In Saccharomyces cerevisiae, the Inositol Polyphosphate Kinase Activity of Kcs1p Is Required for Resistance to Salt Stress, Cell Wall Integrity, and Vacuolar Morphogenesis*

A problem for inositol signaling is to understand the significance of the kinases that convert inositol hexakisphosphate to diphosphoinositol polyphosphates. This kinase activity is catalyzed by Kcs1p in the yeast Saccharomyces cerevisiae. Akcs1Δ yeast strain that was transformed with a specifically “kinase-dead” kcs1p mutant did not synthesize diphosphoinositol polyphosphates, and the cells contained a fragmented vacuolar compartment. Biogenesis of the yeast vacuole also required another functional domain in Kcs1p, which contains two leucine heptad repeats. The kinase activity of Kcs1p was also found to sustain cell growth and integrity of the cell wall and to promote adaptive responses to salt stress. Thus, the synthesis of diphosphoinositol polyphosphates has wide ranging physiological significance. Furthermore, we showed that these phenotypic responses to Kcs1p deletion also arise when synthesis of precursor material for the diphosphoinositol polyphosphates is blocked in arg82Δ cells. This metabolic block was partially bypassed, and the phenotype was partially rescued, when Kcs1p was overexpressed in the arg82Δ cells. This was due, in part, to the ability of Kcs1p to phosphorylate a wider range of substrates than previously appreciated. Our results show that diphosphoinositol polyphosphate synthase activity is essential for biogenesis of the yeast vacuole and the cell's responses to certain environmental stresses.

Environmental and physiological stresses are constantly challenging living organisms. For microorganisms in particular, their immediate environmental conditions, temperature, salinity, and nutrient availability, can fluctuate considerably. The yeast Saccharomyces cerevisiae is a widely used model system for studying the molecular processes that sense and initiate responses to these stressful changes. This work is of general biological interest, because the molecular mechanisms that underlie these processes are highly conserved between yeasts and higher eukaryotes. It has been known for some years that the membrane-bound inositol lipids and their hydrolysis by phospholipase C (Plc1p) both play important roles in adaptations to environmental stress. For example, disruption of the genes encoding inositol lipid phosphatases results in cell wall lysis, vacuolar fragmentation, and impaired homeostatic responses to osmotic stress (4,5). The inositol lipid kinases have also been shown to be required for cell wall integrity and the maintenance of vacuolar morphology (6,7). However, it has not previously been considered that the soluble inositol phosphates might also be important in mediating any of these homeostatic responses.
In yeast, exposure to the toxic effects of elevated salinity (8) and osmotic stress (9) both initiate protective responses that involve vacuolar function (10,11). A vacuolar defect was recently found to arise in S. cerevisiae following disruption of approximately half of the KCS1 gene (12). We have therefore now investigated whether this particular vacuolar defect is associated with any impairment to the homeostatic responses to high salt and osmotic stress. Furthermore, an intact cell wall is essential for yeast to successfully adapt to environmental stress (13). Thus, we have also studied whether Kcs1p has a role in maintaining the integrity of the cell wall. These studies into environmental stress responses represent new areas of research into the function of the KCS1 gene.
One of the functional characteristics of Kcs1p is its diphosphoinositol polyphosphate synthase (DINS) 1 activity (2,12), which yields a specific class of inositol phosphates distinguished by their diphosphate groups (diphosphoinositol tetrakisphosphate (PP-InsP 4 ), diphosphoinositol pentakisphosphate (PP-InsP 5 ), and bis-diphosphoinositol tetrakisphosphate ((PP) 2 -InsP 4 ); see Fig. 1). InsP6K1 is a mammalian homologue of Kcs1p (2); InsP6K1 is pleiotropic, since, in addition to its own DINS activity, it also interacts with a guanine nucleotide exchange factor for Rab3A (14). Kcs1p is at least as likely to be multifunctional, since it is much larger in size (120 kDa (15)) than are the mammalian InsP 6 kinases (50 -60 kDa (2,16)). Indeed, the KCS1 gene was first identified in a screen for suppressing the temperature sensitivity of the pkc1-4 allele (15). Kcs1p also has two groups of four heptad repeats of leucine residues. The latter were suggested to be leucine zippers (15), which in other contexts direct the homo-and heterodimerization of proteins (17). In the current study, we provide the first evidence that there is an important function for these leucine repeats in Kcs1p.
Since Kcs1p may be multifunctional, its severe disruption in an earlier report (12) could have yielded abnormal phenotypes that were not all directly related to its kinase activity. Thus, in this study, we have investigated the physiological impact of specifically targeting the DINS activity of Kcs1p. Our approach was first to create kcs1⌬ cells and then transform them with plasmids encoding mutants in which the inositol phosphate kinase activity was specifically modified.
Diphosphoinositol polyphosphates, both in yeast and in mammalian cells, are synthesized from the InsP 5 and InsP 6 precursor pools (12, 18 -20). Thus, we considered the possibility that the phenotypic consequences of eliminating the DINS activity of Kcs1p from cells could also arise independently from genetic manipulations that deplete the InsP 5 and InsP 6 pools. ARG82 2 is one gene that is required for InsP 5 and InsP 6 synthesis. Arg82p phosphorylates Ins (1, 4, 5)P 3 to InsP 5 (Refs. 2, 3, and 20; see Fig. 1). In addition, Arg82p regulates the expression of enzymes in the arginine metabolism pathway (21,22). This transcriptional role is achieved by Arg82p binding and sequestering two MADS-box proteins, Arg80p and Mcm1p, thereby preventing their rapid degradation (23). These two transcription factors then form a complex on DNA with Arg81p, the arginine sensor (22). In the current study, we use specific genetic manipulations that distinguish the role of Arg82p as a regulator of arginine gene expression from its separate function as a supplier of precursor material for the synthesis of diphosphoinositol polyphosphates. As a consequence, we demonstrate that inositol phosphate phosphorylation by Arg82p has more extensive physiological consequences than have previously been appreciated.

EXPERIMENTAL PROCEDURES
Strains and Media-Escherichia coli strain XL1-B was used for plasmid amplification and for in vitro mutagenesis. The BY4709 (MAT␣, ura3) yeast strain (24) was used to create a series of gene deletions. The long flanking homology strategy (25) was used to perform deletion of ARG82 and KCS1. A long flanking homology replacement cassette was synthesized using a two-step PCR, leading to the kanMX4 cassette flanked by about 500 bp corresponding to the promoter and terminator regions of the target genes. The DNA fragments containing the different cassettes were used to transform strain BY4709 on rich medium plates containing 200 g/ml of Geneticin. The correct targeting of the deletions in G418 r transformants was verified by PCR, using whole cells as a source of DNA, and appropriate primers. The following strains were obtained: arg82::kanMX4 (03127c) and kcs1::kanMX4 (4709 kcs1⌬).
Unless otherwise noted, all yeast strains were grown on minimal medium (pH 6.5), which contained 3% glucose, vitamins, and mineral traces. Unless otherwise stated, the nitrogen source was 0.02 M (NH 4 ) 2 SO 4 (M.am medium).
Genetic Manipulations-The low copy pFL38 plasmid was used in this work to bear wild type and mutated arg82 (26) and kcs1 genes. Since these proteins are expressed under their native promoter and on a low copy number plasmid (one or two copies in the cell), the amount of each protein is comparable with the amount produced by the genomic copy of the wild type strain. The wild type KCS1 gene was cloned on a pFL38 plasmid by insertion of a 4.95-kb EcoRI-EcoRI fragment of KCS1, synthesized by PCR using appropriate oligonucleotides as primers with EcoRI restriction sites, and genomic wild type DNA as template, yielding plasmid pFV241. Oligonucleotides were used to create substitution by in vitro mutagenesis on double-stranded DNA from plasmid pFV241 using the QuikChange site-directed mutagenesis kit from Stratagene (Amsterdam, The Netherlands). In this plasmid, the following amino acid changes were created: L794A/L801A/L857A/ L864A (pFV198) and S887A/L888A/L889A (pFV217). To check that each mutant protein was stable when expressed, the same Kcs1p proteins were expressed fused at the C terminus to the V5 epitope tag, using the pYES2 plasmid (Invitrogen), with a GAL10 promoter. Plasmids pFV249 (GAL10-KCS1-V5), pFV251 (GAL10-kcs1L794A/L801A/ L857A/L864A-V5), and pFV252 (GAL10-kcs1S887A/L888A/L889A-V5) were obtained by insertion in the EcoRI restriction site of pYES2 plasmid of EcoRI KCS1 DNA fragments synthesized by PCR using appropriate oligonucleotides and as templates the plasmids pFV241, pFV217, and pFV198, respectively. We completely sequenced all of the wild type and the different mutated genes to ensure that no additional mutations had been introduced.
To overexpress ARG82 and KCS1 genes from S. cerevisiae, we fused their coding sequence to the tet promoter present in plasmid pCM262 (gift from E. Herrero, Universitat de Lleida, Spain). For ARG82, a 1070-bp BamHI-BamHI fragment was synthesized by PCR using appropriate oligonucleotides as primers with BamHI restriction sites and DNA from plasmid pFV145 as template, blunt-ended with T4 DNA polymerase and inserted in the PmeI restriction site of plasmid pCM262, yielding plasmid pFV186. For KCS1, a 4.15-kb EcoRI-EcoRI fragment was synthesized by PCR using appropriate oligonucleotides as primers with EcoRI restriction sites and DNA from plasmid pFV241 as template, blunt-ended with T4 DNA polymerase and inserted in the PmeI restriction site of plasmid pCM262, yielding plasmid pFV187.
Western Analysis-Exponentially growing cells (25 ml) were harvested by centrifugation. The cell pellet was resuspended in 300 l of buffer containing 20 mM Tris-HCl, pH 8.0, 50 mM ammonium acetate, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and a protease inhibitor mixture tablet (catalog no. 1697498; Roche Diagnostics). Next, 300 l of 20% (w/v) trichloroacetic acid and 300 mg of glass beads were added, and cells were disrupted in a Bead-Beater. Proteins were extracted by centrifugation and resuspended in 200 l of gel loading buffer containing 3.5% (w/v) SDS, 80 mM Tris, 8 mM EDTA, 14% (v/v) glycerol, 1 mM phenylmethylsulfonyl fluoride, and blue bromphenol. Samples were heat-treated (100°C, 10 min). After electrophoresis on NuPAGE 4 -12% Bis-Tris gels, followed by transfer to Highbond membrane, the Kcs1p proteins were visualized using antibodies raised against the V5 epitope. Arg82p proteins were detected using polyclonal antibodies against recombinant GST-Arg82p, which were raised in the mouse, and prepared as previously described using glutathione-Sepharose 4B beads (26). Western blots were performed according to a standard chemiluminescent protocol provided with the WesternBreeze kit from Invitrogen (Merelbeke, Belgium).
Other Enzyme Assays-Ornithine carbamoyltransferase (OTCase) was assayed as described previously (27). Data for this enzyme are presented as repression factors. The repression factor for OTCase corresponds to the ratio of OTCase-specific activities of the arg82⌬ strain grown on M.am versus OTCase-specific activities of the different strains grown on M.am plus 1 mg/ml arginine. Alkaline phosphatase release from cells was recorded on plates that contained M.am plus 40 g/ml BCIP (Sigma). The hydrolysis of the dye yields a blue stain.
Vacuole Analyses-Cells were grown at 29°C in YPD medium (2% peptone, 1% yeast extract, 2% glucose). Cells from 5 ml of early logarithmic phase cultures (A 600 ϭ 0.4 -0.6) were collected and resuspended in 100 l of medium containing 50 mM sodium citrate buffer, pH 5.0, 2% glucose, 10 M carboxy-DCFDA. Cells were incubated at room temperature for 20 min, whereupon the membrane-permeable carboxy-DCFDA enters the vacuoles, the acetate groups are hydrolyzed, and the non-membrane-permeable and fluorescent carboxy-DCF highlights the vacuolar compartment. Intracellular localization of carboxy-DCF was examined with a Zeiss LSM510 microscope with a 100ϫ oil immersion objective (NA ϭ 1.4) under transmitted light or with epifluorescence (488-nm excitation with an argon laser, HFT 488-nm dichroic, and 505-nm long pass filter). The images that are shown in this study are representative of at least 100 cells from every strain.

RESULTS
The Effect of Deletion of the KCS1 Gene upon DINS Activity and Vacuolar Morphology-We initially studied the consequences of deleting the entire KCS1 gene. Since Kcs1p has diphosphoinositol phosphate kinase activity, we monitored the effect of the gene deletion upon the cellular levels of PP-InsP 5 and (PP) 2 -InsP 4 . Individual inositol phosphates were resolved by HPLC of yeast cell extracts that had been prelabeled with [ 3 H]inositol (Fig. 2). To ensure that all of the inositol polyphosphate pools were radiolabeled to equilibrium, their saturation with [ 3 H]inositol during the time course of our experiments was ensured by labeling for at least eight cell generations. As shown previously (28,29), InsP 1 and InsP 6 account for more than 99% of the total spectrum of inositol phosphates in yeast cells (Fig. 2). However, there is considerable interest in the functional significance of the other, more minor inositol phosphates (28), particularly the diphosphoinositol polyphosphates (12). In wild-type cells, we detected one peak of (PP) 2 -[ 3 H]InsP 4 and two peaks of PP-[ 3 H]InsP 5 (Fig. 2). The kcs1⌬ cells showed 93% lower steady-state levels of PP-InsP 5 and (PP) 2 -InsP 4 compared with wild-type cells ( Fig. 2; Table I), confirming the importance of the DINS activity of Kcs1p.
The vacuolar compartment of S. cerevisiae was identified by fluorescence microscopy. Wild-type cells were found to have the normal complement of one large vacuole (Fig. 3). In contrast, the vacuolar compartment in the kcs1⌬ strain was fragmented ( Fig. 3). In addition to being about twice as large as the wildtype cells (Fig. 3), the kcs1⌬ cells also exhibited a growthimpaired phenotype that was evident at either 30 or 37°C (Fig. 4A).
Cell Wall Integrity and the Homeostatic Responses to Salt Stress in the kcs1⌬ Strain-Ion transport processes in the vacuolar membrane are thought to help protect the cell from toxicity caused by exposure to high salt as well as affording protection against sudden osmotic challenges (10, 11). Thus, we investigated whether the defective vacuolar compartment in the kcs1⌬ strain compromised these stress responses. We discovered that these cells were unusually sensitive to NaCl (either 0.6 or 0.8 M), which strongly inhibited colony formation (Fig. 4B). On the contrary, the kcs1⌬ strain was unaffected by an osmotic challenge from 1 M sorbitol (Fig. 4A). Thus, we conclude that these cells have normal osmotic stress responses, but resistance to salt stress is compromised. Another new finding was the fragility of these kcs1⌬ cells; mutants developed blue-colored colonies on a medium that was overlaid with a pH 9.8 buffer supplemented with BCIP (Fig. 5A), which records the leakage of intracellular alkaline phosphatase into the extracellular milieu (7,30). We also observed phosphatase leakage when the yeast cells were grown on media with a less stressful pH of 6.5 (Fig. 5B). The addition of 1 M sorbitol elicited an osmoremedial effect that reduced the leakage of phosphatase from inside the cell (Fig. 5, A and B), indicating that in the kcs1⌬ strain, there was a defect in the maintenance of cell wall structure rather than a defect in the plasma membrane (see Refs. 30 and 31).
We sought an independent means of examining cell wall integrity. It has been reported that perturbations to the integrity of the cell wall can decrease the degree of resistance of the cells to the toxic effects of caffeine (see Ref. 31 and references therein). These experiments were performed on YPD media, since the toxicity of caffeine is enhanced, even in wild-type cells, when they are grown on minimal medium (data not shown). In the absence of caffeine, the greater nutritional support of the YPD media provided the kcs1⌬ cells with some protection against their growth defect, compared with cells grown on minimal media (compare Figs. 4A and 5C). Nevertheless, the growth phenotype of kcs1⌬ cells was still evident at 37°C in YPD media, and they were more sensitive to 10 mM caffeine than were wild-type cells (Fig. 5C). Osmotic stabilization of the plasma membrane by 1 M sorbitol largely protected against this sensitivity to caffeine (Fig. 5C); this osmotic remedial response is typical of a defect in cell wall integrity (31). Clearly, Kcs1p has far more wide ranging consequences for cell function than has been appreciated from earlier studies (12,15).

DINS Activity of Kcs1p Is Required for Vacuolar Biogenesis, Protection against Salt Stress, and Cell Wall Synthesis-Since
Kcs1p is probably pleiotropic (see the Introduction), an important objective in the current study was to investigate the specific role of diphosphoinositol polyphosphate synthesis. Therefore, we transformed kcs1⌬ cells with a plasmid encoding wildtype Kcs1p, which restored PP-InsP 5 and (PP) 2 -InsP 4 levels to 40 -70% of those in wild-type cells (Table I). This was sufficient to rescue vacuolar morphology (Fig. 3), cell wall integrity (Fig.  6B), protective adaptations to salt stress (Fig. 6B), and normal growth rates (Fig. 6A).
We next transformed the kcs1⌬ cells with a plasmid encoding a catalytically inactive form of Kcs1p. To achieve this goal, we took note that all three mammalian diphosphoinositol polyphosphate synthases possess a SLL consensus, which is essential for catalytic activity (16). Multiple sequence alignments indicate that the corresponding sequence in Kcs1p is Ser-887, Leu-888, and Leu-889. We transformed kcs1⌬ cells with a plasmid encoding a mutant form of kcs1p, in which these three residues were changed to Ala. In this yeast strain, designated (kcs1⌬ ϩ pkcs1 SLL3 AAA ), PP-InsP 5 and (PP) 2 -InsP 4 were virtually eliminated (Table I). We also confirmed that the kcs1p SLL3 AAA mutant protein was stable when expressed in these cells (Fig. 6C). In this kcs1⌬ ϩ pkcs1 SLL3 AAA strain, the vacuolar space remained fragmented (Fig. 3), the growth rate was impaired (Fig. 6), there was a defect in cell wall integrity, and the cells were sensitive to salt stress (Fig. 6). This is the  first demonstration, in yeast, that the synthesis of an inositol phosphate is necessary for the homeostatic responses to salt stress and the maintenance of cell wall integrity.
The Importance of the Leucine Heptad Repeats in Kcs1p-The Kcs1p protein contains two groups of four heptad repeats of leucine residues; the first comprises Leu-794, Leu-801, Leu-808, and Leu-815, and the second comprises Leu-857, Leu-864, Leu-871, and Leu-878 (15). These were proposed to be leucine zippers (15), but this idea was not subsequently pursued further. We introduced into kcs1⌬ cells a plasmid encoding a mutant kcs1p protein in which two Ala residues were substituted for either Leu-794 and Leu-801 in the first putative zipper or Leu-857 and Leu-864 in the second putative zipper. In both cases, the cells showed a wild-type phenotype (data not shown). Next, we expressed in the kcs1⌬ strain a plasmid encoding a form of kcs1p in which all four Leu residues were changed to Ala. The corresponding protein was stable when expressed in cells (Fig. 6C). This mutation did not affect the synthesis of diphosphoinositol polyphosphates (Table I). These cells (kcs1⌬ ϩ pkcs1 L1L23 AA ) were also resistant to salt stress (Fig. 6B), but cell wall integrity was compromised (Fig. 6B). The vacuolar space was divided into several smaller compartments, typically 3-5 in number (Fig. 3). This was a less severe defect in vacuolar morphology than was observed in the kcs1⌬ cells (Fig. 3). Nevertheless, these data emphasize the multifunctional nature of Kcs1p and validate our goal of identifying the individual molecular contributions made by different domains in the protein.
Separating the Roles of Arg82p in Synthesizing Precursors for Diphosphoinositol Polyphosphates from Regulation of Expression of Genes in the Arginine Metabolic Pathway-Arg82p has been studied for many years for its role in the transcrip- The inset to each panel shows the levels of arg82p present in 20-g extracts of total cell protein, as determined by Western analysis. The apparent molecular mass of these arg82p proteins was approximately 52 kDa, except the arg82p ⌬282-303 deletion mutant, which was 47 kDa.

TABLE II Mutations in Arg82p influence vacuole biogenesis, responses to salt stress, cell wall integrity, OTCase expression, and levels of InsP 6 , PP-InsP 5 , and (PP) 2 -InsP 4
The vacuolar space was analyzed as described in the legend to Fig. 8. Cell wall integrity and the response to salt stress was assessed as described in the legend to Fig. 9. Some of the data for arginine-dependent repression of OTCase were taken from Ref. 26, and the remainder were determined as described under "Experimental Procedures." Inositol phosphate levels (means Ϯ S.E. from three experiments) were determined by HPLC as described in the legend to Fig. 7 tional control over the expression of enzymes in the arginine catabolic and anabolic pathways (21,22). Recent work has studied whether this particular transcriptional control process requires the inositol phosphate kinase activity of Arg82p (3,26). In the current study, we directly assessed this role for Arg82p by measuring the activity of one of the enzymes (OTCase) that is repressed when arginine is present in the media (26). Others have performed gel shift assays that measure the binding to DNA of the Arg82p-stabilized, transcriptional complex that regulates OTCase expression (3). However, these gel shift assays do not ascertain whether the transcriptional unit is active or not (see Ref. 32). In contrast, our OTCase assays provide a direct readout of the activity of this Arg82p-stabilized transcriptional complex. We transformed arg82⌬ yeast with a plasmid expressing a form of arg82p from which a poly(Asp) region (residues 282-303) was deleted, since this domain is necessary for transcriptional control of gene products that regulate the arginine metabolic pathway (21). In this strain (i.e. arg82⌬ ϩ parg82⌬ 282-303), arginine did not repress expression of OTCase (Table II). There were nearly normal levels of inositol phosphates (Fig. 7), including the diphosphoinositol polyphosphates (Table II). Vacuolar morphogenesis, cell wall integrity, and response to salt stress were all normal (Figs. 8 and 9; Table II). These results show that the role of Arg82p in regulating expression of enzymes in the arginine metabolic pathway is separate from the processes that regulate vacuole morphogenesis.
We also created two yeast strains (arg82⌬ ϩ parg82 G135A , arg82⌬ ϩ parg82 W65R ), which displayed slight metabolic defects, namely elevated levels of InsP 2 , InsP 3 , and InsP 4 (compare Figs. 7 and 2), and slightly impaired synthesis of InsP 6 (Table II). Both strains contained nearly normal levels of diphosphoinositol polyphosphates (Table II) but exhibited a vacuolar defect (Fig. 8), albeit more minor than that seen in arg82⌬ or kcs1⌬ cells. The size of the vacuolar space in both arg82⌬ ϩ parg82 G135A cells and arg82⌬ ϩ parg82 W65R cells was approximately equal to that of wild-type cells but typically comprised between two and four separate compartments (Fig.  8). Further evidence that this phenotype is relatively mild comes from data showing the cells had a normal response to salt stress (Fig. 9), and there was no enzyme leakage through the cell wall (Fig. 9). This phenotype was observed irrespective of whether OTCase was properly repressed (as in arg82⌬ ϩ parg82 G135A cells) or not (as in arg82⌬ ϩ parg82 W65R cells). Again, these data show that the role of Arg82p in regulating OTCase expression is separate from the inositol kinase activity that is necessary for synthesizing precursors for diphosphoinositol polyphosphates.
We created a third category of yeast strains (arg82⌬ ϩ parg82 D131A and arg82⌬ ϩ parg82 K133A ) with gross perturbations to their inositol phosphate profiles (Fig. 7), including  (Table II). These two strains showed normal arginine-dependent repression of OTCase (Table II). In contrast, the vacuolar space was highly fragmented (Fig. 8), the cell wall was leaky, and there was poor adaptation to salt stress (Fig. 9). An important conclusion to derive from these data is that disrupting the inositol phosphate kinase activity of Arg82p yields more extensive phenotypic consequences than have previously been recognized.
Overexpression of Kcs1p Partially Suppresses Some Cellular Defects of the arg82⌬ Strain-The arg82⌬ cells do not contain the InsP 6 and Ins(1,3,4,5,6)P 5 substrates for Kcs1p (Fig. 7). Nevertheless, overexpression of KCS1 in the arg82⌬ background rescued vacuolar morphology (Fig. 8) and resistance to salt stress (Fig. 8). Compared with arg82⌬ cells, phosphatase leakage was less pronounced in arg82⌬ ϩ pTet-KCS1 cells (Fig.  9), indicating that cell wall integrity was improved but not fully restored. The growth rate of the arg82⌬ ϩ pTet-KCS1 cells was improved compared with the arg82⌬ strain ( Fig. 9), but was still less than the growth of wild-type cells. The arginine-dependent expression of OTCase and arginase, which is defective in arg82⌬ cells (21), was not rescued by Kcs1p expression (data not shown).

DISCUSSION
The four most important conclusions to arise from this study are as follows. First, we have shown that PP-InsP 5 and (PP) 2 -InsP 4 synthesis by Kcs1p is a specific functional aspect of this protein that is critical for biogenesis of the yeast vacuole. Second, we have shown that diphosphoinositol polyphosphates are also necessary for stability of the cell wall and for adaptive responses to salt stress. Thus, the significance of PP-InsP 5 and (PP) 2 -InsP 4 is far more wide ranging than previously appreciated. Indeed, no inositol phosphates have previously been implicated in regulating these physiological processes. Third, we have demonstrated that the leucine zipper motifs within Kcs1p also contribute to vacuolar morphogenesis and cell wall stability. Finally, we showed that cell wall stability, cell growth, adaptation to salt stress, and vacuolar morphogenesis were all impaired in cells expressing kinase-defective arg82p mutants, due to the absence of precursor material for the synthesis of diphosphoinositol polyphosphates.
To ascertain the specific importance of the DINS activity of Kcs1p, we first created a strain of yeast from which the entire KCS1 gene had been deleted. We then rescued this strain by transformation with wild-type kcs1p, or we transformed the cells with a mutant form of kcs1p in which the DINS activity was specifically eliminated. The latter cells (kcs1⌬ ϩ pkcs1 SLL3 AAA ) showed defective vacuolar biogenesis, resulting in the formation of small, fragmented vacuoles. The importance of creating a yeast strain with such a specific metabolic defect was borne out by additional experiments in which we discovered the significance of the two leucine heptad repeats in Kcs1p. When both repeats were mutated, the vacuolar space was fragmented, and cell wall integrity was compromised. This result provides the first evidence that these leucine residues are functionally significant, perhaps as leucine zippers, which promote functionally significant homo-and heterodimerization of proteins (17). Indeed, yeast two-hybrid analysis has indicated that Kcs1p may associate with Bmh2 (33), a homologue of the mammalian 14-3-3 family (34). The latter modulate interactions between proteins that regulate signal transduction processes and cell cycle control (34). In any case, our work demonstrates that the contributions of Kcs1p to vacuole morphogenesis involve coordinating the activities of more than one functional domain within this protein.
We also created specific mutant strains of yeast in which the inositol phosphate kinase activity of Arg82p was rendered inactive. These cells displayed impaired vacuolar biogenesis, cell wall integrity was compromised, and there was a defective response to salt stress. Our data indicate that these phenotypes, which have not previously been recognized to arise from arg82p mutations, are largely due to inadequate synthesis of diphosphoinositol polyphosphates, as a result of the elimination of the InsP 5 and InsP 6 precursor pools. Furthermore, the specific absence of DINS activity in the kcs1⌬ ϩ pkcs1 SLL3 AAA strain elicited a general growth defect (Fig. 6). The same growth phenotype was seen in the arg82⌬ strain (Figs. 4 and 5 and Ref. 26), since these cells also cannot synthesize diphosphoinositol polyphosphates (Table II). Our validation of the general growth defect of arg82⌬ cells contrasts with an alternate proposal that growth impairment in this strain is nutritionally dependent (3). This is an important point. Recognition that arg82⌬ cells have a general growth defect undermines the hypothesis that the inositol phosphate kinase activity of Arg82p directly empowers a transcriptional complex to regulate expression of enzymes in the arginine metabolic pathway (see Ref. 32 for a detailed explanation). When this transcriptional role for the kinase activity of Arg82p was first proposed (3), PP-InsP 5 and (PP) 2 -InsP 4 were not considered to be relevant, and InsP 6 was stated as being the end point of the inositol phosphate signaling pathway. In contrast, our current work shows that diphosphoinositol polyphosphates are not only required for cell growth and integrity of the cell wall, but these polyphosphates also direct the very survival of the yeast cell in the face of environmental stress.
In yeast, cell wall biosynthesis is controlled by integrating inputs from several regulatory pathways (13), any of which might not function appropriately in kcs1⌬ cells. Nevertheless, the osmotic remedial sensitivity of the cell wall to caffeine, as is the case in our kcs1⌬ cells (Fig. 4), raises the more specific possibility that a protein kinase C/mitogen-activated protein kinase signaling pathway is perturbed (31). In fact, the KCS1 gene was first identified in a screen for suppressing the temperature sensitivity of the pkc1-4 allele (15). As for the impaired response to salt stress in kcs1⌬ cells, one clue to the underlying mechanism comes from an earlier observation that genes encoding the subunits of the vacuolar H ϩ -ATPase are dramatically up-regulated following salt stress (35); the electrochemical gradient generated by this ATPase is crucial to sequestration of sodium into the vacuole. It is therefore possible that, in the arg82⌬ and kcs1⌬ strains, the defective response to salt stress may be molecularly linked to the deformity of the vacuolar compartment.
What might be the link between diphosphoinositol polyphosphates and biogenesis of the yeast vacuole? Vacuoles derive from the fusion of cytoplasm-derived vesicles and from clathrin-coated and non-clathrin-coated vesicles derived from the endocytic apparatus and the trans-Golgi network (36,37). It has been known for several years that diphosphoinositol polyphosphates bind with high affinity to those clathrin-dependent and non-clathrin-dependent adaptor proteins, such as AP-180 and coatomer, that regulate vesicle traffic in both yeasts and mammalian cells (19,38,39). Furthermore, recent studies indicate that a mammalian diphosphoinositol polyphosphate synthase associates with proteins that regulate vesicle exocytosis (14). To rationalize these effects, it has been suggested that PP-InsP 5 and (PP) 2 -InsP 4 might be phosphate donors for protein phosphorylation (see Ref. 16). Alternately, when the diphosphoinositol polyphosphates bind to target proteins, the phosphotransferase reaction (when the diphosphate ligand is either phosphorylated or dephosphorylated) may act as a molecular switch (2,40,41). This idea may be viewed as being analogous to G-proteins that function as a binary switch between two interconvertible GTP-bound active and GDPbound inactive states. Our study focuses attention on the need to characterize the molecular mechanisms by which diphosphoinositol polyphosphate synthases regulate vacuole biogenesis; future research in this area is also likely to be pertinent to understanding the regulation of vesicle trafficking processes in higher eukaryotic cells.