of our aequorin experiments, we cannot rule out the possibility that subsequent spikes are required to recruit downstream effectors or to maintain the response. (b) Limiting the duration that [Ca 2ϩ ] cyt is high could preclude initiation of mitochondrial death pathways. The time course of NaCl-induced cell death in K616 is not supportive of death processes being directly induced by Ca 2ϩ (data not shown). (c) The time course of the calcium signal could be critical for triggering downstream survival pathways. This hypothesis is appealing, particularly in view of the fact that poor survival is correlated with prolonged transients.
Surviving in high salt medium requires active depletion of Na ϩ from the cytosol. In K601, this is achieved mainly by the Na ϩ -ATPase ENA1 (26), with possible contributions from the plasma membrane Na ϩ /H ϩ antiporter NHA1. ENA1 expression is low under standard growth conditions but is rapidly induced by a number of complex pathways after exposure to saline or osmotic or alkaline stress (54). It is activated by calcineurin at high Na ϩ stress and by the Ca 2ϩ -independent HOG pathway at low osmotic stress (below 300 mM NaCl) (55). Inhibition of the TOR pathway by saline stress also activates ENA1 (56). The sensitivity of K616 to salinity indicates that neither NHA1 nor any calcineurin-independent mechanism of activation of ENA1 contributes significantly to the depletion activity under our experimental conditions. ENA1 is transcriptionally induced in all of the strains tested ( Fig. 4) but significantly more so in K601.
A microarray study found that yeast responds to saline stress by the transcriptional induction of a large number of genes, including ENA1, which shows transient activation over a time period of 20 min (57). Our data acquired over a time scale of hours detects strong induction of ENA1 in K601 after 1 h of stress, increasing for an additional 2 h and sustained for several h thereafter (Fig. 4). Among the strains lacking CNB1, K616 shows lowest activation, whereas K616-ACA2 and KNX616-ACA2 (with or without functional NHX1, respectively) show comparable mild induction of ENA1, in two phases (Fig. 4). Differences in temporal patterns of expression between the wild type strain and K616-ACA2 are indicative of differences in the mode of induction of ENA1. K616-ACA2 survives saline stress, whereas KNX616-ACA2 does not (Fig. 3, E and F), demonstrating that the levels of ENA1 induction in these strains is insufficient to confer tolerance in the absence of NHX1.
NHX1 is also amenable for transcriptional induction upon saline stress (Fig. 4), the temporal pattern of NHX1 induction being comparable between wild type and K616-ACA2, transcript levels increasing significantly from basal levels over several hours to comparable extents. However, accumulation of Na ϩ into internal stores is negligible in K601, suggesting a second level of post transcriptional control in K616-ACA2. It is to be noted that Posas et al. (57) report the rapid induction of several H ϩ -ATPases located on the tonoplast immediately after exposure to salinity.
We used sublethal doses of inhibitors of candidate signaling intermediates to decipher the pathway that couples the stress signal to the tolerance response in K616-ACA2. ACA, which affects both phospholipase A1 and TRP channels, dramatically reduced salt tolerance in both K601 and K616-ACA2, with negligible transcriptional induction of ENA1 in K601, and of NHX1 in both K601 and K616-ACA2 (Fig. 5, A-C). The inhibitor must block a common upstream signaling event in both strains. The [Ca 2ϩ ] cyt transient in response to NaCl arises from internal Ca 2ϩ stores (supplemental Fig. 3, A-C). Ca 2ϩ release from the vacuole in response to hyperosmotic stress has been earlier shown to involve YVC1, a TRP channel (34) that could be a target of ACA (47). Indeed, the [Ca 2ϩ ] cyt transient we see in response to NaCl is attenuated in ACA-pretreated cells (supplemental Fig. 3D). In the absence of CNB1, the downstream pathway in K616-ACA2 must diverge from that of the wild type strain. HOG mitogen-activated protein kinase plays roles in osmotic stress signaling in yeast (55). However, SB 203580, an inhibitor of HOG mitogen-activated protein kinase, had no effect on salt tolerance of K616-ACA2 (data not shown). On the other hand, the significant and specific reduction in viability of K616-ACA2 exposed to salt stress in the presence of KN62 and the resulting block of transcriptional induction of NHX1 (Fig. 5,  D and E) indicates that Ca 2ϩ /CaM kinase function is required in the survival pathway. Surprisingly, yeast strains lacking both CMK1 and CMK2, which encode Ca 2ϩ /CaM-dependent protein kinases, are viable, as are strains in which two other closely associated genes, CLK1 and RCK1, are knocked out along with CMK1 and -2 (58). These enzymes are thus not required for a constitutive function but may have roles during stress signaling. Indeed, the Ca 2ϩ /CaM kinase RCK1 has been reported to be transcriptionally up-regulated in yeast under salt stress (57).
Future studies will focus on this pathway that couples the [Ca 2ϩ ] cyt signature with the activation of NHX1. In plants, the significant contribution of the NHX1 homologue to salt tolerance is well documented (59,60). Interestingly, blocking a related kinase, Ca 2ϩ -dependent protein kinase, results in increases in [Na ϩ ] cyt in rice cells due to the lack of activation of a likely vacuolar Na ϩ /H ϩ antiporter (3). It is, however, to be determined whether ACA2 plays a direct signaling role in activation of plant NHX1.
Thus, heterologously expressed ACA2 sculpts the Ca 2ϩ transient induced by NaCl stress. K616-ACA2 survives saline stress by accumulating Na ϩ in endomembranous compartments in an NHX1-dependent manner. A signaling role for a Ca 2ϩ /CaM kinase in this stress pathway is likely. This is in distinction to wild type yeast, where activation of calcineurin results in transcriptional up-regulation of ENA1, which then pumps Na ϩ out across the plasma membrane.