Two classes of plant cDNA clones differentially complement yeast calcineurin mutants and increase salt tolerance of wild-type yeast.

The salt-sensitive phenotype of yeast cells deficient in the phosphoprotein phosphatase, calcineurin, was used to identify genes from the higher plant Arabidopsis thaliana that complement this phenotype. cDNA clones corresponding to two different sequences, designated STO (salt tolerance) and STZ (salt tolerance zinc finger), were found to increased tolerance of calcineurin mutants and of wild-type yeast to both Li+ and Na+ ions. STZ is related to Cys2/His2-type zinc-finger proteins found in higher plants, and STO is similar to the Arabidopsis CONSTANS protein in regions that may also be zinc fingers. Although neither protein has sequence similarity to any protein phosphatase, STO was able to at least partially compensate for all tested additional phenotypic effects of calcineurin deficiency, and STZ compensated for a subset of these effects. Salt tolerance produced by STZ appeared to be partially dependent on ENA1/PMR2, a P-type ATPase required for Li+ and Na+ efflux in yeast, whereas the effect of STO on salt tolerance was independent of ENA1/PMR2. STZ and STO were found to be expressed in Arabidopsis roots and leaves, whereas only STO message was detectable in flowers. An apparent increase in the level of STZ mRNA was observed in response NaCl exposure in Arabidopsis seedlings, but the level of STO mRNA was not altered by this treatment.

Upon exposure to elevated NaCl levels, plant cells are damaged due to the combined effect of ion toxicity and osmotic stress (1). Improving salt tolerance in crop plants is limited by our poor understanding of the nature of the pathways that allow plants to adapt to salt stress. Identification of genes associated with an increase in salt tolerance would be a first step in engineering halotolerance. The similarity of ion transport systems between plants and Saccharomyces cerevisiae and the genetic advantages of yeast have made this organism a good model system for identifying salt tolerance genes (1).
Yeast genes that are involved in salt tolerance have been identified by their ability to improve growth at increased gene dosage on elevated Na ϩ ion concentrations (2)(3)(4)(5) and by the identification of yeast mutants with increased sensitivity to Na ϩ ions (6 -10). Some of the genes identified encode kinases or phosphatases, suggesting that protein phosphorylation is one of the cellular mechanisms regulating salt tolerance in yeast (4, 6 -10). Calcineurin, a Ca 2ϩ /calmodulin-dependent Ser/Thr phosphoprotein phosphatase type 2B, has been identified as one such modulator of salt tolerance in yeast, because mutants lacking this phosphatase have increased sensitivity to Na ϩ ions (6 -8).
Calcineurin has been identified in a variety of eukaryotic organisms from yeast to mammals (reviewed in Refs. 11 and 12). In T lymphocytes, calcineurin is an essential component in a Ca 2ϩ -dependent signal transduction pathway leading to interleukin-2 production and subsequent T-cell activation (reviewed in Ref. 13). Calcineurin also appears to be required for dephosphorylation of Ca 2ϩ channels (14) and Na ϩ channels in mammalian brain cells (15,16) and for regulating N-methyl-Daspartate-receptor channels in adult neurons (17). Luan et al. (18) showed that in plant (Vicia faba) guard cells a calcineurinlike activity may play a role in regulating K ϩ channels.
Calcineurin is a heterodimer composed of a catalytic A subunit (60 kDa) and a regulatory B subunit (19 kDa) (reviewed in Refs. 12 and 19). Yeast contains two catalytic subunit genes, CNA1 and CNA2, and one regulatory subunit gene, CNB1 (20 -22). Calcineurin null mutants (i.e., null mutants in either both CNA genes (cna1 cna2) or in CNB1 (cnb1)) are sensitive to Na ϩ and Li ϩ ions, indicating that calcineurin is an essential component in the pathway regulating tolerance to these cations (6 -8). Mendoza et al. (7) showed that cnb1 mutants accumulate abnormally high levels of Li ϩ ions due to reduced expression of ENA1/PMR2 (23,24), which encodes a P-type ATPase involved in Na ϩ and Li ϩ efflux in yeast, and a failure of the K ϩ uptake system encoded by TRK1 (25) to convert to the high affinity state of K ϩ transport. These results suggest that calcineurin mediates NaCl tolerance in part by regulating expression of ENA1 and activity of ion transporters.
Cells lacking calcineurin activity are also sensitive to high pH (8) and Mn 2ϩ ions 1 and are more tolerant than wild-type cells to high concentrations of Ca 2ϩ (26). Calcineurin mutants have also been shown to fail to recover from ␣ factor-induced G 1 arrest (20,21) and lose viability during prolonged exposure to ␣ factor. 2 We took advantage of the salt-sensitive phenotype of yeast calcineurin mutants and the availability of an Arabidopsis cDNA library constructed in a yeast-Escherichia coli shuttle * This research was supported in part by National Science Foundation Grant NSF90-58284 (to C. S. G.). 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.
The vector to develop a screen for identification of Arabidopsis cDNA clones that confer increased salt tolerance on yeast. This procedure was successful in identifying two genes, STO (salt tolerance) and STZ (salt tolerance zinc finger), that complement the salt-sensitive phenotype of yeast calcineurin mutants. Although these genes have no apparent sequence similarity to calcineurin, STO complements all tested calcineurinrelated phenotypes in calcineurin-deficient yeast, whereas STZ only rescues the Na ϩ and Li ϩ sensitivity of these mutant cells. Interestingly, the salt-tolerant effect of STZ in yeast appears to depend on ENA function, whereas STO confers salt tolerance even in the absence of ENA.
DNA Manipulation and Sequencing-Subclones were produced by standard techniques (34). Single-stranded templates (35) or doublestranded plasmids were sequenced using the dideoxynucleotide-termination method (36) and Sequenase polymerase (U. S. Biochemical Corp.). Both strands of reported sequences were completely determined. Sequence analysis was performed using the Genetics Computer Group software package (37).
Screening for Arabidopsis cDNA Clones That Complement the Saltsensitive Phenotype of Calcineurin Mutants-Clones that rescued the salt-sensitive phenotype of calcineurin null mutants were isolated from an Arabidopsis cDNA library constructed in YES (a gift from John Mulligan, Stanford University) (38). This phage library was converted into a plasmid (URA3 CEN4 ARS1) library using the cre-lox site-specific recombination system (38). The plasmid library, in which inserts are transcribed under control of the GAL1 promoter, was transformed into the cna1 cna2 strain and into cnb1 cells essentially as described by Gietz et al. (33), and cells were plated on SC-Ura. Ura ϩ colonies were pooled, incubated in 2% Gal, 50 mM MES, 3 pH 5.5 for 3 h, and ϳ5 ϫ 10 5 cells were plated on YPGalRaf medium supplemented with 200 mM LiCl and incubated for 6 days at 30°C. Salt-resistant Ura ϩ colonies were restreaked on YPGalRaf medium supplemented with 260 mM LiCl and on YPD medium supplemented with 200 mM LiCl. Plasmids were recovered from colonies that exhibited galactose-dependent salt tolerance by transformation of E. coli cells with DNA isolated from the yeast strains (39) and selection on ampicillin plates. Plasmids were reintroduced into fresh preparations of the mutant strain used for their initial identification, and transformants were retested for galactose-dependent growth on medium containing 200 mM LiCl. 18 positive isolates were identified, all of which contained inserts encoding either STO or STZ. Differences in the 5Ј-and 3Ј-flanking sequences contained in different clones within each class confirmed that different clones derived from independent cloning events. Four plasmids were used for further experiments; plasmids pVL35 and pVL37 contain STZ cDNA inserts, and pVL36 and pVL38 contain STO cDNA inserts.
Construction of Control Plasmids-As a positive control to test for salt resistance in calcineurin-deficient yeast cells, a plasmid carrying the CNA2 coding sequence (20) under the control of the GAL1 promoter was constructed as follows. A BamHI/HindIII fragment containing the CNA2 coding region was isolated from YEp352-CNA2 (YEp-CNA2) (21) and inserted into these same sites of pBluescript KSϩ (Stratagene, La Jolla, CA), creating pVL10. An XhoI site was engineered 5Ј of the CNA2 coding sequence by the polymerase chain reaction (40) using the primer 5Ј-CCCCTCGAGTCACACAGGAGCCA-3Ј (based on the 5Ј-untranslated region of the CNA2 gene, with the XhoI site underlined) and the T3 primer (Stratagene, La Jolla, CA). The polymerase chain reaction product was inserted into the EcoRV site of pBluescript KSϩ, resulting in pVL12. A 2.7-kb XhoI fragment (the 3Ј XhoI site derives from the cloning vector) that included the CNA2 coding region was transferred from pVL12 into the XhoI site of pVL11, a random isolate from the plasmid-rescued YES library, replacing the insert of this clone. The resulting plasmid, pVL14 (URA3 CEN4 ARS1), referred to as pGAL1-CNA2, contains the GAL1 promoter driving expression of the CNA2 sequence. An empty vector was constructed as a negative control by removing the insert in pVL11 with XhoI followed by ligation, producing pVL15.
Plant Material and NaCl Treatments-Arabidopsis thaliana (L.) Heynh. ecotype Colombia (Co-3) was grown at 20 -25°C under continuous fluorescent and incandescent light as described (41). For salt treatment experiments, surface sterilized Arabidopsis seeds were sown in 100 ϫ 25-mm Petri dishes containing 20 ml of germination medium (42) and were grown for 12 days at 23°C under continuous fluorescent light. The concentration of NaCl in the medium was then increased to 0, 80, 140, and 220 mM (expected concentrations at equilibrium following diffusion) by adding a concentrated solution of NaCl to the medium surface as described by Lehle et al. (43). 30 h after exposure to NaCl, plants were harvested and frozen under liquid nitrogen.
Arabidopsis genomic DNA was isolated as described (46), digested with BamHI, EcoRI, or HindIII, fractionated on 0.8% agarose gels, and transferred to nylon membranes (47). Hybridization was performed in 2 ϫ SSPE, 0.1% SDS, 50 g/ml tRNA, 2 ϫ Denhardt's solution at 65°C. Final washes were as described for RNA blot analysis. Similar hybridization and wash conditions were used to determine the identity of some of the cDNA clones.
Hybridizations were performed with 2-3 ϫ 10 6 cpm/ml of cDNA fragments labeled with 32 P by the random priming method (48). STZ probes were prepared from a 0.65-kb XhoI/EcoRI fragment of pVL37 that lacks 243 bases of 3Ј coding sequence. STO probes were prepared from a 1.1-kb XhoI fragment of pVL38 and included the entire STO cDNA insert.

Isolation of Arabidopsis cDNA Clones That Confer Increased Salt Tolerance to Yeast Calcineurin Mutants-To identify
Arabidopsis clones that rescue the salt-sensitive phenotype of yeast calcineurin mutants, we transformed cna1 cna2 double null mutants and cnb1 null mutants with an Arabidopsis cDNA library that contains cDNA inserts under the control of the galactose-inducible GAL1 promoter (38). To select for salt-resistant colonies, transformants were plated under inducing conditions on rich medium containing LiCl. Transport of Li ϩ across the plasma membrane in yeast is mediated by the same transport system used for Na ϩ (23,49). However, Li ϩ is more toxic than Na ϩ and can therefore be used at lower concentrations than Na ϩ , resulting in a higher growth rate of tolerant cells (7,8). Because the frequency of spontaneous reversion to salt tolerance was high under these selective conditions (ϳ1 in 10,000), salt-tolerant putative transformants were tested for uracil prototrophy and for their ability to grow on medium containing LiCl and galactose but not on medium containing LiCl and glucose. A total of ϳ1 ϫ 10 6 transformants represent-ing 5 ϫ 10 5 independent cDNA clones were screened in each mutant background. In this screen, 34 of 111 salt-resistant colonies showed galactose-dependent salt tolerance in the cna1 cna2 mutant background, and 39 of 104 showed this growth phenotype in the cnb1 mutant background. Plasmids were recovered from the putative positive transformants and reintroduced into the original mutant strains. A total of eight cDNA clones (corresponding to one STO and seven STZ cDNA inserts, see below) retested as increasing salt tolerance of cna1 cna2 cells and ten cDNA clones (corresponding to four STO and six STZ cDNA inserts) retested as increasing salt tolerance of cnb1 cells. Partial sequencing of the first seven positive clones (five from cna1 cna2 mutants and two from cnb cells) revealed that each corresponded to one of two sequences, designated STO and STZ. Hybridization of subsequent positive clones with probes prepared from either STO or STZ cDNA inserts revealed that all clones included either STO or STZ (or possibly closely related sequences). cDNAs from both sequences were recovered in both mutant backgrounds. A total of five STO and thirteen STZ cDNAs were identified of which at least four and eleven were independent isolates, respectively.
On medium containing LiCl and galactose, expression of STO in calcineurin-deficient strains produced colonies that grew faster than either wild-type cells carrying the vector control or cna1 cna2 cells expressing CNA2 under the control of the GAL1 promoter ( Fig. 1). In contrast, calcineurin-deficient cells harboring the STZ expression plasmid grew at approximately the same rate as strains containing calcineurin (Fig. 1). Salt tolerance of mutant cells producing STO or STZ was abolished on medium containing glucose, confirming that expression from the cDNA clones was galactose-dependent ( Fig. 1). STO and STZ also conferred increased tolerance to NaCl in the presence of galactose on both cna1 cna2 and cnb1 mutants (data not shown). As was observed with LiCl, at elevated NaCl concentrations (e.g., 600 mM) expression of STO in the mutant backgrounds increased salt tolerance more than either STZ or GAL1-CNA2 (data not shown).
Sequence Analysis of STZ and STO- Fig. 2A shows the deduced amino acid sequence of the protein encoded by the STZ clones. STZ consists of 228 amino acids with a calculated molecular mass of 24.6 kDa and an estimated pI of 8.3 ( Fig. 2A). STZ is 37-68% identical in amino acid sequence to a family of petunia DNA-binding Cys 2 /His 2 -type zinc finger proteins associated with flowers (termed EPF) being most similar to EPF2-7 (68%) ( Fig. 2A; Refs. 50 and 51). STZ also shows 47% amino acid identity with WZF1 (52), a wheat zinc finger DNA-binding protein that is primarily expressed in the root apex ( Fig. 2A).
STZ, like WZF1 and members of the EPF family, contains two Cys 2 /His 2 zinc finger motifs ( Fig. 2A). As is characteristic for zinc fingers in this family, the Cys-Xaa 2 -Cys and His-Xaa 3 -His regions in each zinc finger in STZ are separated by 12 residues including invariant hydrophobic residues, Phe and Leu (53). Both zinc fingers in STZ and in the other plant members of this family contain 6 conserved consecutive residues (QALGGH), which appear to contact DNA (54, 55). The region between the two zinc fingers ranges from 36 to 61 amino acid residues in these plant protein and is considerably larger than the 7-8 residues observed in many zinc finger proteins reported from yeast and animals (50 -52, 56). The 35-residue separation between the two zinc fingers of STZ is similar to that of the petunia proteins. The basic B-box, hypothesized to be a nuclear localization signal (51), is present in STZ and in all Cys 2 /His 2 zinc finger proteins so far reported from plants.
The STO clones encode a 27.6-kDa hydrophilic protein of 249 amino acids with a calculated pI of 5.4 (Fig. 2B). In a search of the sequence data bases the Arabidopsis CONSTANS (CO) protein (57) showed the greatest similarity to STO. Similarity was confined to two regions (Fig. 2B), which have been hypothesized to represent zinc fingers, although no biochemical information supporting this hypothesis is yet available (57). Near the C terminus of STO is a highly basic region followed by acidic amino acid residues. CO also contains a relatively basic region followed by an acidic region near the C terminus, but there is no sequence similarity between the two proteins in these regions (data not shown). Data base searches also revealed several randomly sequenced cDNAs (EST) with similarity to STO. Arabidopsis clones H36917 and N38572 are identical to STO over regions of several hundred bases, indicating that they represent independent clones of this gene. Arabidopsis clone ATTS3129 was similar to STO but included enough differences to indicate that it derives from a related but distinct gene. Rice clones Ricr1479a, Ricr15772a, Ricc10131a, and Ricr2967a appear to encode at least two different related proteins.

STO and STZ cDNAs Increase Salt Tolerance of Wild-type
Yeast-The increased colony size on salt-containing medium of yeast calcineurin mutants producing STO relative to an isogenic wild-type strain (Fig. 1, A and B) prompted us to investigate effects of STO or STZ expression on salt tolerance of wild-type yeast. Fig. 3 shows that wild-type strain producing STO or STZ grew faster in the presence of galactose and LiCl (260 mM) than the same strain harboring either a vector control or pGAL1-CNA2. When NaCl (600 -750 mM) was used in place of LiCl, similar results were obtained (data not shown). No growth differences were observed on equivalent medium lacking salt (Fig. 3).
In liquid medium (YPGalRaf) containing 260 mM LiCl, the wild-type strain harboring plasmids that encode either STO or STZ grew at faster rates (4.7 Ϯ 0.2 and 4.5 Ϯ 0.2 h doubling times, respectively) than did the same strain harboring a vector control (6.0 Ϯ 0.5 h doubling time). Doubling times were determined from four data sets for each strain. Similar differences were observed in medium containing 0.7 M NaCl. As observed on solid medium, in the absence of salt (Fig. 3, galactose, no LiCl) no differences were detected between growth rates of wild-type strains in the presence or the absence of plasmids expressing either STO or STZ (all doubling times, approxi-

FIG. 2. Protein sequence of STZ and STO.
A, deduced amino acid sequence of STZ aligned with the published sequences of petunia zinc finger DNA-binding proteins EPF2-7, Epf2-5b, and EPF2-4 (50), petunia EPF1 (51), and wheat zinc finger DNA-binding protein WZF1 (52). The basic B-box in STZ is indicated by an overline, and zinc finger motifs are indicated by a double underline. Dots indicate gaps introduced to allow for optimal alignment of the sequences. Single-letter codes for amino acid residues are used and asterisks indicate termination codons for translation. B, deduced amino acid sequence of STO aligned with the putative zinc finger regions of the Arabidopsis CO protein (57). Basic amino acids in a basic region of the protein are overlined, and the acidic region is underlined. Double underlines indicate amino acid residues that were altered in CO mutant proteins.
Vertical bars indicate identical amino acids in the two sequences and colons and periods indicate more or less similar amino acids, respectively.

FIG. 3. Expression of STO or STZ increases LiCl tolerance of wild-type yeast.
Wild-type (WT; YPH499) strain transformed with plasmids carrying the sequence coding for STO (pVL36), STZ (pVL35), or GAL1-CNA2 (pVL14) or containing a vector control (pVL15) were streaked onto YPGalRaf medium containing 0 or 260 mM LiCl and incubated at 30°C for 3 days (galactose, no LiCl) or 8 days (galactose ϩ LiCl). mately 2.8 h). These data indicate that STO and STZ provide a growth advantage to wild-type yeast only in the presence of salt.
STO and STZ Correct Other Effects of Calcineurin Mutations-Calcineurin mutants are not only sensitive to Li ϩ and Na ϩ ions but also to elevated levels of Mn 2ϩ ions. 1 In addition, calcineurin mutants are more tolerant to elevated Ca 2ϩ levels than are wild-type cells (26) and are unable to recover during constant exposure to the ␣ factor mating pheromone (20,21). To address the question of STO and STZ function, it was of interest to determine if these proteins affected these other calcineurin phenotypes. As shown in Fig. 4, production of STO increased Mn 2ϩ tolerance of cna1 cna2 and cnb1 mutants to near wild-type levels. STO slightly increased the sensitivity of cna1 cna2 and cnb1 cells to elevated (300 mM) CaCl 2 concentrations, thus displaying a similar phenotypic effect as presence of calcineurin activity (data not shown). In addition, expression of STO led to a small, but reproducible recovery of cna1 cna2 and cnb1 mutants from ␣-factor-induced growth arrest (data not shown). In contrast to STO, expression of STZ in calcineurin-deficient cells produced increased sensitivity to both Mn 2ϩ (Fig. 4) and Ca 2ϩ ions (data not shown), and had no effect on recovery of cells upon ␣-factor treatment. STO, therefore, can at least partially compensate for the absence of calcineurin in all tested processes, whereas STZ compensates for some, but not all, effects of calcineurin deficiency.
Role of ENA in Effects of STO and STZ-To test whether effects of STO and STZ are dependent on the ENA gene products, we determined the effects of STO and STZ expression on salt sensitivity of cells lacking ENA activity. Fig. 5 shows that STO increased Li ϩ tolerance of ena mutants, whereas expression of STZ had no detectable effect. Both STO and STZ increased tolerance of an isogenic wild-type strain to 60 mM LiCl (Fig. 5), demonstrating that both gene products function in this yeast background. Together, these results suggest that ENA activity is required for an increase in salt tolerance by STZ, whereas it is not essential for STO function.
Arabidopsis Genomic DNA Blot Analysis-Blots containing Arabidopsis genomic DNA digested with BamHI, EcoRI, or HindIII were hybridized to 32 P-labeled STO or STZ cDNA probes at high stringency (see "Experimental Procedures"). Fig.  6 shows that a single band hybridized to STO cDNA in BamHI and EcoRI digests with an intensity similar to a single-copy control. Two less intense bands were visible in the HindIII digest. These data suggest the presence of a single STO gene in Arabidopsis. In contrast, the STZ probe hybridized strongly to one gene and more weakly to two to four additional genes (Fig.  6), consistent with our hypothesis that STZ is a member of a multi-gene family.
Expression Patterns of STO and STZ Genes in Arabidopsis Organs-On blots of total RNA from roots, leaves, and flowers, STO hybridized to a 1.1-kb transcript that was similar in size to the isolated STO cDNA clones (Fig. 7A). The steady-state level of STO mRNA was highest in leaves and was significantly lower in roots and flowers (Fig. 7A). In contrast, two hybridizing bands of 0.9 and 0.7 kb were observed when an equivalent RNA blot was hybridized to STZ (Fig. 7A). Both STZ-hybridizing transcripts were present at 3-fold higher levels in roots than in leaves, and neither transcript was detected in flowers  ). C, the wild-type strain (W303.1B) containing pVL15 (WT/cont.), pVL35 (WT/STZ), or pVL36 (WT/STO) was plated on YPGalRaf and incubated at 30°C for 2 days. D, the same strains in C plated on YPGalRaf containing 60 mM LiCl and incubated for 9 days. All the above experiments were performed in parallel with independent transformants of each line, and essentially identical results were obtained (data not shown). The interpretation of the results was sometimes complicated by the fact that all tested lines spontaneously gave rise to fast-growing Li ϩ -tolerant colonies at a significant frequency (not shown).
FIG. 6. Genomic DNA blot analysis of STO and STZ. 1.5 mg of genomic DNA from Arabidopsis ecotype Colombia was digested with BamHI (B), EcoRI (R), or HindIII (H) and fractionated as described under "Experimental Procedures." Blots were hybridized with probes prepared from either STO or STZ cDNAs, washed as described under "Experimental Procedures," and exposed to x-ray film for 6 days. Numbers indicate the sizes of marker DNA fragments in kilobases. (Fig. 7A). When the RNA blot hybridized with STZ was washed at higher stringency (0.3 ϫ SSPE, 0.1% SDS at 75°C), hybridization to the lower molecular mass band was eliminated, whereas the higher molecular mass band was only slightly reduced in intensity (data not shown). These results indicate that the high molecular mass band (0.9 kb) is encoded by STZ with the other RNA species deriving from a related gene (consistent with our DNA blot data; Fig. 6). The 0.9-kb transcript was similar in size to the isolated STZ cDNA. Hybridization of the 0.7-kb message at 65°C but not at 75°C indicates that this message is between 85 and 90% identical in sequence to STZ. Prior hybridization of these blots with an Arabidopsis CyP (ROC1) probe demonstrated equal loading in all lanes (58).
Expression of STO and STZ in NaCl-treated Arabidopsis Plants-Because the Arabidopsis STO and STZ cDNAs conferred increased salt tolerance in yeast (Figs. 1 and 3), it was of interest to determine if the corresponding genes were induced in plants exposed to elevated levels of salt. Blots were prepared with RNA isolated from control and NaCl-treated Arabidopsis plants and hybridized with either STO or STZ cDNA probes (Fig. 5B). After normalizing for the amount of RNA loaded in each lane, we determined that the steady-state levels of STO mRNA were essentially unchanged in plants treated with increasing NaCl concentrations (Fig. 5B). When a similar blot was hybridized with the STZ probe, two hybridizing bands were observed (Fig. 5B), similar to the results observed in Fig.  5A with the same probe. On the basis of a higher stringency wash, the 0.9-kb transcript corresponded to STZ (data not shown). The steady-state levels of the 0.9-kb transcript were similar in plants treated with 0, 80, or 140 mM NaCl but were measured as being approximately 2.5-fold higher in plants treated with 220 mM NaCl (Fig. 5B, STZ probe). In contrast, the steady-state levels of the 0.7-kb fragment were low in 0 mM NaCl-treated plants and increased approximately 3-fold in the presence of 80 mM NaCl. At the higher NaCl concentrations (140 and 220 mM) the 0.7-kb mRNA levels were lower than at 80 mM NaCl being ϳ2-fold higher than in the 0 mM NaCl treatment. In conclusion, unlike STO, both the STZ gene and the gene encoding the 0.7-kb transcript that hybridized to STZ, appear to respond to NaCl in the medium and seem to respond differentially at any given salt concentration.. DISCUSSION We have identified two Arabidopsis genes, STO and STZ, which confer increased tolerance to LiCl and NaCl on yeast calcineurin mutants and wild-type yeast. Previous work supports the notion that plants and yeast have developed similar mechanisms for adapting to salt stress and that both organisms have conserved salt-sensitive components in cellular metabolism (1). Serrano and co-workers have identified yeast genes HAL1 (2), HAL2 (3), and HAL3 (5) in S. cerevisiae by selecting for genes whose overexpression led to improved growth on NaCl medium. HAL1 and HAL3 encode novel proteins that appear to modulate cation transport systems (2,5). A HAL1 homolog was detected in plants where it is induced by NaCl and abscisic acid, a plant hormone known to mediate adaptation of plants to osmotic stress (2). HAL2 encodes a 3Ј(2Ј),5Ј-bisphosphate nucleotidase required for recycling of adenine nucleotides in sulfate transfer reactions, and its phosphatase activity was inhibited by Li ϩ and to a lesser extend by Na ϩ ions (59). Tomato also contains a 3Ј,5Ј-bisphosphate nucleotidase that is sensitive to these ions (59). Our findings further support the use of yeast as a model organism for identifying genes that might be relevant to salt tolerance in plants.
Calcineurin is known to play a role in adaptation to elevated Na ϩ and Li ϩ concentrations in yeast (7,8). When grown in the presence of NaCl or LiCl, cna1 cna2 and cnb1 mutants accumulate abnormally high levels of Na ϩ or Li ϩ ions. STO and STZ have no sequence similarity to calcineurin and therefore are unlikely to be directly replacing the phosphatase activity of calcineurin in this salt tolerance cascade.
We observed significant differences between the effects of expression of STO or STZ on other consequences of calcineurin deficiency in yeast (20,21,60). In a calcineurin-deficient background both proteins were able to restore sensitivity to Ca 2ϩ ions. In contrast, only STO allowed recovery of these cells from ␣-factor arrest, and although STO increased resistance to Mn 2ϩ ions, STZ had the opposite effect. Thus, expression of STO in the calcineurin mutants produced phenotypes similar to those resulting from expression of active calcineurin, suggesting that STO might modulate a calcineurin-dependent pathway in yeast. The observation that STZ exhibits a phenotype only in response to cations suggests that STZ is involved in ion adaptation but does not eliminate the possibility that this gene product is regulating a subset of calcineurin-mediated pathways.
ENA1 encodes the major sodium and lithium efflux system in yeast and is the first repeat of a tandem array of four genes encoding almost identical proteins (23,24,61). Although the four ENA genes contribute to salt tolerance, the contribution of ENA1 to salt tolerance is highest because, unlike the other genes, which are expressed constitutively and at low levels, the expression of ENA1 can be high and is subject to a complex regulation (5, 7). 4 STZ did not rescue the Li ϩ sensitivity of a yeast strain lacking the ENA locus but did improve growth of an isogenic wild-type strain, suggesting that STZ is at least partially dependent on ENA gene products for manifestation of salt tolerance. The inducible expression of ENA1 makes this gene a potential target for the action of STZ. Our finding that STO increased tolerance to Li ϩ ions of ena mutant cells suggests that ENA activity is not essential for STO function. Because the effects of calcineurin on salt tolerance do not 4 A. Rodriguez-Navarro, personal communication. , and roots (R) were hybridized with probes prepared from either STO or STZ cDNAs as described under "Experimental Procedures." B, expression of STO and STZ in NaCl-treated plants. Arabidopsis plants were grown for 12 days on germination medium (see "Experimental Procedures") at 23°C under continuous fluorescent light at which point the concentration of NaCl was increased to 0, 80, 140, or 220 mM (expected concentrations at equilibrium following diffusion) by adding a concentrated solution of NaCl to the medium surface. Total RNA was extracted from plants after 30 h of exposure to NaCl, and 10 g of each treatment was fractionated on gels, transferred to nylon membranes, and hybridized to STO or STZ probes. The ϳ1.5-fold increase in STO mRNA concentration at 140 and 220 mM NaCl relative to 0 and 80 mM NaCl can be accounted for by the amount of total RNA in each lane as visualized by the amounts of ribosomal RNA on the gel. The blot hybridized to the STZ probe was prepared from a gel where lanes from all salt treatments contained essentially identical amounts of ribosomal RNA. Blots were exposed to x-ray film for 6 days. Approximate sizes of hybridizing bands are indicated in kilobases. appear to be mediated entirely through ENA1 but also via regulation of TRK1 activity (7), the effect of STO in ena mutant cells is not inconsistent with STO modulating a calcineurin-dependent pathway.
The primary structure of the STZ protein provides insight into its possible activity. STZ belongs to a family of DNAbinding Cys 2 /His 2 zinc finger proteins, which have thus far only been identified in plants, suggesting that STZ is also a DNA-binding protein. Thus, it is possible that STZ may act to directly regulate transcription and could represent the first eukaryotic transcription factor identified to play a role in salt tolerance. The effects of STZ on salt tolerance in yeast may be the result of regulation of expression of a critical gene (or genes) involved in the yeast salt-stress response. The effect of deletion of ENA genes on STZ effects could indicate that the ENA genes are directly or indirectly regulated by STZ.
Homologs to STZ have been identified in both monocots and dicots. STZ is most similar to petunia EPF2-7 (68% amino acid identity) which, in contrast to STZ and another STZ-like gene, is preferentially expressed in floral organs and is expressed at very low levels in roots and leaves (50). Thus, EPF2-7 is unlikely to be the petunia ortholog of STZ. STZ is more similar to the wheat member of this family of proteins than it is to the most diverged member of the petunia sequences (EPF1) (51), indicating that this is a family whose divergence precedes the separation of monocots and dicots. It is worth noting that in our screen for genes that confer increased salt tolerance, we only recovered cDNAs encoding STZ and not other distant members of this family. Blots containing Arabidopsis RNA hybridized at high stringency identified an additional transcript that is at least 85% identical to STZ and that has a similar expression pattern to STZ. Because some STZ cDNA clones were identified as being members of this class only by hybridization under high-stringency conditions, it is possible that some of these clones do not encode STZ but encode a closely related sequence. Our data suggest that STZ and a related gene might be responsive to NaCl in Arabidopsis. The high expression of STZ and an STZ homolog in Arabidopsis roots, the organ through which all ions enter the plant, and the induction of these genes by NaCl support the hypothesis that STZ controls steps in Na ϩ ion balance in plants.
Sequence comparison of STO to the existing data base indicates that STO also contains putative zinc finger regions that are homologous to the putative zinc finger regions of the Arabidopsis CO protein. Although there is yet no demonstration that CO is a DNA-binding protein, the similarity of the motifs observed in CO and STO to the zinc finger motifs of the GATA1 protein family (62) indicates that both proteins may be transcription factors. We also found that Arabidopsis contains an additional STO-like gene and rice contains at least two STOlike genes. STO was expressed at the highest levels in leaves but was also expressed in roots and flowers. Our data suggest that STO expression in Arabidopsis was insensitive to NaCl treatments under the conditions tested.
Increased salt tolerance has previously been engineered in tobacco using either a bacterial gene encoding mannitol 1-phosphate dehydrogenase (63) or a ⌬-pyrroline-5-carboxylate synthetase gene from another plant species (64). Expression of STO and STZ not only increased salt tolerance of calcineurin mutants but also of wild-type yeast. Because steps in the salt stress response and salt-sensitive components in metabolic pathways appear to be conserved between yeast and plants (1), overexpression of STO and STZ has the potential to be another method of increasing salt tolerance in plants.