Phosphoglucomutase is a key enzyme of glucose metabolism that interconverts glucose-1-phosphate and glucose-6-phosphate. Loss of the major isoform of phosphoglucomutase in Saccharomyces cerevisiae results in a significant increase in the cellular glucose-1-phosphate-to-glucose-6-phosphate ratio when cells are grown in medium containing galactose as carbon source. This imbalance in glucose metabolites was recently shown to also cause a six- to ninefold increase in cellular Ca2+ accumulation. We found that Li+ inhibition of phosphoglucomutase causes a similar elevation of total cellular Ca2+ and an increase in 45Ca2+ uptake in a wild-type yeast strain grown in medium containing galactose, but not glucose, as sole carbon source. Li+ treatment also reduced the transient elevation of cytosolic Ca2+ response that is triggered by exposure to external CaCl2 or by the addition of galactose to yeast cells starved of a carbon source. Finally, we found that the Ca2+ overaccumulation induced by Li+ exposure was significantly reduced in a strain lacking the vacuolar Ca2+-ATPase Pmc1p. These observations suggest that Li+ inhibition of phosphoglucomutase results in an increased glucose-1-phosphate-to-glucose-6-phosphate ratio, which results in an accelerated rate of vacuolar Ca2+ uptake via the Ca2+-ATPase Pmc1p.
- calcium influx
- calcium signal
- glucose phosphate
lithium is commonly used clinically either alone or in combination with valproic acid or carbamazepine for the treatment of bipolar disorder (11, 31). Also, it is the oldest mood stabilizer that remains in clinical use (38). Despite extensive research, the exact molecular mechanism(s) of Li+ action remain(s) unknown (24, 37). Two molecular targets for Li+ at therapeutically relevant concentrations (0.6–1.2 mmol/l serum) are inositol monophosphatase and glycogen synthase kinase (GSK)-3β, both important enzymes of intracellular signal transduction pathways (7, 17). The third molecular target of Li+, phosphoglucomutase (PGM) (18, 37), is a key metabolic enzyme of reserve polysaccharide synthesis and galactose (Gal) metabolism (14, 25).
The effect of Li+ on inositol metabolism was described during the 1970s (4), and a relationship between altered inositol metabolism and cellular Ca2+ signaling was elucidated during the early 1980s (8). Briefly, Li+ inhibits inositol monophosphatase. Therefore, Li+ therapy results in the accumulation of inositol monophosphate and a decrease of free inositol in the brain, which is thought to result in decreased neuronal Ca2+ mobilization (1). Alternative proposals to explain how Li+ modifies cellular signaling have suggested that it decreases protein kinase C activity, G protein activation, or cAMP generation (1, 16, 43). It has also been reported that Li+ inhibits GSK-3β activity (23, 40). Some developmental, metabolic, and neuroprotective actions of Li+ could be readily explained by a negative regulation of GSK-3β function (18, 23, 40). Li+ inhibits the activity of these proteins (and other enzymes not directly involved in signal transduction) by competitively displacing Mg2+ from its binding site (35, 36). Enzymes such as inositol phosphatases, fructose 1,6-bisphosphatase, and bisphosphate nucleotidases all share a structural domain for Mg2+ binding (44). The Mg2+ binding domain of PGM is different, but Li+ also competitively displaces Mg2+ from its binding site on PGM, thus reducing its enzymatic activity (26, 36).
PGM catalyzes the reversible conversion of glucose (Glc)-1-phosphate (Glc-1-P) to Glc-6-phosphate (Glc-6-P). This function leads to its involvement in the synthesis and degradation of UDP-hexoses and glycogen, as well as in Gal metabolism. The direction of metabolic flow through PGM depends on the carbon source available (Fig. 1). Saccharomyces cerevisiae has two PGM genes, PGM1 and PGM2, with the latter accounting for ∼80–90% of the total activity (14). At least one of the PGM genes is required for cell growth in media containing Gal, but not Glc, as sole carbon source. Recently, we reported (15) that a pgm2Δ strain lacking the major isoform of PGM accumulates six- to ninefold more total cellular Ca2+ (Cat) than a wild-type strain. Furthermore, this Ca2+ overaccumulation occurred in Gal-containing media, in which the intracellular Glc-1-P level is severalfold higher than normal, but not in media containing Glc as carbon source. The increase in Cat observed in cells grown with Gal as carbon source is due to an elevated rate of Ca2+ influx across the plasma membrane (15). We found (2) that the altered Ca2+ homeostasis phenotypes (including high Cat and increased Ca2+ uptake through the plasma membrane) in the pgm2Δ strain are suppressed in a mutant S. cerevisiae strain that lacks the regulatory subunit of phosphofructokinase (PFK2). This pgm2Δ/pfkΔ strain contained elevated levels of both Glc-1-P and Glc-6-P when grown in media containing Gal as carbon source. This finding led to the conclusion that cellular Ca2+ homeostasis is linked to the intracellular Glc-1-P-to-Glc-6-P ratio, and thus to PGM activity, in S. cerevisiae. Recently, we reported (3) that the disruption of the PMC1 gene encoding the vacuolar Ca2+-ATPase Pmc1p also suppressed the Ca2+-related phenotypes in the pgm2Δ strain, suggesting an important role for vacuolar Ca2+ compartmentalization. We also noted that the grossly increased unfolded protein response in the pgm2Δ mutant is probably a consequence of unbalanced (predominantly vacuolar) Ca2+ storage, which results in a reduced endoplasmic reticulum (ER) Ca2+ level.
These observations prompted us to investigate whether the inhibition of PGM activity by Li+ could influence cellular Ca2+ homeostasis and signaling in S. cerevisiae. Our results indicate that exposure to Li+ induces excessive Ca2+ uptake and accumulation in a manner that is extremely similar to the phenotype previously observed in the pgm2Δ mutant. Because these perturbations in Ca2+ homeostasis can be suppressed by disruption of the gene encoding the vacuolar Ca2+-ATPase Pmc1p (pmc1Δ), we conclude that they result from hyperactivation of the vacuolar Ca2+-ATPase Pmc1p.
MATERIALS AND METHODS
Cell lines and culture conditions.
S. cerevisiae strains used in the present study include SC252 (Sj21R) (MATa ade1 leu2, 3–112 ura3–52 MEL), YDB200 (MATa ade1 leu2,3–112 ura3–52 pgm1Δ::URA3) and YDB171 (MATa ade1 leu2, 3–112 ura3–52 pgm2Δ::LEU2, MEL). S. cerevisiae strains were grown in standard yeast extract-peptone (YP) medium or synthetic minimal (SM) medium supplemented with either 2% d-Glc or 2% d-Gal as carbon source. In all experiments, cultures were grown for at least 5–6 generations to a cell density of ≤1.0 absorbance at 600 nm (A600) unit/ml.
Measurement of Cat levels, Ca2+ uptake rates, and Ca2+ exchange.
Cat levels were measured with an Eppendorf Efox 5053 flame photometer (27, 28). Briefly, cells were grown to a cell density of 0.8–1.0 A600 units/ml and then harvested by centrifugation at room temperature (RT) for 5 min at 10,000 g. A single sample contained ∼100 A600 units of cells. Measurements were routinely carried out in triplicate. The samples were transferred into microcentrifuge tubes of known weight and centrifuged at RT for 10 min at 15,000 g. The supernatants were carefully removed, and the sample was measured gravimetrically on an analytical balance. Each sample was dried in a Speed Vac (Savant) vacuum refrigerator for 3 h, and the dry weight of the samples was measured. One molar HCl (0.6 ml) was added to the dry samples and vortexed. The samples were extracted on a rocker table for 24 h and then centrifuged at 15,000 g for 5 min. Ca2+ measurements were carried out on the supernatants.
45Ca2+ uptake was measured as described previously (28). Cells were harvested from a culture in exponential growth (cell density <0.8 A600 unit/ml), washed three times in distilled water, and resuspended at a density of 1 A600 unit/ml in 40 mM MES-TRIS buffer, pH 5.5, and 20 mM d-Gal or d-Glc. The Ca2+ uptake experiment was started with the addition of 1 μCi/ml 45Ca2+. At the indicated times, 1-ml aliquots were filtered through 0.45-μm Millipore filters prewashed with a solution containing 10 mM LaCl3 and 20 mM MgCl2. Samples were then washed with 5 ml of wash solution before the membranes were collected for scintillation counting. Nonspecific 45Ca2+ binding at the zero time point was subtracted for each sample.
After the isolation of vacuole-rich membrane fraction, the 45Ca2+ uptake was measured similarly as described earlier (3). The test medium was slightly different (in mM): 20 MES-TRIS buffer, pH 6.7, 1 ATP, 2 MgCl2, and 2 NaN3, with 0.5 μCi/ml 45Ca2+. The incubation was carried out at 30°C for 10 min, and the samples were filtered rapidly through 14,000 molecular weight cutoff membranes with a Millipore filter manifold. The filters were washed twice with ice-cold 20 mM MES-TRIS buffer, pH 6.7, 145 mM KCl, and the membrane-associated counts per minute values were measured.
Ca2+ is present in yeast in two kinetically distinguishable pools (9). The 45Ca2+ exchange was measured as described previously (13). Briefly, wild-type yeast was grown in the absence or presence of 15 mM LiCl in the YP-Gal growth medium for five generations. The medium was supplemented with 5 μCi/ml 45Ca2+. The cells were harvested, washed, and resuspended in YP-Gal supplemented with 20 mM CaCl2. The cultures were then incubated at 30°C for 40 min, and aliquots of cells were filtered and processed for scintillation counting as described above.
Measurements of cytosolic free Ca2+ concentrations.
Cytosolic free Ca2+ was measured in yeast cells essentially as described previously (5). Briefly, the wild-type strain carrying the pEVP11 plasmid (which carries the apoaequorin gene and the LEU2 gene as selectable marker) was grown in SM-Glc or SM-Gal medium supplemented with the appropriate amino acids. The cells were then centrifuged and loaded with coelenterazine. For in vivo cytosolic Ca2+ measurements, a Berthold 9050 Lumat luminometer was used. Cultures were grown to 0.7–1.0 A600 unit/ml. For Gal readdition experiments the cells were preincubated in hexose-free test medium for 2 h. Two A600 units of cells were used for a single measurement. After the preincubation period, cells were transferred into sample holders, and the measurement was initiated. After the baseline light emission was measured for ∼40 s, 100 mM CaCl2 or 100 mM Gal was injected directly into the sample cuvette. Each experiment was repeated at least three times before results were accepted. Standardization and calculation of results were performed as described previously (28).
Measurement of cellular Glc-1-P and Glc-6-P levels.
Wild-type S. cerevisiae cells were grown and maintained as described above. For assays of Glc-1-P and Glc-6-P levels, the cells were harvested at a density of 0.8–1 A600 unit/ml by centrifugation and resuspended in YP-Gal medium (100 A600 units/ml). The cells were preincubated at 30°C for 15 min, and the experiment was initiated by the addition of LiCl. The samples were continuously kept in motion (200 rpm/min) in an environmental shaker incubator. Samples were harvested by pipetting 1 ml of cells into microcentrifuge tubes containing 0.11 ml of 6.67 M perchloric acid. The microcentrifuge tubes also contained ∼200 mg of glass beads. After sample addition, the microcentrifuge tubes were placed on ice for 20 min. Subsequently, samples were vortexed vigorously three times for 1 min and centrifuged at 10,000 g for 5 min, and the supernatant was removed. After the neutralization of the supernatant with 5 M KOH (final pH 6.0–6.5), the Glc-6-P and Glc-1-P substrate measurements were carried out according to the method of Bergmayer et al. (6). Because some samples contained Li+ that can competitively displace Mg2+ from PGM, we raised the Mg2+ concentration to 5 mM in the test buffer. We found that the highest sample Li+ concentrations did not significantly affect the results.
Measurement of ATP, ADP, and AMP levels.
ATP, ADP, and AMP levels were measured from the same extracts used for Glc-1-P and Glc-6-P determinations. We used a reversed-phase HPLC method slightly modified from that described by Stocchi et al. (41). Briefly, 20 μl of supernatant was mixed with 80 μl of buffer A (0.1 M KH2PO4, pH 6.0) and filtered through a 0.2-μm syringe filter. Samples were loaded on a Beckman Ultrasphere ODS C18 reversed-phase column (pore size 5 μm, length 250 mm, diameter 4.6 mm). Elution was performed under the following conditions: 9 min at 100% buffer A, 6 min at up to 100% buffer B [0.1 M KH2PO4 containing 10% (vol/vol) HPLC-grade methanol, pH 6.0], and hold for 5 min. The gradient was then returned to buffer A in 2 min, and the initial conditions were restored for 5 min. The flow rate was 1.3 ml/min, and detection was performed at 254 nm. Peaks identified were confirmed by coelution with standards. Data were calculated from the peak areas and expressed as “energy charge,” using the following formula: energy charge = (ATP + ADP)/(ATP + ADP + AMP).
Lithium elevates Cat in yeast cells grown in media containing Gal as carbon source.
We previously demonstrated (15) that a S. cerevisiae mutant lacking the major PGM isoform (pgm2Δ) accumulates six to ninefold more Cat than a wild-type strain when grown in medium containing Gal, but not Glc, as sole carbon source. Others demonstrated that the presence of Li+ inhibits PGM activity in wild-type S. cerevisiae (26). To determine the effect of Li+ on cellular Ca2+ homeostasis, a wild-type yeast strain and a mutant strain lacking the minor PGM isoform (pgm1Δ) were grown in YP media containing Gal or Glc as carbon source. Cultures were also supplemented with 0–15 mM LiCl. Figure 2 shows that the presence of Li+ in the growth medium caused an elevated Cat level in wild-type yeast when Gal was utilized as carbon source (Fig. 2A). The maximum level of Cat was 4.1-fold higher than that in control (untreated) cells. Also similarly to the pgm2Δ mutant, Li+ slowed the growth rate of the cells in a concentration-dependent manner. In contrast, the presence of Li+ in the growth medium did not lead to a significant increase in Cat in cells grown with Glc as carbon source (Fig. 2B). To exclude the possibility that Ca2+ influx may occur as a general response to salt stress, we repeated the above experiments with 150 mM NaCl in the medium. No increase in Cat was observed under these conditions (Fig. 2).
A strain lacking the minor PGM isoform (pgm1Δ) was previously shown to exhibit a modest (10–15%) decrease in PGM activity that did not cause any growth defects in Gal-containing media (14, 15). Similarly, we found that Cat levels were comparable when the wild-type and pgm1Δ strains were grown under these conditions. However, the pgm1Δ strain accumulated significantly more Ca2+ when this growth medium was supplemented with Li+ (Fig. 2A). The maximum level of Cat was 6.3-fold higher than the values measured when this strain was grown in the absence of Li+. As before, the addition of Li+ had no effect on Cat when the pgm1Δ strain was grown with Glc as carbon source (Fig. 2B). We also attempted to determine the effect of Li+ treatment on the pgm2Δ mutant strain. We found that even the lowest (1 mM) Li+ level causes near-complete growth inhibition in Gal-containing media. This observation is not surprising because only ∼10% of the wild-type PGM activity remains in the pgm2Δ mutant, and it suggests that Li+ inhibition of this remaining PGM activity cannot be tolerated. Together, these results suggest that the metabolic bottleneck in the conversion of Glc-1-P to Glc-6-P caused by Li+ inhibition of PGM activity results in a large increase in Cat. Not surprisingly, in strains that hold a reduced level of PGM activity (such as the pgm1Δ mutant) this Ca2+ accumulation is further enhanced.
Li+ increases the rate of Ca2+ uptake in yeast cells but does not alter its intracellular distribution.
Our observation that Li+ elevated Cat in cells grown with Gal as carbon source (Fig. 2A) suggested that increased Ca2+ uptake through the plasma membrane occurs in Li+-treated cells grown under these conditions. To test this hypothesis, wild-type S. cerevisiae cells were grown in media containing Gal as carbon source in the absence or presence of 1 mM LiCl. Cells were then harvested, washed, and resuspended in a Ca2+-uptake buffer with or without 1 mM LiCl. We found that the uptake of 45Ca2+ was accelerated roughly twofold when 1 mM LiCl was present in both the growth medium and the uptake buffer (Fig. 3A). A similar increase in 45Ca2+ uptake was observed when cells were grown in Li+-free medium as long as the Ca2+-uptake buffer contained 1 mM LiCl. This demonstrates that only a 1-min exposure to Li+ was sufficient to mediate this increase in Ca2+ uptake. In contrast, cells grown with Glc as the carbon source did not show any significant Li+-dependent alterations in 45Ca2+ uptake (data not shown). These results indicate that Li+ exposure leads to an increase in Ca2+ uptake by a rapid mechanism that does not require de novo protein synthesis.
Intracellular Ca2+ in yeast exists in two kinetically distinguishable pools (9). More than 95% of the Cat is compartmentalized in the vacuole in a relatively stable polyphosphate-bound form (the “nonexchangeable” or slowly exchangeable pool). The exchangeable pool represents the cytosol-, ER-, and Golgi-localized form of this divalent cation, as well as a small fraction of the huge vacuolar Ca2+ pool. We showed previously (15) that the distribution of Ca2+ between these two kinetically distinguishable pools was not significantly altered in a pgm2Δ mutant. Here, we found no significant difference between the relative size of the exchangeable Ca2+ pools in wild-type yeast grown in the presence or absence of 15 mM LiCl (Fig. 3B). These results indicate that the Li+-treated cells grown with Gal as carbon source continued to distribute cellular Ca2+ normally between these kinetically distinguishable pools.
Magnesium reduces Li+-induced accumulation of Ca2+ in S. cerevisiae.
Li+ inhibits PGM activity by displacing essential Mg2+ (35, 36). In addition, Masuda et al. (26) found that Mg2+ addition can reverse the Li+-mediated inhibition of PGM activity in cell homogenates and protect against the toxic effects of Li+ exposure on yeast cell growth. In light of these findings, we next investigated the effect of Mg2+ on Li+-induced Ca2+ accumulation. Figure 4A shows that the addition of 100 mM MgCl2 to the growth medium had no significant effect on the Cat in wild-type yeast cells grown in medium containing Gal as carbon source. The addition of 15 mM LiCl to the medium resulted in a fourfold increase in Cat, whereas the addition of Mg2+ largely reversed the Li+-induced Ca2+ accumulation. These results demonstrate that although the addition of Mg2+ does not significantly affect Cat in control cells, it alleviates the high-Cat phenotype found in Li+-treated cells grown with Gal as carbon source.
Furthermore, a 4.7-fold reduction in the cellular Li+ level was observed in LiCl + MgCl2-treated cells compared with cells treated with LiCl alone (Fig. 4A). We also found that the addition of 100 mM MgCl2 caused a 2.7-fold increase in the cellular Mg2+ level (Fig. 4B), whereas growth in the presence of 15 mM LiCl reduced the cellular Mg2+ concentration 1.8-fold below the level observed in the wild-type control. The cellular Mg2+ level remained at 2.3-fold higher than control when both 100 mM MgCl2 and 15 mM LiCl were present in the medium instead of 100 mM MgCl2 alone. The slightly decreased Mg2+ level observed when both Li+ and Mg2+ were added together is likely due to competition for entry into the cell. Accordingly, we found that both the elevation of intracellular Mg2+ (4.2-fold) and the reduction of intracellular Li+ (4.7-fold) contribute to the high intracellular Mg2+-to-Li+ ratio (17.8) found in LiCl + MgCl2-treated cells over LiCl-treated cells (Mg2+-to-Li+ ratio = 0.97).
Out of these data two conclusions may be drawn. First, as predicted, extracellular Mg2+ prevents the Li+-induced elevation of Cat. Second, unlike in cell lysates, the addition of Mg2+ to the medium not only elevates the intracellular Mg2+ concentration but also reduces Li+. Consequently, two accommodating effects result in a very high intracellular Mg2+-to-Li+ ratio and decrease PGM inhibition within the cell.
Li+ reverses cellular Glc-1-P-to-Glc-6-P ratio without altering cellular energy charge.
It has been shown that either the deletion of the PGM2 gene or Li+ inhibition of PGM activity elevates the cellular Glc-1-P level in media containing Gal as carbon source (2, 15, 26). To further address whether a direct link exists between Glc-1-P accumulation and increased Ca2+ uptake, we next asked whether the metabolic consequences of Li+ inhibition of PGM developed within a comparable time period.
Wild-type and pgm2Δ yeast strains were grown at 30°C in medium containing Gal as carbon source. An aliquot of the wild-type cells was harvested, and the remainder was immediately resuspended in fresh growth medium containing 15 mM LiCl. Additional aliquots were harvested from the Li+-treated culture 30 s and 5 min later, and each aliquot was then processed to determine cellular Glc phosphate levels. Figure 5 shows the Glc-6-P and Glc-1-P levels and Glc-1-P-to-Glc-6-P ratios measured in control cells (harvested before the addition of Li+) and in cells harvested after 30 s or 5 min of Li+ treatment, respectively. The Glc-1-P-to-Glc-6-P ratio more than doubled in these cells within 30 s of addition of 15 mM LiCl (from 0.33 to 0.79). By the 5-min time point, the Glc-1-P-to-Glc-6-P ratio had increased more than fivefold (from 0.33 to 1.71). Longer incubation times resulted in additional modest increases in the Glc-1-P-to-Glc-6-P ratio in wild-type cells (data not shown). Although this reversal of the Glc-1-P-to-Glc-6-P ratio is due primarily to a large increase in the Glc-1-P concentration, a small decrease in the level of Glc-6-P also occurred. These results demonstrate that the Li+ inhibition of PGM activity leads to a significant increase in the cellular Glc-1-P-to-Glc-6-P ratio within 30 s of exposure, with a further increase occurring on longer incubation. In addition, the rapid changes in the Glc-1-P-to-Glc-6-P ratio occur within a short time frame similar to that of the increase in 45Ca2+ uptake (see Fig. 3), suggesting that these two responses are related. We also note that the Glc-1-P-to-Glc-6-P ratio is significantly lower in Li+-treated wild-type cells compared with the pgm2Δ mutant. This indicates that a more severe metabolic bottleneck occurs when the strain is completely unable to express the major isoform of this enzyme.
A variety of physiological conditions that alter steady-state cellular ATP levels or, more importantly, the cellular energy charge [(ATP + ADP)/(ATP + ADP + AMP)] could result in significant alterations in the relative balance of various intracellular ions. Consequently, energy-depleted cells tend to contain more Na+ and Ca2+ and less K+ and Mg2+. To determine whether this could be responsible for the observed changes in intracellular Ca2+ levels, we next asked whether Li+ treatment causes an alteration in the cellular energy charge. Wild-type yeast cells were again grown with either Glc or Gal as carbon source, and 15 mM LiCl was added to the cultures two generations (3 h) before harvesting. The cells were then processed for the determination of ATP, ADP, and AMP levels. We found that the energy charge of Li+-treated cells was not significantly different from those incubated without Li+ when grown with either Glc (0.618 in untreated vs. 0.609 in Li+ treated), or Gal (0.682 in untreated vs. 0.678 in Li+ treated) as carbon source. These results indicate that the changes in the cellular Ca2+ level following Li+ treatment were not due to a perturbation of the energy charge.
Li+ reduces transient elevation of cytosolic Ca2+ response in S. cerevisiae.
Yeast cells respond to a variety of stimuli, such as exposure to mating pheromones or changes in carbon source, osmolarity, and ionic composition of the environment, with a transient elevation of the cytosolic Ca2+ (TECC) response (5, 19, 21, 32, 42). As described above, we observed an accelerated rate of 45Ca2+ uptake when yeast cells were exposed to Li+ (Fig. 3). We also previously reported (15) that a similar increase in Ca2+ uptake was associated with a pgm2Δ mutant and showed that the TECC response induced by Gal readdition to yeast cells starved for a carbon source is reduced in the pgm2Δ strain.
To determine whether exposure to Li+ also alters the TECC response, we preincubated cells in a medium lacking a carbon source in the presence or absence of 10 mM LiCl and then measured free cytosolic Ca2+ levels in response to the addition of 100 mM Gal. The basal cytosolic Ca2+ level under both conditions before Gal addition was roughly 65 nM (although the basal level in Li+-treated cells was slightly lower than that in untreated cells). We found that the normal TECC response stimulated by the addition of Gal resulted in a peak cytosolic Ca2+ concentration of 200 nM (Fig. 6A). In contrast, the TECC response induced in Li+-treated cells in the same manner resulted in a peak cytosolic Ca2+ concentration of only 130 nM. The recovery phase following the TECC response was similar in both Li+-treated and control cells.
Normal yeast SM medium contains 1 mM CaCl2. However, wild-type yeast strains can grow normally in the presence of 100 mM CaCl2. Previous studies have shown that the abrupt exposure of yeast cells to this high level of CaCl2 results in a TECC response that differs from that induced by hexose addition to cells starved of a carbon source (42). Under these conditions, the sudden elevation of extracellular Ca2+ results in an immediate Ca2+ surge into the cytosol that results in a sharp TECC response followed by a rapid reduction in the level of free cytosolic Ca2+ as plasma membrane Ca2+ channels close and organellar Ca2+ transporters are activated to increase sequestration into intracellular compartments. Figure 6B shows how S. cerevisiae cells grown with Gal as carbon source respond to the injection of 100 mM CaCl2 into the incubation medium. As before, the basal cytosolic Ca2+ level under both conditions was roughly 65 nM. We found that the TECC response following the addition of 100 mM CaCl2 to untreated cells resulted in a peak cytosolic concentration of ∼1,010 nM. In contrast, the response observed in Li+-treated cells was reduced to a peak level of only 310 nM. Thus the TECC response was again significantly reduced in Li+-treated cells. These data indicate that Li+ exposure results in a reduction in the magnitude of the TECC response to various stimuli. Because we had previously shown that Li+ treatment increases the basal rate of cellular Ca2+ uptake, this decrease in the magnitude of the TECC response suggests that Li+ activates organellar Ca2+ sequestration. Therefore, despite the increased Ca2+ uptake, the TECC response observed is smaller than normal.
Increased Cat level in Li+-treated cells is caused by excessive vacuolar Ca2+ uptake.
Intracellular Ca2+ storage occurs primarily in two organelles in S. cerevisiae. Previous studies showed that >90% of the Cat is located in the vacuole, where the Ca2+-ATPase Pmc1p and the Ca2+/H+ antiporter Vcx1p facilitate Ca2+ uptake and accumulation (10, 20, 34). In addition, the Golgi-located Ca2+-ATPase Pmr1p assists in intracellular Ca2+ compartmentalization under certain conditions (12, 33, 39). Previous studies showed that a vcx1Δ/pmc1Δ mutant that is defective for vacuolar Ca2+ uptake exhibits a reduced Cat level and decreased Ca2+ tolerance (28, 29), whereas a pmr1Δ mutant that is impaired in Golgi Ca2+ uptake has an increased rate of cellular Ca2+ uptake (20).
To determine whether an elevated rate of vacuolar Ca2+ sequestration is responsible for the observed effects of Li+ on cellular Ca2+ homeostasis, wild-type, vcx1Δ, pmc1Δ, and vcx1Δ/pmc1Δ strains of S. cerevisiae were grown in YP-Gal medium in the presence or absence of 15 mM LiCl. Both the wild-type and vcx1Δ strains exhibited a similar fourfold increase in Cat when 15 mM Li+ was present in the medium (Fig. 7A). The pmc1Δ mutant showed a smaller increase (1.8-fold) under these conditions, whereas the vcx1Δ/pmc1Δ strain grown in the presence of LiCl exhibited a level of Cat that was only 1.2-fold higher than the untreated control. We also found that a pmr1Δ strain exhibited an increase in Cat in response to Li+ treatment that was similar to the wild-type strain (data not shown). These data suggest that the primary consequence of Li+ exposure is an increased level of vacuolar Ca2+ accumulation and the increased rate of Ca2+ uptake across the plasma membrane is a secondary consequence of excessive vacuolar uptake. Of the two known vacuolar Ca2+ transport mechanisms, Ca2+ transport by Pmc1p appears to be more important for this response than the Ca2+/H+ exchanger Vcx1p. However, Vcx1p appears to be capable of contributing to a small increase in vacuolar Ca2+ transport in the pmc1Δ mutant, because the vcx1Δ/pmc1Δ double mutant showed a lower Cat level than the pmc1Δ strain after Li+ exposure. This conclusion is consistent with the results of previous studies that described partially overlapping roles for these Ca2+ transporters (28, 29).
Isolated and partially purified vacuolar vesicles were also used to examine the effect of Li+ on 45Ca2+ uptake into this organelle (Fig. 7B). Wild-type and pmc1Δ strains were grown in YP-Gal media in the absence or presence of 1 mM LiCl, and a vacuole-enriched membrane fraction was prepared as previously described (3, 39). 45Ca2+ uptake measurements were carried out in test buffer in the absence or presence of 1 mM LiCl, respectively. We observed a 1.9-fold increase in 45Ca2+ uptake in vesicles isolated from wild-type cells cultured in the presence of Li+, but a much smaller 1.2-fold increase was induced in the pmc1Δ mutant by the same treatment. These results further support the role of PMC1 in Li+-induced Ca2+ accumulation.
The low level of PGM activity caused by the deletion of the PGM2 gene results in a slow-growth phenotype when a pgm2Δ strain is grown in medium containing Gal as carbon source (15). This slow-growth phenotype is related to the elevated Ca2+ accumulation that occurs in this strain, and this phenotype is partially suppressed in a pgm2Δ/pfk2Δ double mutant, in which a more normal Glc-1-P-to-Glc-6-P ratio is restored and Cat is reduced (2). We next asked whether the pfk2Δ mutation can also reverse the elevated Cat caused by Li+ treatment. We found that the pfk2Δ mutation effectively suppressed the high-Cat phenotype of the Li+-treated cells (Fig. 8). A significant increase in Cat was not observed in the pfk2Δ strain in the 0–5 mM Li+ range, and at 15 mM Li+ a Cat increase of only 1.7-fold was observed. The level of Ca2+ accumulation in the pfk2Δ strain grown in the presence of 15 mM Li+ was <50% of the level observed in the wild-type strain grown under the same conditions.
It has also been shown that Li+ reduces the growth rate of the wild-type strain to a level that is similar to the growth rate observed in a pgm2Δ strain grown in medium containing Gal as carbon source (26). Because the results described above indicated that the pmc1Δ mutation suppresses the high-Cat phenotype, this suggested that the pmc1Δ mutant might also grow better in media containing LiCl than the wild-type strain. This prediction was found to be correct, as the pmc1Δ strain was found to grow faster on YP-Gal plates supplemented with 15 mM LiCl than either the wild-type or vcx1Δ strain (Fig. 9). Consistent with the observation that the vcx1Δ/pmc1Δ strain largely reverses the high-Ca2+ phenotype (Fig. 7A), we found that the vcx1Δ/pmc1Δ strain grew significantly faster than a strain carrying the pmc1Δ mutation alone. Finally, we also observed an intermediate level of suppression of the Li+-mediated growth inhibition in a pfk2Δ strain (data not shown). Together these results suggest that Li+ exerts its effects primarily by enhancing Ca2+ transport into the vacuole. Because the targeted deletion of the PMC1 gene partially suppresses not only the high-Cat but also the slow-growth phenotype of Li+-treated cells, it appears that the vacuolar Ca2+-ATPase Pmc1p plays a pivotal role in this process.
Li+ has been shown to compete with Mg2+ for a functionally important binding site in mammalian PGM because of similarities in their hydrated radii (35, 36). The resulting Li+-PGM complex is almost completely devoid of enzymatic activity. Recently, Masuda et al. (26) reported that Li+ mediates an inhibition of PGM activity in the yeast S. cerevisiae. We previously showed (15) that reduced PGM activity results in elevated Cat in a pgm2Δ mutant grown in media containing Gal as the carbon source.
In the present study we found that Li+ elevates Cat 4.1-fold in wild-type S. cerevisiae grown in media containing Gal as carbon source and 6.3-fold in a mutant strain (pgm1Δ) that has a modest (10–20%) decrease in PGM protein (Fig. 2A). Consequently, Cat appears to relate inversely to PGM activity in cells grown with Gal as the carbon source. In contrast, the Cat level remained unaltered when these cells were grown in the presence of identical Li+ concentrations in media containing Glc as carbon source (Fig. 2B), where the PGM enzyme levels are much lower because cells do not depend on the activity of this enzyme for the entry of metabolites into the glycolytic pathway (Fig. 1; Ref. 14). The carbon source dependence of this effect is consistent with the previous conclusion that alterations in Ca2+ homeostasis caused by changes in PGM activity are determined by the metabolic flow through the reversible PGM enzymatic reaction (and the resulting relative levels of Glc-1-P and Glc-6-P) (2, 15).
We also demonstrated that the uptake of Ca2+ into Li+-treated cells is significantly accelerated over the Ca2+ uptake measured in control cells (Fig. 3A). Remarkably, this increase was found to begin within 1 min of Li+ exposure. Because the cellular Glc-1-P-to-Glc-6-P ratio more than doubles within 30 s of Li+ addition (Fig. 5), we propose that the rapid increase of Glc-1-P, and the resulting high Glc-1-P-to-Glc-6-P ratio, is the most likely inducer of this response. These observations also suggest that the effect on Ca2+ homeostasis is not due to the induction of de novo gene expression and is probably due to a signaling mechanism.
We previously observed (15) a similar correlation between the cellular Glc-1-P-to-Glc-6-P ratio and Cat in a S. cerevisiae pgm2Δ mutant that lacks the major isoform of PGM. It was also shown that the disruption of a second gene (pfkΔ) encoding the β-subunit of phosphofructokinase in a pgm2Δ strain results in dual metabolic blocks that elevated the levels of both Glc-1-P and Glc-6-P (2). Remarkably, the Ca2+-related phenotypes of the pgm2Δ strain were largely eliminated in the pgm2Δ/pfk2Δ double mutant. Therefore, one might assume that the effect of Li+ on Cat should also be abrogated in a pfk2Δ mutant, which would similarly correct the cellular Glc-1-P-to-Glc-6-P ratio. Indeed, we found that the pfk2Δ strain had significantly lower Cat levels than the wild-type strain when grown in Gal media supplemented with LiCl (Fig. 8).
The hypothesis that Li+ mimics the Ca2+ homeostasis-related phenotype of the pgm2Δ strain through its ability to inhibit PGM activity is further strengthened by the observation that Mg2+ reverts the Li+-induced elevation of Cat (Fig. 4). Unlike cell homogenates in which competition for the functionally important Mg2+ binding site of PGM would correspond to the actual Mg2+ and Li+ levels added, the addition of extracellular Mg2+ both elevates the intracellular Mg2+ levels and decreases the intracellular Li+ concentration. Importantly, extracellular Mg2+ alone had little if any effect on Cat, indicating that it affects Cat in Li+-treated cells by decreasing the inhibition of PGM activity.
A transient elevation of cytosolic Ca2+ occurs in S. cerevisiae immediately after the addition of mating pheromone, Ca2+, or hexoses or an abrupt change in the osmolarity of the environment (5, 32, 42). In the present work, we found that the basal cytosolic Ca2+ levels were slightly lower in Li+-treated cells relative to untreated control cells. The TECC responses caused by CaCl2 shock (sudden Ca2+ overflow) or the addition of Gal to cells starved of a carbon source (Glc phosphate-induced Ca2+ uptake through the plasma membrane) were significantly reduced when Li+ was present in the culture medium (Fig. 6). These results indicate that Li+ alters a common component involved in the propagation of various TECC responses. Because Ca2+ uptake across the plasma membrane increases almost immediately on Li+ exposure, a reduced rate of Ca2+ uptake cannot account for the diminished TECC response. Instead, we propose that the observed reduction in the TECC response results from an increased rate of Ca2+ removal from the cytosol into an intracellular compartment. Consistent with this hypothesis is the finding that the excessive Ca2+ accumulation and slow-growth phenotype induced by Li+ treatment could be reduced by disruption of the gene encoding the vacuolar Ca2+-ATPase Pmc1p (Figs. 7 and 9). These results are also consistent with our recent observation (3) that the disruption of the PMC1 gene partially suppresses the Ca2+-related phenotype of the pgm2Δ mutant. The results of the current work also suggest that deletion of the genes encoding both the vacuolar Ca2+/H+ antiporter Vcx1p and the Ca2+-ATPase Pmc1p causes a more complete suppression of the Li+-mediated Ca2+ phenotype than disruption of the PMC1 gene alone. This observation is consistent with an earlier report that the two structurally and functionally distinct vacuolar Ca2+ transporters play partially overlapping roles. For example, it was shown previously that the Ca2+-sensitive phenotype of the pmc1Δ mutant is further exacerbated in the vcx1Δ/pmc1Δ double mutant (15).
These results, and our earlier observation that the Glc-1-P-to-Glc-6-P ratio is important for proper Ca2+ homeostasis, lead us to propose the following sequence of events. First, Li+ inhibits PGM activity and alters the cellular Glc-1-P-to-Glc-6-P ratio. Second, the altered Glc phosphate levels stimulate vacuolar Ca2+ uptake by the Ca2+-ATPase Pmc1p by a currently unknown mechanism. Third, the increased rate of vacuolar Ca2+ uptake reduces basal cytosolic (and subsequently ER) Ca2+ levels, leading to an induction of the unfolded protein response within the ER lumen (3). Finally, plasma membrane Ca2+ channels open to increase Ca2+ uptake in an attempt to restore the normal resting steady-state cytosolic Ca2+ level.
As yet, we do not understand how Glc phosphates regulate vacuolar Ca2+ uptake through the action of PMC1. In a previous study, it was shown that Glc phosphates were unable to activate 45Ca2+ uptake directly into vesicles derived from vacuolar membranes (3). It has also been shown that the inhibition of calcineurin with cyclosporin A prevents growth of the pgm2Δ strain, suggesting that some downstream consequence of calcineurin activity is essential for the viability of this strain (15). However, we measured only a modest increase in PMC1 mRNA abundance in the pgm2Δ strain relative to the wild-type strain (unpublished results). Because Ca2+ accumulation is triggered within a minute of Li+ exposure, de novo protein synthesis is clearly not required for this response. These results lead us to conclude that the calmodulin-calcineurin signaling pathway must be involved in this regulatory circuit. However, the downstream signaling targets that modulate vacuolar Ca2+ uptake remain to be identified.
In a recent study, Mulet and coworkers (30) reported that trehalose 6-phosphate synthase (TPS)1 activity alters the ion transport characteristics of the K+ channels TRK1 and TRK2 through the modulation of cellular Glc phosphate levels. They also noted that like TPS1, PGM and hexokinase 2 are also capable of altering the levels of Glc phosphates and that elevated Glc phosphate levels activate TRK by way of the calcineurin pathway (30). Furthermore, the involvement of HAL4 and HAL5 protein kinases was also described in that study. In light of the current work and another recent study (3), Glc phosphates are clearly emerging as important regulators of intracellular cation homeostasis.
Despite lithium being the “oldest” drug used in the treatment of bipolar disorder, the complex molecular mechanisms of Li+ action remain a scientifically challenging problem. Because yeast and mammalian cells share many common characteristics at the level of basic metabolism, the use of the yeast model system may represent an extremely useful tool in elucidating how Li+ specifically influences Ca2+ signaling pathways through its effect on PGM. In light of the results of the current study, the relative importance of PGM as a moderator of intracellular signal transduction pathways in mammalian cells should clearly be explored in greater detail. Such studies could provide further insights into the mechanism(s) coupling Glc metabolism to Ca2+ homeostasis and should further elucidate whether the ability of Li+ to inhibit PGM activity plays a significant role in its therapeutic benefits for patients with bipolar disorders.
This work was supported by Hungarian Scientific Research Fund Grants F-038149 (to P. Csutora), T-038144 (to A. Miseta), T-032043 to (K. Sipos), and American Heart Association Southeast Affiliate Grant 0255121B (to D. M. Bedwell).
We thank Sarolta Bácskai, Ibolya Lamár, and Éva Moravcsik for valuable technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society