Regulation of Cellular Ca2+ by Yeast Vacuoles*

The role of vacuolar Ca2+ transport systems in regu- lating cellular Ca2+ was investigated by measuring the vacuolar CaZ+ transport rate, the free energy available to drive vacuolar Ca2+ transport, the ability of the vacuole to buffer lumenal Ca2+, and the vacuolar CaZ+ efflux rate. The magnitude of the Ca2+ gradient generated by the vacuolar H+ gradient best supports a 1 Ca2+:2 H+ coupling ratio for the vacuolar Ca2+iH+ exchanger. This coupling ratio along with a cytosolic Ca2+ concentration of 125 I ~ M would give a vacuolar free Ca2+ concentration of -30 pra The total vacuolar Ca2+ concentration is -2 111~ due to Ca2+ binding to vacuolar polyphosphate. The Ca2+ efflux rate from the vacuole is less than the growth rate indicating that the steady-state Ca2+ loading level of the vacuole is dependent mainly on the Ca2+ transport rate and the rate that vacuolar Ca2+ is diluted by growth. Based on the kinetic parameters of vacuolar Ca2+ accumulation in vitro, the maximum rate of Ca2+ accumulation in uiuo is expected to be "0.2 nmol of Ca2+ min" mg protein", a rate that is similar to the cellular Ca2+ accumulation rate. The cytosolic Ca2+ concentration increases from 0.1 to 1-2 as the extracellular Ca2+ concentration is raised from 0.3 u to 50 111~. The rise in cytosolic Ca2+ concentration increases cellular Ca2+ from 10 to 300 nmol Caz+/mg by increasing the rate of vacuolar

t h e Ca2+ pumps and Ca2+ channels which regulate the cytosolic CaZ+ concentration. The best-Characterized Ca2+ transport system in S. cerevisiae is the vacuolar Ca2+/H+ exchanger (3)(4)(5)(6). The importance of vacuolar Ca2+ transport in regulating cellul a r Ca2+ is suggested by the observation that mutants defective in the vacuolar H+-ATPase (7)(8)(9)(10) or in processing of vacuolar proteins (11) have reduced growth rates i n 100 mM Ca2+, whereas 100 mM Ca2+ has little influence on the wild-type growth rate. Since cytosolic Ca2+ concentration is elevated in some of the mutants that have vacuolar defects (8,12), it could be argued that vacuolar Ca2+ transport regulates cytosolic Ca2+. However, it is also possible that the elevated cytosolic Ca2+ is an indirect consequence of the vacuolar defect.
The goal of the study presented here is to evaluate the potential role of the vacuole in regulating the cellular Ca2+ con-Institutes of Health and Grants C07lbK and RO71AG from the Uni-* This work was supported by Grant GM 46495 from the National formed Services University of the Health Sciences. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. EXPERIMENTAL PROCEDURES Materials-Arsenazo 111 was obtained from Aldrich. Zymolyase lOOT was supplied by Seikagaku Kogyo, Rockville, MD. ,Va2+ was purchased from DuPont-New England Nuclear. All other chemicals were purchased from Sigma.
Yeast StrainslGrowth-The yeast strain used in this work was CuH3 (S288C, MATa his4419 ura3-52). Cells were g r o w n according to standard procedures (13). Standard yeast extract bacto-peptone dextran (YF' D) and synthetic media were prepared according to Sherman (13). Low phosphateYPD media was made according to the method described by Rubin (14).
Measurement of Ca2+ Accumulation by Whole Cells--Two methods were used to measure Ca2+ accumulation by cells. In one method, cells were incubated in media containing tracer ,Va2+ (0.1-1 pCi/ml) for different periods of time, and aliquots were removed, washed three times by centrifugation in YPD medium at 4 "C, and filtered through Millipore HA 0.45-pm filters that were prewashed with 20 m MgSO,.
The filters were washed three times with 5 ml of 20 m MgSO,, dried, placed in vials with Beckman Ready OrganicTM scintillation mixture, and the amount of Ca2+ in the cell was determined by scintillation counting.
In the other method, cells were incubated in media containing 1-100 m Ca2+ for different periods of time and then harvested and washed by centrifugation (4 "C). The cells were resuspended in 0.1 M KC], 10 ~l l~ PIPES,' pH 7.0, 0.1 M arsenazo I11 at a density 107/ml (36 "C). Enough digitonin (1 mg/ml) to permeabilize all the cell membranes was added, and the amount of Ca2+ released from the cell was determined spectrophotometrically using an SLM-Aminco DW2c dual wavelength spectrophotometer by measuring the increased absorbance caused by the formation of the arsenazo III-Ca2+ complex at 660 nm using 685 nm as a reference wavelength. When both methods (filtration and spectrophotometric) were compared under the same experimental conditions, the results were essentially the same. Isolation of Vacuole Membrane Vesicles-Vacuole membrane vesicles (vesiculated vacuoles) were prepared by the method developed by Anraku and co-workers (3) and stored at -70 "C in 0.1 M potassium glutamate, 10 m PIPES, pH 7.0, and 10% glycerol.
Permeabilization of Yeast Cells-Ca2+ accumulation by intact vacuoles was measured using either osmotically shocked partially regenerated spheroplasts (15) or cells permeabilized by treatment with the detergent digitonin. Treatment of cells with 0.2 mg digitonidmg protein for 10 min at 26 "C selectively permeabilizes the plasma membrane.
Measurement of Polyphosphate-Polyphosphate was extracted from the cells by the method described by Clark et al. (16) and measured by the procedure described by Ames (17).

RESULTS
Effect of ea2+ on the Growth Rate and Ca2+-loading &vel of S. cereuisiae cells-The role of the vacuole in regulating cellular Ca2+ was investigated. Varying the Ca2+ concentration in the growth medium from 1 1.1~ to 100 mM has little effect on the growth rate of S. cerevisiae cells (18)(19)(20)(21)(22)(23). The tolerance of S. cerevisiae to high Ca2+ concentrations is not due to the inability of Ca2+ to enter the cell since the cellular Ca2+ increases as the extracellular Ca2+ concentration is raised (Fig. 1, inset). In normal YPD medium (-0.3 mM Ca2+), the cellular Ca2+ is 10-20 nmoVmg protein. Addition of 50 m~ Ca2+ to the YF' D growth medium causes a relatively slow increase in the cellular Ca2+ to -225 nmoVmg protein after 5 h (Fig. 1) without changing the growth rate.
Ca2+ Accumulation by Permeabilized Cells and Vacuolar Membrane Vesicles-The following experiments were designed to determine the relationship between the cytosolic Ca2+ concentration and vacuolar Ca2+ accumulation. The kinetic and thermodynamic parameters for vacuolar Ca2+ transport were investigated.
ATP initiates Ca2+ accumulation by isolated vacuole membrane vesicles (4) as well as by permeabilized partially regenerated spheroplasts (Fig. 2). Since nigericin and other proton ionophores inhibit ATP-dependent Ca2+ accumulation (4), it was proposed by Ohsumi and Anraku that ATP hydrolysis by the vacuolar H+-ATPase establishes a proton gradient that drives Ca2+ transport. The Ca2+ KM of the Ca2+ transporter in both the isolated vacuole membrane vesicles (data not shown) and permeabilized cells is 25 p~ at pH 7.0 and 1 mM Mg2+ (Fig.  2 B ) . Permeabilized cells have a maximum Ca2+ accumulation rate (Vmax) of 35 nmol of Ca2+ min-' (mg cell protein)-' (Fig.  2B). Therefore, the vacuolar Ca2+ accumulation rate in the presence of 125 IIM Ca2+ is expected to be -0.2 nmol Ca2+ min-l (mg cell protein)-I which is comparable to the rate of whole cell Ca2+ accumulation in YPD medium (0.1-0.2 nmol Ca2+ min-I (mg cell protein)-'). Thus, the overall rate of cellular Ca2+ accumulation could depend on the rate of vacuolar Ca2+ accumulation and not on the rate of Ca2+ influx across the plasma membrane.
Effect of the Proton Gradient on Ca2+ Accumulation by Vacuole Membrane Vesicles-To determine if ATP is required directly by the Ca2+ transport system, H+ gradients across the yeast vacuolar membrane were formed in the absence ofATP by diluting K+-equilibrated vacuolar membrane vesicles into solutions containing differing ratios of Tris' and K+, and the ionophore, nigericin. Nigericin mediates the exchange of one proton for one K+ causing the formation of a H+ gradient which is expected to equal the K+ gradient. Ca2+ accumulation was measured spectrophotometrically using the Ca2+ indicator arsenazo 111. As shown in Fig. 3 A , the rate of Ca2+ accumulation by vacuole membrane vesicles is dependent on the K+ gradient (and therefore on the H+ gradient). No Ca2+ accumulation is observed in the absence of nigericin or when nigericin is replaced with the K+ ionophore, valinomycin, demonstrating that neither a K+ gradient nor K+ diffusion potential (inside-negative) drives Ca2+ transport. An apparent H+ K,,, of 2 p~ is determined from the relationship between the K+ gradient and the Ca2+ transport rate (Fig. 3B).
A 225-fold Ca2+ gradient across the vacuolar membrane in vivo would predict a free vacuolar Ca2+ concentration of about 30 p~ for cells growing in YPD. Since the total Ca2+ concentration in the vacuole is estimated to be 2 -4 mM, most of the vacuolar Ca2+ must be bound.
Effect of Vacuolar Polyphosphate on Ca2+ Uptake by S. cerevisiaePince yeast vacuoles contain large amounts of polyphosphate that range in size from 3 to 260 units with most being 7 to 20 units (251, the role of polyphosphate in vacuolar Ca2+ sequestration was investigated. The strain used in this study accumulates -0.1 mg polyphosphate/mg protein when grown in YPD medium. Ca2+ binding to polyphosphate was investigated using several experimental approaches. First, properties of Ca2+ binding to synthetic polyphosphate were determined. Next, Ca2+ binding to cellular polyphosphate was compared with that of synthetic polyphosphate. Finally, the effect of the vacuolar polyphosphate content on Ca2+ accumulation by whole cells and by cells with the plasma membrane permeabilized was determined. The affinity and capacity of polyphosphate to bind Ca2+ was determined spectrophotometrically using the Ca2+-indicator arsenazo 111. The free Ca2+ concentration in the presence of 80 pg/ml polyphosphate was measured as the total Ca2+ concentration was varied (Fig. 5).
Synthetic polyphosphates containing averages of 5, 18, and 31 phosphate residues were used. At pH 6.0 (close to the vacuolar pH in vivo (26,2711, the Ca2+ dissociation constant is 6-8 p (Fig. 5). The 31-residue polyphosphate binds 1 Ca2+/2.5 phosphate residues with high affinity suggesting that, except for the 3 phosphate residues on each end, the high affinity Ca2+-binding site is formed by 2 phosphate residues. The binding of Ca2+ does not change significantly when the pH is decreased to 5.7 or increased to 6.8 (data not shown); however, Mg2' is a strong competitor for Ca2+ binding (Fig. 5). In the presence of 1 mM M e , the dissociation constant is increased to -1 mM. The data indicate that M$+ and Ca2+ bind to polyphosphate with similar affinities.
Treatment of yeast cells with high detergent concentrations exposes intracellular high affinity Ca2+-binding sites (Fig. 6). The detergent-exposed latent Ca2+ binding activity of S. cerevisiae can be primarily attributed to polyphosphate for the following reasons. 1) Latent Ca2+ binding activity is directly proportional to the polyphosphate content of the cells which varies with the phosphate concentration (0.01-1 m~) of the growth media.
2) The Ca2+/phosphate residue stoichiometry (0.4 Ca2+/ phosphate residue in 50 p Ca") of the latent Ca2+ binding is The effect of the cellular polyphosphate content on Ca2+ uptake was determined both in vitro (Fig. 7 A ) and in vivo (Fig.   7B 1. The in vitro measurements were performed using cells in which the plasma membrane was permeabilized by digitonin treatment which preserves vacuolar polyphosphate. The polyphosphate content of the vacuoles was vaned by incubating cells in phosphate-free YPD for various times. ATP-dependent (1 pCi/ml) was then added (t = 0) and at the indicated times 1-ml aliquots were removed, filtered, and the filters were washed four times with 5 ml of cold 20 m~ MgSO,. Aliquots were also removed for polyphosphate determination; the polyphosphate content continuously changes during the time course of CaZ+ loading due to the accumulation and metabolism of the phosphate by the cells (which did not depend on Ca2+). The polyphosphate content at 420 min was 1.1, 2.4,3.7, and 5.9 pmol of phosphate residuelmg protein for cells placed in 25, 125,250, and 500 p~ phosphate, respectively. Ca2+ accumulation by permeabilized cells was proportional to their polyphosphate content (Fig. 7A). The ratio of Ca2+ accumulatedphosphate residue was -0.3. The Ca2+-loading capacity of permeabilized cells (470 nmol Ca2+/mg) is -47 times greater than the Ca2+-loading level in vivo reflecting the difference in the cytosolic Ca2+ concentration in vivo (0.1-0.2 p~) and the Ca2+ concentration in the assay medium (-27 p~ free Ca2+).
The role of polyphosphate in sequestering cellular Ca2+ is further demonstrated by measuring the effect of cellular polyphosphate levels on the in vivo Ca2+ content (Fig. 7B). The amount of cellular Ca2+ increases with increasing cellular polyphosphate. The ratio of cellular Ca2+ to polyphosphate residues remains constant at about 0.003, once again demonstrating that the in vivo vacuolar Ca2+ level for cells grown in 0.3 m~ Ca2+ is -100 times lower than the level of polyphosphate Ca2+binding sites. The CaZ+ content of cells growing in 50 m~ Ca2+ is also proportional to cellular polyphosphate content; however, the ratio of Ca2+ to polyphosphate residues increases to -0.2 (data not shown).
Vacuolar Cu2+ EflLuc-The above data indicate that the ratelimiting step of Ca2+ accumulation by s. cerevisiae cells is vacuolar transport and that polyphosphate acts as a Ca2+ sink within the vacuole lumen. The next question that was addressed was whether Ca2+ sequestration in the vesicles was reversible. The rate of Ca2+ efflux from yeast vacuoles was measured both in vitro and in vivo. Permeabilized cells equilibrated with 25 p~ Ca2+ sequester -6 nmol Ca2+/mg protein (about 0.14 nmoYmg would be free, most of the rest is bound to polyphosphate) (Fig. 8). This observation further supports the conclusion that the free vacuole Ca2+ concentration in vivo is -30 p~ since cells accumulate 10-20 nmol Ca2+/mg protein.
The rate of Ca2+ efflux from the vacuole of permeabilized cells was measured following dilution into an EGTA solution. The Ca2+ efflux rate was only 0.016 nmol Ca2+ min" (mg protein)-l; therefore, the half-time for Ca2+ release from the vacuole exceeds the doubling time for cells under normal growth conditions. 45Ca2+ release was observed when unlabeled Ca2+ was added to the dilution medium, but this is due to 45Ca2+-Ca2+ exchange mediated by the Ca2+/H+ exchanger since in parallel experiments it was observed that extravacuolar 45Ca2+ was accumulated by vacuoles in exchange with lumenal unlabeled Ca2+.
The Ca2+ ionophore, A23187, was able to induce the release of all sequestered Ca2+ in the absence of extravacuolar Ca2+. Like the intrinsic Ca2+/2H+ exchanger, A23187 mediates the transmembrane exchange of one Ca2+ for two protons. Perhaps the reason that the intrinsic exchanger fails to release Ca2+ from the vacuole lumen while A23187 does is that A23187 has a relatively high affinity for protons (high pK) so that it transports protons in both directions, while the intrinsic exchanger may have a low affnity for protons (low pK) on the cytosolic Panel A, the predicted steady-state vacuolar Ca2+ is plotted as a function of the cytosolic CaZ+ concentration. The model used to predict the steady-state vacuolar Ca2+ (inset) assumes that vacuolar Ca2+ transport rate is given by the Michaelis-Menten equation, and that vacuolar Ca2+ efflux is essentially zero so that steady-state is reached when the vacuolar Ca2* transport rate (reaction 2) is equal to the rate that vacuolar Ca2+ is diluted by cell growth. Panel B . The cytosolic CaZ+ concentration that is predicted to give the experimentally observed cellular CaZ+ content ( Fig.  1) is plotted against the extracellular Ca2+ concentration.
side blocking the exchange of cytosolic protons with lumenal Ca2+.
To compare the rate of Ca2+ efflux from whole cells with that of permeabilized cells, cells were grown for several generations in YPD medium containing a variety of Ca2+ concentrations with 45Ca2+ and then shifted to YPD medium without 45Ca2+.
The 45Ca2+ efflux rate was measured by filtration. As is seen in Fig. 8B, most of the cellular Ca2+ is in a stable nonreleaseable pool, and the amount of Ca2+ in this stable pool increases with increasing extracellular Ca2+ concentration. This experiment indicates that the in vivo vacuolar Ca2+ emux rate is also very low.

DISCUSSION
Model of Ca2+ Metabolism by S. cerevisiae-Amodel for Ca2+ metabolism by S. cerevisiae cells is presented in the inset of Fig.  9. Steady-state Ca2+ gradients across the plasma membrane are rapidly established (within seconds) following changes in the extracellular Ca2+ concentration. The mechanisms by which Ca2+ enters the cell and is actively pumped out are not known, but a 25,000-fold Ca2+ gradient across the plasma membrane can apparently be maintained. The rate of Ca2+ accumulation by the vacuole (the major Ca2+-sequestering organelle) is dependent on the rate of Ca2+ transport mediated by the Ca2+/2H+ exchanger and the cell growth rate. The vacuolar Ca2+ uptake rate is predicted to increase linearly with the cytosolic Ca2+ concentration because of the relatively high ICM (25 p~) and the high Ca2+ binding capacity of the vacuolar lumen (-400 nmol Ca2+/mg protein). Because most of the cellular Ca2+ is found in the vacuole, the rate of cellular Ca2+ accumulation is dependent on the rate of vacuolar Ca2+ accumulation. Because the vacuolar Ca2+ efflux rate is less than the growth rate, the steady-state vacuolar Ca2+ level is reached when the rate of vacuolar Ca2+ transport is equal to the rate of vacuolar Ca2+ dilution by growth. min-l, the total vacuolar Ca2+ content is predicted to be directly proportional to the cytosolic Ca2+ concentration (233[Ca2+lcytoso1 nmol ( p~ mg protein) (Fig. 9A). According to the proposed model, the cytosolic Ca2+ concentrations that would give the observed cellular Ca2+ levels of 10-300 nmol Ca2+/mg protein would be 0.1-1.4 p~ Ca2+, meaning the cytosolic Ca2+ would have to increase from 0.1 to 1.4 p~ as the extracellular Ca2+ increases from 0.3 to 100 m~ (Fig. 9B).
Addition of 100 m M Ca2+ to the growth medium has no significant effect on the growth rate suggesting that increasing the cytosolic Ca2+ concentration to 1-2 p~ does not alter the cellular physiology of exponentially growing yeast cells. In this regard, it is interesting that at these cytosolic Ca2+ concentrations, calmodulin would be expected to be activated by Ca2+ and yet there is no influence on cell growth. These results complement the findings of Geiser et al. (28) which demonstrate that, while calmodulin is required for growth, Ca2+ binding to calmodulin is apparently not required. Taken together these results indicate that Ca2+ binding to calmodulin does not affect vegetative cell growth.