Regulation of Inositol Metabolism Is Fine-tuned by Inositol Pyrophosphates in Saccharomyces cerevisiae*♦

Background: Regulation of inositol metabolism is crucial for cellular functions. Results: Inositol pyrophosphate-deficient cells exhibit defective inositol biosynthesis. Protein levels of the inositol pyrophosphate biosynthetic enzyme Kcs1 are dynamically altered in response to inositol. Conclusion: INO1 transcription and inositol biosynthesis are regulated by modulation of inositol pyrophosphate synthesis. Significance: Inositol pyrophosphates are novel regulators of biosynthesis of inositol and inositol phospholipids. Although inositol pyrophosphates have diverse roles in phosphate signaling and other important cellular processes, little is known about their functions in the biosynthesis of inositol and phospholipids. Here, we show that KCS1, which encodes an inositol pyrophosphate kinase, is a regulator of inositol metabolism. Deletion of KCS1, which blocks synthesis of inositol pyrophosphates on the 5-hydroxyl of the inositol ring, causes inositol auxotrophy and decreased intracellular inositol and phosphatidylinositol. These defects are caused by a profound decrease in transcription of INO1, which encodes myo-inositol-3-phosphate synthase. Expression of genes that function in glycolysis, transcription, and protein processing is not affected in kcs1Δ. Deletion of OPI1, the INO1 transcription repressor, does not fully rescue INO1 expression in kcs1Δ. Both the inositol pyrophosphate kinase and the basic leucine zipper domains of KCS1 are required for INO1 expression. Kcs1 is regulated in response to inositol, as Kcs1 protein levels are increased in response to inositol depletion. The Kcs1-catalyzed production of inositol pyrophosphates from inositol pentakisphosphate but not inositol hexakisphosphate is indispensable for optimal INO1 transcription. We conclude that INO1 transcription is fine-tuned by the synthesis of inositol pyrophosphates, and we propose a model in which modulation of Kcs1 controls INO1 transcription by regulating synthesis of inositol pyrophosphates.

the synthesis of inositol pyrophosphates. This finding suggested that inositol pyrophosphates may function in the regulation of inositol metabolism.
In this study, we report that inositol pyrophosphates carry out a novel function in the regulation of inositol metabolism. To elucidate the mechanism whereby inositol pyrophosphates regulate inositol synthesis, as suggested by the kcs1⌬ phenotype, we determined the effects of disruption of inositol pyrophosphate synthesis on inositol homeostasis. Our findings suggest that inositol pyrophosphates synthesized from IP 5 by Kcs1 are required for the optimal transcription of INO1 but not for activity of the Opi1-Ino2-Ino4 regulatory complex. Moreover, the Kcs1 protein levels are dynamically altered by addition or removal of exogenous inositol, suggesting that rapid turnover of inositol pyrophosphates generated by Kcs1 regulates inositol synthesis. We propose a model in which regulation of Kcs1catalyzed synthesis of 5PP-IP 4 modulates INO1 transcription.

Yeast Strains, Plasmids, and Growth Media
The yeast S. cerevisiae strains used in this study are listed in Table 1. Wild type (WT) strain with the GFP-HIS3MX6 cassette integrated at the carboxyl-terminal end of the KCS1 open reading frame was obtained from the Yeast-GFP Clone Collection (Invitrogen). Single deletion mutants with the GFP tag and double mutants were obtained by tetrad dissection. Synthetic complete (SC) medium contained adenine (20.25 mg/liter), arginine (20 mg/liter), histidine (20 mg/liter), leucine (60 mg/liter), lysine (200 mg/liter), methionine (20 mg/liter), threonine (300 mg/liter), tryptophan (20 mg/liter), uracil (20 mg/liter), yeast nitrogen base without amino acids (Difco), all the essential components of Difco vitamin (inositol-free), 0.2% ammonium sulfate, and glucose (2%). Inositol was supplemented separately where indicated. Synthetic dropout media contained all ingredients mentioned above except for the amino acid used as a selectable marker and were used to culture strains containing a plasmid. Synthetic complete or dropout medium containing 75 M inositol is denoted as Iϩ, whereas medium lacking inositol is denoted IϪ.
The plasmids used in this study are listed in Table 2. The plasmids pFL38, pFV198, pFV217, and pFV241 (22) were gifts from Dr. Evelyne Dubois, and the UAS INO reporter plasmid (12) was a gift from Dr. Christopher Loewen. All the plasmids  were amplified and extracted using standard protocols. The plasmids were transformed into yeast strains using a one-step transformation protocol (33).

Measurement of Intracellular Inositol
Intracellular inositol was measured as described previously (19) with minor modifications. Briefly, cells were harvested at 4°C by centrifugation, washed once with ice-cold water, and resuspended in ice-cold 7.5% perchloric acid. Each sample was lysed by vortexing with acid-washed glass beads for 10 min at 30-s intervals, alternating with a 30-s incubation on ice. Perchloric acid was removed by titration to pH 7.0 with ice-cold 10 M potassium hydroxide. The cell extracts were clarified by centrifugation for 5 min at 2000 ϫ g at 4°C. The supernatants were collected, and intracellular inositol was measured by enzymecoupled fluorescence assay (34). Inositol content (picomoles) was normalized to units of A 550 .

Determination of PI by TLC
Yeast cells were grown to the mid-logarithmic growth phases (A 550 ϭ 1.0) at 30°C. Cells were then washed once with ice-cold water, and total lipids were extracted with chloroform/methanol (2:1) (v/v) as described previously (35). The extracted lipids were applied onto silica gel plates (Partisil K6F 60 Å, Whatman) pretreated with 1.8% boric acid and separated in the onedimension solvent system chloroform/triethylamine/ethanol/ water (30:35:35:7) as described previously (36). Phospholipids were visualized by carbonization at 120°C for 10 min after dipping plates into 3.2% H 2 SO 4 and 0.5% MnCl 2 and subsequent staining with iodine vapor. Stained silica plates were quantified using ImageJ software (National Institutes of Health). Total PI levels in each strain were normalized to total PC levels.

Spotting Assay
Cells were precultured in Iϩ to the mid-logarithmic growth phase at 30°C, counted using a hemocytometer, and washed with sterile water. 3-l aliquots of a series of 10-fold dilutions were spotted onto Iϩ or IϪ plates and incubated for 3 days at the indicated temperatures.

Real Time Quantitative PCR (RT-qPCR) Analysis
Cells were grown to the indicated growth phase and immediately harvested at 4°C. Total RNA was extracted using hot phenol (37) and purified using the RNeasy mini plus kit (Qiagen, Valencia, CA). Complementary DNA (cDNA) was synthesized using the first strand cDNA synthesis kit (Roche Applied Science) according to the manufacturer's manuals. RT-qPCRs were performed in a 20-l volume using Brilliant III Ultra-Faster SYBR Green qPCR master mix (Agilent Technologies, Santa Clara, CA). Triplicates were included for each reaction. The primers for RT-qPCR are listed in Table 3. RNA levels were normalized to ACT1. Relative values of mRNA transcripts are shown as fold change relative to indicated controls. Primer sets were validated according to the Methods and Applications Guide from Agilent Technologies. Optimal primer concentrations were determined, and primer specificity of a single product was monitored by a melt curve following the amplification reaction. All the primers were validated by measurement of PCR efficiency. All the primers used in this study have calculated reaction efficiency between 95 and 105%.

Quantification of INO1 Expression
RT-qPCR Analysis-Cells were pregrown in Iϩ to the midlogarithmic phase and inoculated into fresh Iϩ medium at A 550 of 0.05. When the A 550 reached 0.5, cells were harvested by centrifugation at 3500 rpm for 3 min at 30°C, washed with prewarmed IϪ or Iϩ, and resuspended to fresh IϪ or Iϩ, respectively. Samples were harvested for RT-qPCR analysis at the indicated times by centrifugation at 3500 rpm for 3 min at 4°C. Cells grown in Iϩ to an A 550 of 0.5 were collected at 4°C and used as the 0-h time point.
␤-Galactosidase Reporter Assay-WT and mutant cells that were transformed with the UAS INO reporter plasmid were precultured in Iϩ to the mid-logarithmic growth phase (A 550 of 0.5-0.8), washed with prewarmed IϪ, and resuspended in fresh IϪ. After continuous growth for 4 h, cells were harvested, and ␤-galactosidase was assayed as described previously (12,38).

SDS-PAGE and Western Blot Analysis
Cells grown to the indicated growth phase were harvested at 4°C and subjected to mechanical breakage at 4°C with acidwashed glass beads in lysis buffer containing 50 mM Tris, 125 mM sodium chloride, 1% Nonidet P-40, 2 mM EDTA, and 1ϫ protease inhibitor mixture (Roche Applied Science). Protein extracts were clarified twice by 5 min of centrifugation at 13,000 ϫ g at 4°C to remove cell debris and glass beads. Protein concentration was determined using the BCA TM protein assay (Pierce Protein), with bovine serum albumin as the standard. Extracts containing 50 g of protein were boiled with protein gel sample buffer, separated on 8% SDS-PAGE, and electrotransferred to a polyvinylidene difluoride (PVDF) membrane (Millipore). The membrane was incubated with antibodies (1:3000 anti-GFP; 1:3000 anti-tubulin; 1:5000 appropriate secondary antibodies conjugated with HPR) and visualized using ECL Plus substrate (Pierce Protein), with ␣-tubulin as the loading control. ImageJ software was used to quantify the intensities of bands.

Visualization of Opi1p-GFP Using Fluorescence Microscopy
To visualize the localization of Opi1p-GFP in WT and kcs1⌬ cells, fluorescence microscopy was performed using an Olympus BX41 epi-fluorescence microscope. Images were acquired using an Olympus Q-Color3 digitally charge coupled device camera operated by QCapture2 software. All pictures were taken at ϫ1000.

Deletion of KCS1 Results in Decreased Inositol Biosynthesis-
To identify potential regulators of inositol biosynthesis, we carried out a targeted screen for the growth of mutants hypersensitive to the inositol-depleting drug valproate. Yeast mutants carrying deletions in genes with reported functions in inositol metabolism (Saccharomyces Genome Database) were grown on Iϩ or IϪ plates. We screened 26 deletion mutants in the categories expected to affect inositol metabolism, including inositol polyphosphate kinases, protein kinases and protein phosphatases, vacuolar proteins, and endoplasmic reticulum membrane proteins. Deletion mutants that exhibited defective growth on IϪ were further tested for growth on medium supplemented with valproate. One of the mutants identified in this screen was kcs1⌬. Inositol auxotrophy of kcs1⌬ was also reported in genome-wide studies of inositol auxotrophy (39,40). To further investigate the role of KCS1 in the regulation of inositol metabolism, we analyzed the growth of the kcs1⌬ mutant. As seen in Fig. 2A, kcs1⌬ cells showed an extended lag phase when inocu-lated into Iϩ medium compared with isogenic WT cells. Importantly, they did not significantly grow in IϪ medium. Furthermore, growth of the mutant was diminished relative to that of WT cells at elevated temperatures, even in the presence of inositol (Fig. 2B). Consistent with inositol auxotrophy, intracellular inositol levels in kcs1⌬ were reduced to less than 30% of WT levels (Fig. 2C), and PI were about 42% of WT (Fig. 2D). Inositol biosynthesis is activated in WT cells in inositol-deficient medium by dramatically up-regulating INO1 transcription (2). However, up-regulation of INO1 mRNA was not observed in kcs1⌬ (Fig. 2E), suggesting that transcription of INO1 is defective in the mutant. We addressed the possibility that defective INO1 transcription resulted from a global repression of transcription by comparing mRNA expression of a variety of genes in WT and kcs1⌬, including genes in glycolysis (PDA1 and TDH3), basal transcription (TAF10, TFC1, and SPT15), and protein processing (RDN18 and UBC6). None of these genes exhibited decreased expression in kcs1⌬ (Fig. 2F). Taken together, these studies suggested that decreased INO1 transcription in kcs1⌬ diminishes biosynthesis of inositol and PI, leading to inositol auxotrophy.
Decreased Inositol Biosynthesis in kcs1⌬ Is Not Because of Perturbation of the UAS INO Regulatory Complex Opi1-Ino2-Ino4-The native promoter of INO1 contains the UAS INO element that is widely found in the promoter regions of many genes, including genes involved in phospholipid metabolism (14). As shown in Fig. 3A, transcription of genes containing the  Among genes regulated in this manner, INO1 is the most responsive (2,13). In the absence of inositol, localization of Opi1 on the endoplasmic reticulum is stabilized by interaction with Scs2 and PA (12). In response to exogenous inositol, Opi1 is translocated to the nucleus, inhibiting INO1 transcription (12,44). We addressed the possibility that decreased transcription of INO1 in kcs1⌬ is caused by retention of the transcription repressor Opi1 in the nucleus. As shown in Fig. 3B, in IϪ conditions, GFP-tagged Opi1 locates on the nuclear rim in kcs1⌬ as observed in WT cells, indicating that the translocation of Opi1 is not perturbed in kcs1⌬. Therefore, KCS1 does not regulate INO1 transcription by affecting the localization of Opi1.
We further investigated if the INO1 transcription defect in kcs1⌬ was caused by perturbation of the transcriptional activators Ino2 and Ino4. INO2 is known to be up-regulated in IϪ, whereas INO4 is constitutively expressed in both Iϩ and IϪ (45). Although INO2 transcripts were decreased in kcs1⌬ relative to WT cells, expression in IϪ was greater than in Iϩ in both strains (Fig. 3C), indicating that decreased INO1 expression in kcs1⌬ is not due to an inability to up-regulate INO2. Levels of the constitutively expressed INO4 were not significantly diminished in kcs1⌬ (Fig. 3D). These experiments suggest that decreased transcription of INO1 in kcs1⌬ is most likely not due to decreased availability of Ino2 and Ino4, although levels of INO2 transcription were somewhat decreased relative to WT.
KCS1 Is Required for Optimal INO1 Transcription-As mentioned, OPI1 is a transcriptional repressor of INO1, and deletion of OPI1 leads to overproduction of inositol (11). Not surprisingly, deletion of OPI1 restored growth of kcs1⌬ on IϪ at 30 and 37°C (Fig. 4A). Interestingly, deletion of OPI1 also alleviated the growth defect of kcs1⌬ on Iϩ at 30°C (Fig. 4A), suggesting that the Opi1-controlled repression of other genes may also be deleterious to the growth of kcs1⌬. Deletion of OPI1 in kcs1⌬ restored PI levels (Fig. 2D). Relatively higher PI levels in opi1⌬ than WT were most likely due to overproduction of inositol in opi1⌬. To determine whether INO1 transcription is also restored in kcs1⌬opi1⌬, we analyzed INO1 expression in the double deletion mutant transformed with the INO1-lacZ reporter. Surprisingly, although deletion of OPI1 increased INO1-lacZ expression in kcs1⌬, expression in kcs1⌬opi1⌬ was only 20 -30% of that in WT and opi1⌬ cells (Fig. 4B), suggesting that KCS1 is required for optimal INO1 transcription.
Both bZIP and DINS Domains of Kcs1 Are Required for INO1 Transcription-As depicted in Fig. 5A, Kcs1 has two functional domains (46,47) as follows: the kinase domain (also referred as DINS) (47,48) and two bZIP domains containing four leucine heptad repeats (22,46). Plasmids containing the full-length KCS1 or KCS1 with site mutations in each functional domain were constructed and characterized previously (Fig. 5A) (22). To determine whether these domains are required for INO1 transcription, we assayed growth and INO1 expression in kcs1⌬ cells transformed with these plasmids. In contrast to the full-length KCS1 (pKCS1), the kinase-mutated KCS1 (pKCS1 SLL3 AAA ) did not rescue inositol auxotrophy or restore INO1 transcription in kcs1⌬ (Fig. 5, B and C). It has been dem-onstrated that synthesis of inositol pyrophosphates 5-IP 7 and 5PP-IP 4 is virtually eliminated by mutation of the kinase domain (22). Therefore, Kcs1 kinase activity, which catalyzes the synthesis of inositol pyrophosphates, is required for inositol biosynthesis as well as optimal INO1 transcription. Previous studies also indicated that site mutations in the bZIP domain did not affect the generation of inositol pyrophosphates (22). Unexpectedly, kcs1⌬ cells containing the bZIP-mutated KCS1 exhibited decreased growth on IϪ, which was rescued by inositol (Fig. 5B). Consistent with the defective growth on IϪ, the strain also exhibited a 50% decrease in INO1 expression compared with WT (Fig. 5C)   anti-GFP, most likely corresponding to full-length and truncated Kcs1 proteins, as reported previously (49). WT cells cultured in IϪ (Fig. 6A), conditions in which INO1 transcription is increased, exhibited elevated levels of Kcs1 protein compared with WT cells cultured in Iϩ. In addition, both Kcs1 protein and INO1 transcription levels were decreased at elevated temperature compared with those observed at 30°C (Fig. 6, A and  B). Interestingly, decreased Kcs1 protein levels in Iϩ relative to IϪ were not observed in opi1⌬ cells (Fig. 6A), indicating that OPI1 is required to regulate Kcs1 protein in response to inositol.
To determine whether Kcs1 protein levels respond specifically to inositol, we observed the effects on Kcs1 protein of shifting cells from Iϩ to fresh IϪ medium, which are conditions that increase INO1 expression. WT cells were grown in Iϩ to the mid-logarithmic phase (A 550 of 0.5), then shifted to prewarmed Iϩ or IϪ medium, and harvested for analysis of Kcs1 protein levels and INO1 expression. As shown in Fig. 7A, by 2 h after the shift to IϪ, levels of the full-length Kcs1 protein increased more than 10-fold. Levels decrease after 4 h, and Kcs1 was not detected at 6 h. This pattern is consistent with the pattern of INO1 expression (Fig. 7B), which peaked at 2 h and was significantly diminished at 6 h. Kcs1 protein was not increased significantly in cells shifted to fresh Iϩ medium (Fig.  7A). These findings indicated that Kcs1 protein levels and INO1 transcription levels are regulated similarly in WT cells in response to exogenous inositol. In contrast to WT cells, opi1⌬ cells did not exhibit an increase in Kcs1 protein in response to inositol (Fig. 7A), indicating that Opi1 regulates Kcs1 protein levels. Interestingly, despite the dramatic increase in Kcs1 protein in response to the shift from Iϩ to IϪ, transcription of KCS1 was not altered (Fig. 7C).
In reciprocal experiments, we assayed Kcs1 protein levels in cells shifted from IϪ to Iϩ (Fig. 7D). WT cells were precultured in IϪ to the mid-logarithmic phase (A 550 of 0.5); inositol was then added, and cells were harvested for analysis of Kcs1 protein levels at the indicated times. In control cells (IϪ), Kcs1 protein exhibited a steady decrease after 1 h and was reduced to less than 10% of the initial level within 4 h. In cells supplemented with inositol, the decrease in Kcs1 protein levels was greater than in IϪ controls. The decrease in Kcs1 protein is consistent with the well established rapid decrease in INO1 transcription observed in response to inositol (12,50). Taken together, these experiments indicate that Kcs1 protein, but not the transcription of KCS1, is regulated in response to exogenous inositol, and this modulation of Kcs1 protein requires Opi1.
Inositol Pyrophosphates 5PP-IP 4 Synthesized from IP 5 by Kcs1 Are Required for INO1 Transcription-The findings that Kcs1 protein is required for INO1 expression and that levels of INO1 transcription correspond to levels of Kcs1 protein suggest that Kcs1-catalyzed synthesis of inositol pyrophosphates regulates INO1 expression. We analyzed well characterized inositol pyrophosphate mutants to determine which inositol pyrophosphates are responsible for the regulation of INO1 transcription. The biosynthetic pathways for generating soluble inositol polyphosphates are depicted in Fig. 1. Hydrolysis of phosphatidylinositol 4,5-bisphosphate by Plc1 provides IP 3 as a precursor for the synthesis of inositol polyphosphates. Ipk2 catalyzes the synthesis of IP 4 and IP 5 , and Ipk1 catalyzes the synthesis of IP 6 . Kcs1 catalyzes the pyrophosphorylation of IP 5 to 5PP-IP 4 and further to (PP) 2 -IP 3 (not shown) and IP 6 to 5-IP 7 (48,51,52). Vip1 catalyzes the synthesis of inositol pyrophosphates at the 1-hydroxyl site of the inositol ring (26 -28). To assess which inositol poly-and/or pyrophosphates are involved in the regulation of inositol biosynthesis, we assayed inositol auxotrophy and INO1 expression in all the single and double mutants shown in Table 4. Inositol poly-/pyrophosphates synthesized by the WT and deletion strains shown in Table 4 have been characterized previously by high performance liquid chromatography (HPLC) (22,23,51,53). As seen in Fig. 4, A and B, ipk1⌬ did not exhibit growth defects on IϪ plates, although deletion of KCS1 and/or IPK2 caused inositol auxotrophy consistent with severe defects in INO1-lacZ expression. Deletion of KCS1 in ipk1⌬, which additionally depletes inositol pyrophosphates synthesized from IP 5 , led to inositol auxotrophy. Consistent with this, INO1-lacZ expression was greatly reduced in kcs1⌬ipk1⌬ compared with both WT and ipk1⌬. These find-  ings suggest that Kcs1-generated 5PP-IP 4 is required for optimal inositol biosynthesis. Inositol defects resulting from deletion of VIP1 were less severe than defects observed in kcs1⌬. Intracellular inositol was decreased by 20% in vip1⌬ but 70% in kcs1⌬ compared with WT (Fig. 2C), and INO1-lacZ expression was decreased about 50% in vip1⌬ but almost not detected in kcs1⌬ (Fig. 4B). The severe inositol defects in kcs1⌬, but not in vip1⌬, led to inositol auxotrophy. The double mutant kcs1⌬vip1⌬ has severe inositol defects as an inositol auxotroph. It exhibited a 60 -80% decrease in intracellular inositol (Fig. 2C) and greatly decreased INO1-lacZ expression (Fig. 4B). Therefore, we conclude that kcs1⌬ is epistatic to vip1⌬ with respect to inositol biosynthesis.

DISCUSSION
This is the first demonstration that Kcs1, which catalyzes the synthesis of inositol pyrophosphates, regulates inositol biosynthesis by controlling INO1 expression. We report the following: 1) kcs1⌬ cells exhibit reduced intracellular inositol and PI, decreased INO1 expression, and decreased growth on inositolfree media; 2) disruption of either functional domain of Kcs1 protein causes inositol deficiency; 3) Kcs1 protein, but not transcription, is regulated in response to inositol; and 4) deletion of KCS1, but not IPK1, causes inositol deficiency, suggesting that synthesis of inositol pyrophosphates from IP 5 but not IP 6 is necessary for inositol synthesis. Based on these findings, we propose a model in which Kcs1-catalyzed synthesis of inositol pyrophosphates modulates INO1 transcription.
Inositol pyrophosphate-deficient kcs1⌬ cells exhibited defective inositol metabolism. Deletion of KCS1 led to an extended lag phase and nearly no growth in IϪ ( Fig. 2A). Consistent with this, intracellular inositol in kcs1⌬ cells was decreased to less than 30% of WT (Fig. 2C), whereas PI was decreased to about 42% of WT. In response to inositol depletion, kcs1⌬ cells displayed severely reduced INO1 derepression compared with WT cells (Fig. 2E). We conclude that the inositol defects in kcs1⌬ are caused by defective INO1 transcription.
Disruption of either of the two functional domains DINS/ kinase and bZIP of Kcs1 resulted in defective inositol biosynthesis (Fig. 5). Although site mutations in either domain resulted in defective INO1 expression and inositol auxotrophy,   Our findings indicate that Kcs1 protein, but not transcription, is regulated in response to inositol. A novel mechanism underlying the regulation of KCS1 transcription in response to phosphate signals was identified previously (49). Pho4-mediated transcription of the antisense and intragenic RNAs in KCS1 leads to the production of truncated Kcs1 protein and down-regulation of Kcs1 kinase activity (49). This mechanism of regulation of phosphate signaling involves a positive feedback loop, in which species of the mRNAs and proteins of KCS1 are regulated by transcription of the antisense and intragenic RNAs. In contrast to Pho4-mediated regulation of KCS1, the KCS1 mRNA levels did not change in response to inositol (Fig.  7C), and the full-length and truncated Kcs1 proteins were similarly increased in IϪ (Fig. 7A) and decreased in Iϩ (Fig. 7D). These findings suggest a different mechanism underlying regulation of Kcs1 protein in inositol biosynthesis compared with phosphate signaling. We speculate that Kcs1 protein may be controlled by translation or post-translational modification and/or stability of Kcs1 protein.
Analysis of inositol pyrophosphate mutants indicates that inositol pyrophosphates synthesized from IP 5 but not IP 6 are the most likely regulators of inositol biosynthesis. As summarized in Table 4, ipk1⌬, which lacks IP 6 and IP 7 , did not exhibit inositol defects, whereas kcs1⌬ipk1⌬, which lacks 5PP-IP 4 , IP 6 and IP 7 , exhibited severe inositol defects. These findings suggest that 5PP-IP 4 , synthesized from IP 5 , is required for inositol biosynthesis. However, we cannot completely rule out the possibility that 5-IP 7 is required for inositol biosynthesis. Indeed, deletion of ipk1⌬ caused only about a 30% decrease in INO1 expression (Fig. 4B), consistent with the findings of Wu and co-workers (3). Therefore, 5PP-IP 4 is sufficient for inositol regulation, but IP 7 also contributes to regulation. This is consistent with the moderate inositol defects observed in vip1⌬. Because of the difficulty of constructing a strain that can generate 5PP- IP 4 and IP 6 , but not IP 7 , it is difficult to elucidate the specific role of IP 7 in regulating INO1 transcription. Interestingly, deletion of PLC1, the gene encoding phospholipase C that hydrolyzes phosphatidylinositol 4,5-bisphosphate and generates IP 3 as precursors for inositol poly-/pyrophosphates, exhibited elevated INO1 expression (55,56). It is likely that regulation of INO1 gene expression and inositol biosynthesis is coordinated with phospholipase C activation in addition to the negative feedback circuit in response to exogenous inositol. However, deletion of PLC1 is lethal in some genetic backgrounds (42). This complicates our understanding of the regulation of INO1 expression by PLC1. Interestingly, inositol polyphosphates IP 5 and IP 6 , produced from phosphorylation of IP 3 , have roles in Ino80-mediated chromatin remodeling, a process also required for INO1 expression (3,4). Regulation of INO1 expression by synthesis of inositol pyrophosphates from IP 5 and IP 6 will further complicate the regulation of INO1 expression as altered levels of IP 5 and IP 6 may affect chromatin structure. We propose a model, depicted in Fig. 8, in which optimal INO1 transcription is modulated by the synthesis of inositol pyrophosphate, 5PP-IP 4 (derived from IP 5 ). Under derepressing conditions (IϪ), Opi1 is excluded from the nucleus (2, 12), although Kcs1 protein levels are increased (Fig.  7A). Increased Kcs1 protein accelerates production of 5PP-IP 4 , which is required for optimal INO1 expression. Nuclear Opi1 most likely decreases Kcs1 protein as increased Kcs1 was observed in opi1⌬ and in IϪ (during which Opi1 is excluded from the nucleus). Consistent with this, under repressing conditions (Iϩ), Kcs1 is rapidly decreased (Fig. 7D), most likely due to Opi1 translocation into the nucleus where it represses INO1 expression (2, 12) and decreases Kcs1 protein. Kcs1 and Opi1 may compete for a common binding site via the bZIP domain in the nucleus. Therefore, Opi1-dependent modulation of Kcs1 protein allows one or the other to interact with the common sites of specific nuclear proteins required for INO1 transcription in the nucleus, leading to repression or transcription of INO1, respectively. In this scenario, Kcs1 protein levels control INO1 transcription by regulating the synthesis of inositol pyrophosphates. We speculate that 5PP-IP 4 may be required to recruit transcriptional activators to the INO1 promoter region or stabilize the interaction among those activators.
In conclusion, we identified a novel mechanism whereby inositol biosynthesis is regulated by modulation of Kcs1 protein and suggested a model in which Kcs1-catalyzed synthesis of inositol pyrophosphates regulates INO1 transcription.