Extracellular Ca²⁺ sensing contributes to excess Ca²⁺ accumulation and vacuolar fragmentation in a pmr1Δ mutant of S. cerevisiae

Changes in the cytosolic Ca²⁺ concentration are thought to participate in a variety of physiological processes in yeast, including cell-cycle control (Hartley et al.,1996; Iida et al.,1990a); escape from G₀/G₁ arrest following carbon source limitation (Eilam and Othman, 1990; Eilam et al.,1990; Kaibuchi et al.,1986); glucose and galactose sensing(Tokes-Fuzesi et al., 2002);the mating response (Nakajima-Shimada et al., 1991; Withee et al.,1997); protein processing in the secretory pathway(Durr et al., 1998); and adaptation to environmental stress (Batiza et al., 1996; Denis and Cyert,2002; Mori et al.,1998). The basic features of Ca²⁺ signaling in Saccharomyces cerevisiae appear to be remarkably similar to the mechanisms used in mammalian cells. Ca²⁺ acts as a signaling ion in yeast through the activation of the calmodulin/calcineurin pathway, which leads to alterations in the transcription of a large number of genes(Matheos et al., 1997; Stathopoulos and Cyert, 1997; Yoshimoto et al., 2002).

A relatively small number of Ca²⁺ transporters appear to maintain cellular Ca²⁺ homeostasis in yeast. Among these are the vacuolar Ca²⁺-ATPase Pmc1p(Cunningham and Fink, 1994b);the vacuolar Ca²⁺/H⁺ exchanger Vcx1p/Hum1p(Cunningham and Fink, 1996; Pozos et al., 1996); the endoplasmic reticulum (ER) Ca²⁺-ATPase Cod1p/Spf1p(Bonilla et al., 2002; Cronin et al., 2000; Cronin et al., 2002; Suzuki and Shimma, 1999); and the Golgi Ca²⁺-ATPase Pmr1p(Antebi and Fink, 1992; Rudolph et al., 1989; Sorin et al., 1997). Remarkably, the action of this small group of Ca²⁺ transporters maintains the resting cytosolic Ca²⁺ level between 50 and 200 nM when environmental Ca²⁺ concentrations range from <1 μM to>100 mM (Batiza et al.,1996; Iida et al.,1990b; Miseta et al.,1999b). When the level of environmental Ca²⁺ is elevated, the calmodulin/calcineurin signaling pathway activates the expression of many genes (including those encoding Pmr1p and Pmc1p) by the transcription factor Tcn1p/Crz1p. By contrast, the level of Vcx1p is slightly reduced by high environmental Ca²⁺, and its activity is further repressed by calcineurin activation through a post-translational mechanism(Cunningham and Fink, 1996; Matheos et al., 1997; Stathopoulos-Gerontides et al.,1999; Yoshimoto et al.,2002).

The vacuole is the principle Ca²⁺ storage organelle in yeast,and normally contains >95% of the total cellular Ca²⁺(Dunn et al., 1994; Eilam et al., 1985). However,it has been shown that the loss of the Golgi-localized Ca²⁺transporter Pmr1p causes an increased sensitivity to high environmental Ca²⁺ when vacuolar Ca²⁺ transport is compromised,indicating that the Golgi apparatus also plays an important role in Ca²⁺ sequestration (Miseta et al., 1999b; Tanida et al.,1995). Furthermore, Pmr1p has also been reported to be involved in maintaining the resting Ca²⁺ concentration within the ER(Strayle et al., 1999),whereas both Pmr1p and Pmc1p influence Ca²⁺-dependent functions within the secretory pathway such as protein degradation in the ER and protein sorting in the Golgi apparatus (Durr et al., 1998). The ability of these transporters to influence Ca²⁺-dependent processes in multiple organelles might be due to their movement through these compartments during the transit to their final cellular destinations. However, it has been reported that the distribution of Pmc1p in Golgi fractions increases in a pmr1Δ strain,suggesting that its abundance in compartments of the secretory pathway might be influenced by the lumenal Ca²⁺ concentration(Marchi et al., 1999). These findings illustrate both the complexity and the sophisticated regulatory mechanisms that control cellular Ca²⁺ homeostasis in yeast.

In mammalian cells, free Ca²⁺ located in the ER serves as a mobilizable pool that can be released into the cytosol in response to an appropriate stimulus. The resulting increase in cytosolic Ca²⁺ can then activate signaling pathways that alter the expression of many genes in a coordinated manner (Putney,1992). In many non-excitable cells, the release of ER Ca²⁺ can also induce a store depletion signal that results in the influx of Ca²⁺ ions across the plasma membrane in a process termed capacitative Ca²⁺ entry (CCE). Recent studies have provided evidence that yeast cells might also utilize a mechanism that couples intracellular store depletion to Ca²⁺ uptake across the plasma membrane. A pmr1Δ mutant has been shown to exhibit a higher rate of Ca²⁺ uptake than the WT strain(Antebi and Fink, 1992; Halachmi and Eilam, 1996; Rudolph et al., 1989; Sorin et al., 1997). This led to the model that the depletion of Golgi Ca²⁺ stores can stimulate Ca²⁺ uptake into yeast cells in a manner analogous to CCE in mammalian cells (Csutora et al.,1999; Durr et al.,1998; Locke et al.,2000). In contrast to the pmr1Δ mutant, a pmc1Δ strain exhibits a reduced level of total cellular Ca²⁺ (Cunningham and Fink,1994b). Since Pmc1p is the only known vacuolar Ca²⁺-ATPase in yeast and the vacuole normally contains the bulk of total cellular Ca²⁺, this suggests that Ca²⁺ uptake across the plasma membrane is coupled to the ability of the cell to remove it efficiently from the cytosol.

In the current study, we examined how a reduced level of divalent cations in the growth medium influences cellular Ca²⁺ homeostasis in the pmr1Δ strain. We found a large increase in Ca²⁺uptake and accumulation under these conditions, which led to activation of the calcineurin signaling pathway. Consistent with this high level of Ca²⁺ uptake, we found that the PMC1 gene was required for growth of the pmr1Δ mutant under these growth conditions. Our observation that cellular Ca²⁺ uptake increases as the concentration of environmental Ca²⁺ decreases suggests that an extracellular Ca²⁺ sensor is capable of coupling Ca²⁺uptake to extracellular Ca²⁺ levels. Finally, we found that conditions of cellular Ca²⁺ stress result in a vacuolar fragmentation phenotype in both WT and mutant yeast strains. This might serve as an adaptive mechanism to maintain cellular Ca²⁺ homeostasis under these stress conditions.

The yeast strains used in this study are described in Table 1. The Pmc1p-GFP fusion plasmid was constructed as follows. A 4.9 kb HindIII fragment containing the PMC1 gene was isolated from pKC44 (gift from Kyle Cunningham) and ligated into pBlueKS+. Site-directed mutagenesis was performed according to the QuickChange Site Directed Mutagenesis (Stratagene) protocol to generate a BssHII site at the 3′ end of PMC1 (5 codons upstream of the stop codon) using the oligonucleotides DB571(5′-CTGATAGTCC TTGGCGCGCC AACTTTTATT AATAGACGC-3′) and DB572(5′-GCGTCTATTA ATAAAAGTTG GCGCGCCAAG GACTATCAG-3′). A 635 bp BglII/SnaBI fragment was isolated from the resulting plasmid and cloned back into pBlueKS+/PMC1 for DNA sequencing. A BssHII DNA fragment carrying the GFP gene was generated using primers DB573 (5′-CGATAAGGAT TGGCGCGCCA AATTCATGAG-3′) and DB574(5′-CGATAAGGAT TGGCGCGCCA AATTCATGAG-3′) and the template pRSETB-GFP (gift from Kelly Thatchell), and ligated into the BssHII site of pBlueKS+/PMC1. The correct GFP coding sequence was confirmed by automated DNA sequencing and a HindIII fragment carrying the PMCI-GFP construct was ligated into the low-copy yeast shuttle vector pSEYC58 (Emr et al., 1986). This final plasmid was then transformed into the pmc1Δ(YDB0224) and the pmr1Δ/pmc1Δ (YDB0276) strains. The HindIII fragment carrying the complete PMC1-GFP open reading frame was also inserted into the high-copy shuttle vector pSEY8(Emr et al., 1986), digested with ApaI, and the digested linear DNA was transformed into the pmc1Δ (YDB0224) and the pmr1Δ/pmc1Δ (YDB0276) strains to integrate it into the genome of each strain. With both approaches, the construct reversed the EGTA-sensitive phenotype of the pmr1Δ/pmc1Δstrain and the high Ca²⁺-sensitive phenotype of the pmc1Δ strain.

Bacterial strains were grown on standard media(Miller, 1992). Yeast strains were maintained on YP medium containing 2% D-glucose (YPD) or synthetic minimal medium containing 2% D-glucose (SMD) and other supplements as required(Burke et al., 2000). Culture media were routinely buffered with 40 mM Mes-Tris, pH 5.5.

Total cellular Ca²⁺ was measured using flame photometry as described previously (Miseta et al.,1999a). Cells were grown in 30-40 ml of YPD or the same medium supplemented with either 1 mM EGTA or 50 mM CaCl₂. Cultures were harvested at a cell density of ∼1 OD₆₀₀/ml, harvested by centrifugation and washed with 20 ml of YPD. Cells were then transferred to previously weighed microfuge tubes and harvested by centrifugation in a micro-centrifuge (16,000 g) for 5 minutes. After aspirating the supernatant, a second spin was conducted for 3 minutes. Samples were weighed, then dried in a Speedvac. The dried samples were weighed again and resuspended in HCl for flame photometric measurements.

Ca²⁺ uptake measurements were similar to a method described previously (Halachmi and Eilam,1996). Cells were grown in YPD to 0.7-1.0 OD₆₀₀/ml,harvested, washed twice with ddH₂O, then re-suspended in buffer containing 25 mM Mes-Tris, pH 6.0 supplemented with 20 mM glucose to 1 OD₆₀₀/ml and incubated at 30°C for 10 minutes. Uptake was initiated by the addition of 1 μCi/ml ⁴⁵Ca²⁺. At the indicated time-points, 1 ml aliquots were collected by filtration through 0.45μm membrane filters (Gelman Sciences) using a vacuum manifold. Membranes were immediately washed with two 5 ml aliquots of ice-cold wash buffer (20 mM MgCl₂, 0.2 mM LaCl₃), dried, and the cell-associated radioactivity was measured by liquid scintillation counting.

RNA extraction and northern blot analysis were carried out as described previously (Bonetti et al.,1995). Strains were grown overnight in YPD medium or YPD supplemented with either 1 mM EGTA or 50 mM CaCl₂ to ∼1 OD₆₀₀/ml. A 0.44 kb probe for the PMC1 mRNA was generated by PCR using primers DB490 (5′-ATGTCTAGACAAGACGAAAA-3′) and DB491(5′-ATACTGTGGAGGTTGCATCC-3′). As control, an ACT1 probe was generated using the primers DB154(5′-GCGCGGAATTCAACGTTCCAGCCTTCTACG-3′) and DB155(5′-GGATGGAACAAAGCTTCTGG-3′). Probes were labeled with[α-³²P]dATP using the random hexamer method(Sambrook et al., 1989). The specific band representing the PMC1 mRNA was confirmed by its absence in RNA extracted from the pmc1Δ strain YDB0224. Gels were quantitated by PhosphorImager analysis (Molecular Dynamics). The relative mRNA levels were normalized using the ACT1 mRNA as internal control after background correction.

Aequorin assays were carried out as described earlier(Miseta et al., 1999a). The two-micron-based plasmid pDB617 expressing a functional apoaequorin gene(pAEQ) was transformed into yeast. Cells containing pAEQ were grown in SMD medium and harvested in the early logarithmic growth phase (0.5-1.0 A₆₀₀ units/ml). 10 A₆₀₀ units of cells were harvested and re-suspended in 0.2 ml of aequorin test medium, which consists of SMD medium (which normally contains 1 mM Ca²⁺) supplemented with 2 mM EGTA and 40 mM MES-Tris, pH 6.5. To convert the apoaequorin to aequorin, 20μl of 590 μM coelenterazine (dissolved in methanol) was added, and the cells were incubated for 20 minutes at room temperature. The cells were then briefly centrifuged, and the supernatant containing excess coelenterazine was removed. The cell pellets were washed again in 0.5 ml of aequorin test medium,re-suspended to a cell density of 1 OD₆₀₀/0.1 ml and incubated at room temperature for 20 minutes before initiating the experiment. Bafilomycin A₁ (5 μM) was added from a 100 μM stock solution (dissolved in DMSO) 10 minutes prior to the measurements(Abe and Horikoshi, 1995). After detecting the baseline light emission, 100 mM Ca²⁺ was administered into the chamber to generate a Ca²⁺ shock. A Berthold Lumat 9050 luminometer was used to collect aequorin light emission (L) data at 200 millisecond intervals. The data were downloaded directly to a computer and transferred to Microsoft Excel 5.0 for analysis. After measuring maximal light emission from crude cell extracts upon Ca²⁺ addition(L_max), L/L_max values were plotted on a standard curve to estimate the free cytosolic Ca²⁺ concentrations, as previously described (Miseta et al.,1999a).

Cells expressing the Pmc1p-GFP fusion protein and/or stained with FM 4-64 were collected, re-suspended in fresh YPD medium, or YPD medium supplemented with Ca²⁺ or EGTA, to 10 OD₆₀₀ units/ml and 10 μl of the culture was mounted on slides coated with concanavalin A, covered with a coverslip and viewed immediately. FM 4-64 staining was carried out as published (Vida and Emr, 1995)with minor modifications. Briefly, yeast cells were grown in the indicated media to mid-log phase, 2 OD₆₀₀ units of cells were collected and re-suspended in 100 μl of YPD (Ca²⁺ was omitted at this step because it decreased total fluorescence). A 1 μl aliquot of an FM 4-64 stock (4 mM in DMSO) was added to the cells, and they were stained for 10-15 minutes at 30°C. Cells were collected and re-suspended in 200 μl fresh media supplemented with Ca²⁺ or EGTA and incubated at 30°C for 40-50 minutes. Light microscopy was conducted using a Leitz Orthoplan microscope with epifluorescence optics and Hoffman Modulation Contrast optics. The images were acquired with a Photometrics CH250 liquid-cooled CCD high-resolution monochromatic camera (Roper Scientific; Tucson, AZ) and analyzed by IPLab Spectrum software from Scanalytics (Fairfax, VA).

We first examined whether the loss of Pmc1p function altered the ability of the pmr1Δ strain to grow in the presence of chelating agents that reduced the level of Ca²⁺ and other divalent cations. We found that the pmr1Δ strain could grow on YPD plates containing the Ca²⁺ chelator 1,2 bis(2-aminophenoxy) ethane-N,N,N,N-tetraacetic acid (BAPTA; 1 mM), whereas the pmr1Δ/pmc1Δstrain could not (Fig. 1A,B). Similar results were obtained with YPD plates containing 2 mM EGTA (data not shown). These results were consistent with the previous conclusion that Pmc1p plays a role in Ca²⁺ uptake into secretory compartments(Bonilla et al., 2002; Durr et al., 1998; Locke et al., 2000). However,it was surprising that Pmc1p influenced cell growth when the availability of Ca²⁺ and other divalent cations was reduced, since previous studies have shown that PMC1 expression and activity is normally induced by calcineurin when cellular Ca²⁺ stress increases(Cunningham and Fink, 1994b; Marchi et al., 1999). To determine whether calcineurin activation is required for growth of the pmr1Δ strain under these conditions, we challenged these strains with cyclosporin A (CsA) on YPD plates containing 1 mM BAPTA(Fig. 1C). We found that neither the pmr1Δ nor the pmr1Δ/pmc1Δ strains were able to grow in the presence of both CsA and BAPTA, demonstrating that calcineurin activity is necessary for growth of the pmr1Δ mutant when the environmental concentration of divalent cations is decreased. CsA inhibition was also observed when these strains were grown on YPD plates containing 2 mM EGTA, but not on YPD plates containing CsA that lacked one of these chelating agents(data not shown). When taken together, these results indicate that one or more genes regulated by the calcineurin signaling pathway, including Pmc1p, plays an important role in Ca²⁺ homeostasis in the pmr1Δstrain when the concentration of divalent cations in the environmental is reduced.

Previous studies have shown that PMC1 expression increases through a calcineurin-dependent mechanism as the level of environmental Ca²⁺ increases (Cunningham and Fink, 1994b; Marchi et al.,1999). However, the results described above suggest that the normal pattern of PMC1 expression is altered in the pmr1Δ strain. To understand these findings better, we next examined the level of PMC1 mRNA in the WT and pmr1Δstrains using northern blot analysis (Fig. 2A). As expected, we found that the level of PMC1 mRNA in the WT strain increased with increasing environmental Ca²⁺. The PMC1 mRNA level in the pmr1Δ strain was only slightly higher than the WT strain when grown in either YPD (which contains 0.3 mM Ca²⁺) or YPD supplemented with 50 mM CaCl₂. By contrast,the steady-state level of PMC1 mRNA in the pmr1Δstrain was ∼5-fold higher than normal when these strains were grown in YPD medium containing 1 mM EGTA. In fact, the PMC1 mRNA level in the pmr1Δ strain was higher in cells grown in the presence of a reduced level of divalent cations than under any other condition tested. These results confirm that the pattern of PMC1 expression is significantly altered in the pmr1Δ strain.

Since the expression of the PMC1 gene is controlled by calcineurin activity (Cunningham and Fink,1994b; Marchi et al.,1999), the results above suggest that the pmr1Δstrain might experience Ca²⁺ stress even when grown in the presence of low environmental Ca²⁺. To test this possibility, we compared the total cellular Ca²⁺ levels in WT and pmr1Δstrains when grown under different environmental Ca²⁺ conditions. Consistent with previous studies (Halachmi and Eilam, 1996; Sorin et al.,1997), we found that the level of cellular Ca²⁺ in the pmr1Δ strain was 1.4-fold higher than the WT strain when grown in YPD and 1.5-fold higher than WT in YPD containing 50 mM CaCl₂. By contrast, we found that the pmr1Δ strain contained∼7.5-fold more total cell Ca²⁺ than the WT strain when grown in YPD medium supplemented with 1 mM EGTA(Fig. 2B). This level of total cellular Ca²⁺ was actually higher than the level measured when this strain was grown in YPD supplemented with 50 mM CaCl₂. These results confirm that the pmr1Δ strain exhibits excessive Ca²⁺ accumulation when grown in media containing a reduced level of divalent cations, which results in the transcriptional activation of genes controlled by the calmodulin/calcineurin signaling pathway.

The results of previous studies have indicated that Pmc1p contributes to the filling of ER Ca²⁺ stores in a pmr1Δ strain(Bonilla et al., 2002; Durr et al., 1998). Given our finding that total cellular Ca²⁺ is significantly higher in a pmr1Δ strain when the level of divalent cations in the growth medium is reduced, we next examined Ca²⁺ uptake by the pmr1Δ/pmc1Δ strain when grown in YPD medium. Consistent with previous reports, we found that the rate of Ca²⁺uptake in the pmr1Δ strain was 1.8-fold higher than the WT strain (Halachmi and Eilam,1996; Sorin et al.,1997), whereas the pmc1Δ strain had a Ca²⁺ uptake rate that was ∼20% lower than normal(Fig. 3A). By contrast,Ca²⁺ uptake in the pmr1Δ/pmc1Δ mutant was 3.5-fold higher than the WT strain (and almost 2-fold higher than the pmr1Δ strain).

We next examined total cellular Ca²⁺ levels following the growth of these strains in YPD medium. We found that the total cellular Ca²⁺ level in the pmc1Δ strain was roughly 2-fold lower than the WT strain, as previously reported(Cunningham and Fink, 1994b). By contrast, the pmr1Δ/pmc1Δ strain contained 3.8-fold more Ca²⁺ than the WT strain, and 2.2-fold more total cellular Ca²⁺ than the pmr1Δ strain(Fig. 3B). These results demonstrate that the pmc1Δ mutation further exacerbates the Ca²⁺ hyper-accumulation phenotype of the pmr1Δstrain and suggest that the growth defect observed when the pmr1Δ/pmc1Δ strain is grown in a low Ca²⁺ environment is caused by an excessive cellular Ca²⁺load in combination with a reduced ability to sequester this excess Ca²⁺ adequately into intracellular compartments.

It was previously shown that Vcx1p activity is downregulated upon calcineurin activation, suggesting that this protein does not play a significant role in Ca²⁺ sequestration under conditions of high Ca²⁺ stress (Cunningham and Fink, 1996; Pozos et al.,1996). Consistent with this conclusion, we previously demonstrated that the presence of a vcx1Δ mutation did not have any effect on the level of total cellular Ca²⁺ when the growth medium was supplemented with more than 5 mM CaCl₂(Miseta et al., 1999a). However, some residual Vcx1p activity remains under such repressing conditions, since a pmc1Δ/vcx1Δ strain is more sensitive to high extracellular Ca²⁺ than a pmc1Δmutant (Cunningham and Fink,1994a; Miseta et al.,1999a; Pozos et al.,1996). To gain further insight into the role of Vcx1p in the maintenance of Ca²⁺ homeostasis under conditions of Ca²⁺stress, we measured total cellular Ca²⁺ levels in the pmc1Δ and pmc1Δ/vcx1Δ strains. We found that the total cellular Ca²⁺ level in the vcx1Δ/pmc1Δ strain was ∼20% lower than the pmc1Δ mutant when these strains were grown in YPD medium. When grown in YPD supplemented with 50 mM CaCl₂, the total cellular Ca²⁺ level increased in both strains, but the level measured in the vcx1Δ/pmc1Δ strain was ∼36% lower than the pmc1Δ strain (Fig. 4A). These results directly implicate Vcx1p in Ca²⁺sequestration in the pmc1Δ mutant under conditions of high Ca²⁺ stress, and suggest that a functionally significant level of Vcx1p activity is maintained in this strain when grown in the presence of high extracellular Ca²⁺.

The Vcx1p Ca²⁺/H⁺ exchanger utilizes the proton gradient across the vacuolar membrane to help maintain cytosolic Ca²⁺ levels within a narrow physiological range. The vacuolar proton gradient is maintained by the vacuolar H⁺-ATPase(Forster and Kane, 2000),which is sensitive to the inhibitor bafilomycin A₁. We previously used a cytosolic aequorin reporter system to characterize how bafilomycin A₁ influenced the regulation of cytosolic Ca²⁺ levels(Miseta et al., 1999a). We found that Vcx1p plays a key role in rapidly restoring basal cytosolic Ca²⁺ levels following a rise in the cytosolic Ca²⁺concentration. Since the pmr1Δ/pmc1Δ mutant lacks the two major Ca²⁺-ATPases involved in the maintenance of cellular Ca²⁺ homeostasis, we next used bafilomycin A₁and the aequorin reporter system to examine the role of Vcx1p in controlling cytosolic Ca²⁺ levels in the pmr1Δ/pmc1Δ mutant. We found that bafilomycin A₁ treatment increased the cytosolic Ca²⁺ level in this strain significantly (Fig. 4B). Using a standardization procedure described previously(Miseta et al., 1999a), we found that the cytosolic Ca²⁺ level in the pmr1Δ/pmc1Δ mutant increased from ∼l60 nM to∼260 nM following bafilomycin A₁ treatment. By contrast, a similar treatment did not alter the basal cytosolic Ca²⁺ levels in WT, pmr1Δ and pmc1Δ strains (data not shown). These results suggest that Vcx1p plays an important role in the maintenance of the resting cytosolic Ca²⁺ level in the pmr1Δ/pmc1Δ strain.

To examine further the role of Vcx1p in Ca²⁺ homeostasis of the pmr1Δ/pmc1Δ strain, we next tested the ability of Vcx1p to regulate cytosolic Ca²⁺ levels following the exposure of this strain to a 100 mM CaCl₂ shock(Fig. 4B). Immediately following the addition of this Ca²⁺ bolus, the cytosolic Ca²⁺ level rapidly increased to ∼300 nM. In the absence of bafilomycin A₁ treatment, we found that the pmr1Δ/pmc1Δ mutant could successfully recover from this rapid increase in cytosolic Ca²⁺, with the resting cytosolic Ca²⁺ level returning to ∼180 nM within 30 seconds. By contrast, a brief pre-treatment with bafilomycin A₁ prior to the Ca²⁺ shock completely eliminated the ability of this strain to compensate for this abrupt increase in cytosolic Ca²⁺. These results demonstrate that Ca²⁺/H⁺ exchange by Vcx1p plays a key role in the maintenance of cellular Ca²⁺ homeostasis in the pmr1Δ/pmc1Δ strain.

Previous studies have suggested that newly synthesized Pmc1p contributes to the maintenance of Ca²⁺ levels within the secretory pathway during its transit to the vacuole in the pmr1Δ mutant(Bonilla et al., 2002; Durr et al., 1998; Locke et al., 2000) (this study). Subcellular fractionation of a pmr1Δ strain has shown that Pmc1p is present not only in vacuolar fractions, but also overlaps fractions containing Golgi markers (Marchi et al., 1999). These results suggest that a significant amount of Pmc1p may be retained in the Golgi apparatus in the pmr1Δmutant. To determine whether the Golgi localization of Pmc1p could be visualized directly in yeast cells, we constructed a Pmc1p-GFP fusion protein. Following the integration of this construct into the nuclear genome of pmc1Δ and pmr1Δ/pmc1Δ strains, we found that the Pmc1p-GFP fusion restored normal Pmc1p function, as indicated by its ability to complement both the Ca²⁺ sensitivity of the pmc1Δ mutant and the EGTA sensitivity of the pmr1Δ/pmc1Δ mutant (data not shown). Thus, these complemented pmc1Δ, and pmr1Δ/pmc1Δ strains were functionally equivalent to WT and pmr1Δ strains, respectively.

We first asked whether the majority of the Pmc1p-GFP fusion protein co-localized with vacuoles as indicated by FM 4-64, a vital stain that accumulates in the yeast vacuolar membrane(Vida and Emr, 1995). As shown in Fig. 5, we observed complete co-localization of these two markers in the expected vacuolar pattern in WT cells grown in YPD medium. Whereas FM 4-64 and Pmc1p-GFP fluorescence also co-localized completely in the pmr1Δ strain, we found that both markers produced an identical pattern of highly fragmented staining. FM 4-64 staining was also used to examine the vacuolar morphology in the pmc1Δ and pmr1Δ/pmc1Δ strains. The vacuolar morphology was normal in the pmc1Δ strain. By contrast, the pmr1Δ/pmc1Δ strain exhibited a highly fragmented, frequently tubular, pattern of fluorescence. No Pmc1p-GFP fluorescence could be detected that was distinct from the FM 4-64 staining in either the WT or pmr1Δ strain, indicating that any accumulation in the Golgi apparatus (or another subcellular location) was below the resolution of this assay. However, these results demonstrate that significant vacuolar fragmentation occurs in strains containing the pmr1Δmutation when grown in standard YPD medium. On the basis of our finding that the pmr1Δ and pmr1Δ/pmc1Δ strains both undergo Ca²⁺ stress under these conditions, these results suggest that Ca²⁺ stress might induce vacuolar fragmentation.

To test this hypothesis, we next used the Pmc1p-GFP fusion protein to monitor vacuolar morphology in the WT and pmr1Δ strains grown in media containing different Ca²⁺ concentrations(Fig. 6). We found that WT cells grown in low and intermediate environmental Ca²⁺concentrations (YPD supplemented with 1 mM EGTA, YPD alone, or YPD supplemented with 5 mM CaCl₂) exhibited normal vacuolar morphology(Fig. 6A). However, this strain exhibited an increasing degree of vacuolar fragmentation when grown in media containing 50 or 200 mM CaCl₂. We found that 70-80% of WT cells grown YPD supplemented with 200 mM CaCl₂ contained four or more vacuoles, while 70-80% of cells grown in standard YPD contained three or fewer vacuoles. Furthermore, as the number of the vacuolar structures increased,their size decreased. This vacuolar fragmentation phenotype was independent of osmotic stress, since WT cells grown in YPD medium supplemented with 300 mM NaCl₂ did not exhibit any significant change in vacuolar morphology(data not shown).

As shown above, we found that the pmr1Δ strain contained highly fragmented vacuoles when grown on YPD medium(Fig. 6B). This fragmented vacuole phenotype was even more severe when grown in the presence of 1 mM EGTA, consistent with our finding that this strain undergoes excessive Ca²⁺ stress under these conditions. Growth of the pmr1Δ strain in YPD supplemented with 5-50 mM CaCl₂resulted in a much less severe alteration in vacuolar morphology. Increasing the extracellular CaCl₂ concentration to 200 mM CaCl₂led to a fragmented vacuolar morphology similar to that observed in WT cells under these conditions. Overall, these results indicate that vacuolar fragmentation is observed in both WT and pmr1Δ strains in conjunction with an elevated level of cellular Ca²⁺ stress.

Previous studies have shown that a pmr1Δ strain lacks the ability to properly maintain normal Ca²⁺ and Mn²⁺ levels in compartments of the secretory pathway. This Ca²⁺ transport defect leads to an increased rate of Ca²⁺ uptake and accumulation by a mechanism that has many characteristics of the mammalian CCE response(Csutora et al., 1999; Lapinskas et al., 1995; Locke et al., 2000). In the current study, we examined how a pmr1Δ strain responds to growth in media containing a reduced level of divalent cations. We found that the steady-state level of total cellular Ca²⁺ in the WT strain decreased by twofold when the level of divalent cations in YPD medium was reduced by the addition of 1 mM EGTA. By contrast, we found that the level of total cellular Ca²⁺ increased almost threefold in the pmr1Δ strain when the level of divalent cations in YPD medium was reduced. This substantial increase in Ca²⁺ accumulation was sufficient to activate the calcineurin pathway, which increased the expression and activity of intracellular transporters that sequester Ca²⁺ from the cytosol. Consistent with this interpretation, we found that the steady-state level of PMC1 mRNA increased nearly fivefold in the pmr1Δ strain under these conditions.

A previous study found that PMC1 expression from a multicopy plasmid was capable of suppressing phenotypes associated with the pmr1Δ mutation (Durr et al., 1998). This suggested that Pmc1p contributes to the filling of Ca²⁺ stores within the ER or Golgi apparatus during its transit through the secretory pathway. In support of this hypothesis, a pmr1Δ/pmc1Δ strain was also shown to exhibit a more severe unfolded protein response (UPR) than a pmr1Δstrain, suggesting that the loss of Pmc1p function causes a further depletion of ER Ca²⁺ that inhibits protein folding(Bonilla et al., 2002). It was also shown that the depletion of Golgi Ca²⁺ stores induces cellular Ca²⁺ uptake by a high-affinity Ca²⁺ uptake transporter encoded by the CCH1 and MID1 genes. These results led to the hypothesis that yeast cells possess a mechanism that couples the level of Ca²⁺ in the Golgi apparatus to Ca²⁺ uptake that is analogous to the CCE response in mammals(Locke et al., 2000). This model predicts that a pmr1Δ strain grows poorly in an environment containing a reduced level of divalent cations because a limiting level of extracellular Ca²⁺ (and/or Mn²⁺) further reduces its ability to transport these cations efficiently into the cell so that it can replenish the depleted pool of these ions in the Golgi apparatus.

Our finding that Ca²⁺ accumulation is significantly increased in the pmr1Δ strain when the availability of divalent cations is reduced cannot be explained solely by a CCE-like mechanism(Csutora et al., 1999; Locke et al., 2000), since it is highly unlikely that the Ca²⁺ level in the Golgi apparatus will be further depleted under conditions where the level of total cellular Ca²⁺ is threefold higher. We also found that the pmr1Δ strain exhibited less vacuolar fragmentation in the presence of 5 mM or 50 mM CaCl₂ than either higher or lower concentrations (see Fig. 6),suggesting that a moderate increase in extracellular Ca²⁺ can reduce the rate of cellular uptake (and the resulting level of Ca²⁺stress). When taken together, these results suggest that a mechanism exists that can couple the rate of cellular Ca²⁺ uptake to the extracellular Ca²⁺ concentration. The results of a previous study also support the existence of an extracellular Ca²⁺-sensing mechanism. We have shown that the loss of the major isoform of phosphoglucomutase (encoded by the PGM2 gene) causes a large increase in Ca²⁺ uptake and accumulation when a pgm2Δ strain is grown in YP medium containing galactose as carbon source(Fu et al., 2000). Furthermore, we found that a pgm2Δ/pmr1Δ strain is unable to grow in YP galactose medium, presumably because the combination of these mutations leads to excessive Ca²⁺ uptake and accumulation that results in an inhibition of cell growth. However, we found that growth of the pgm2Δ/pmr1Δ strain could be restored on YP galactose medium when 100 mM CaCl₂ was added to the growth medium(Fu et al., 2000). In light of our current results, we propose that this increase in the concentration of CaCl₂ in the growth medium can restore growth of the pgm2Δ/pmr1Δ strain by reducing Ca²⁺uptake and accumulation through the action of an extracellular Ca²⁺-sensing mechanism.

While several mechanisms could be used to monitor the level of extracellular Ca²⁺, the most straightforward method would utilize a Ca²⁺ sensor on the cell surface(Fig. 7). This cell-surface Ca²⁺ sensor could be functionally equivalent to the extracellular Ca²⁺-sensing receptor of mammalian cells, which can respond to extremely small changes in the free Ca²⁺ concentration in the blood(Brown et al., 1995; Hebert et al., 1997). The results of the current study provide strong evidence that an extracellular Ca²⁺-sensing mechanism can also play an important role in coupling the level of environmental Ca²⁺ to cellular Ca²⁺ uptake,homeostasis and signaling in yeast. To our knowledge, such a mechanism has not been proposed previously for yeast cells. This might be due to the fact that this mechanism works in conjunction with other redundant processes that together tightly control cellular Ca²⁺ homeostasis. Such overlapping mechanisms could explain why yeast mutants that maintain abnormally high levels of Ca²⁺ uptake and accumulation (such as the pmr1Δ and pgm2Δ strains) were necessary to obtain evidence of this novel control mechanism. These mutant strains have shed new light on the mechanisms that couple cellular Ca²⁺ uptake and accumulation. For example, the inability of the pmr1Δstrain to grow in the presence of a reduced level of divalent cations allowed us to show that excessive Ca²⁺ accumulation is responsible for this growth defect. This problem is further exacerbated in the pmr1Δ/pmc1Δ strain in two ways. First, the loss of Pmc1p further aggravates the reduced ability to properly fill the ER store depletion caused by the pmr1Δ mutation. Second, the pmc1Δ mutation diminishes the ability of the cell to sequester excess Ca²⁺ in the vacuole. As a result, the cytosolic Ca²⁺ load becomes more severe in the pmr1Δ/pmc1Δ strain, whereas its ability to sequester excess Ca²⁺ adequately into the vacuole is decreased. Together, these consequences result in an inability to grow in a Ca²⁺-depleted environment.

In order to understand better the functional interplay between Ca²⁺ transporters in yeast, we also examined the role of Vcx1p in maintaining cytosolic Ca²⁺ homeostasis in the pmr1Δ/pmc1Δ strain. We found that Vcx1p plays an important role in Ca²⁺ homeostasis over a much broader range of extracellular Ca²⁺ concentrations than previously appreciated(Pozos et al., 1996). Genetic studies have suggested that Vcx1p might play a less important role in Ca²⁺ homeostasis than the Ca²⁺-ATPases Pmc1p and Pmr1p. Consistent with this conclusion, we have observed that a pmr1Δ/vcx1Δ strain is no more sensitive to chelating agents than a pmr1Δ strain. This result indicated that Pmc1p alone is sufficient to sequester the high cellular Ca²⁺that accumulates under these conditions, and suggested that Vcx1p plays only a secondary role in Ca²⁺ homeostasis (R. Kellermayer and D. Bedwell,unpublished). However, in the current study, we show that a pmr1Δ/pmc1Δ mutant can successfully cope with a considerable level of Ca²⁺ stress, whereas cellular Ca²⁺homeostasis is completely disrupted when its vacuolar proton gradient is disrupted by bafilomycin A₁. Since it has been shown that vacuolar vesicles derived from a strain lacking Vcx1p do not possess any Ca²⁺/H⁺ exchange activity(Pozos et al., 1996), these findings strongly indicate that Vcx1p plays the predominant role in maintaining Ca²⁺ homeostasis in the pmr1Δ/pmc1Δ mutant. In addition, the previous observation that the combination of the pmr1Δ, pmc1Δ and vcx1Δ mutations together causes synthetic lethality reinforces the importance of Vcx1p in Ca²⁺regulation in the absence of these two Ca²⁺-ATPases(Cunningham and Fink, 1996; Miseta et al., 1999a). These results suggest that the Pmr1p, Pmc1p and Vcx1p Ca²⁺ transporters are all capable of independently maintaining cellular Ca²⁺homeostasis in many yeast strains. Apparently, subtle differences in the extent of Vcx1p inhibition by calcineurin are responsible for conflicting reports regarding the viability of pmr1Δ/pmc1Δstrains in different genetic backgrounds(Cunningham and Fink, 1996; Locke et al., 2000) (this report).

We also used a Pmc1p-GFP fusion protein to show that vacuolar fragmentation coincides with Ca²⁺ stress. Our finding that vacuolar fragmentation occurs under diverse conditions that lead to cellular Ca²⁺ stress in both WT and mutant strains strongly suggests a direct relationship between these two events. The data presented does not allow us to determine whether it is high Ca²⁺ in the cytosol or vacuole that leads to this fragmentation phenomenon. However, in other experiments we found that a pmc1Δ/vcx1Δ strain exhibits fragmented vacuoles in media containing 50 mM CaCl₂ in a manner similar to the WT strain (R. Kellermayer and D. Bedwell, unpublished). Since this mutant accumulates much less Ca²⁺ in the vacuole than the WT strain(Cunningham and Fink, 1996; Miseta et al., 1999a; Pozos et al., 1996), these results strongly suggest that an elevated cytosolic Ca²⁺ causes vacuolar fragmentation in S. cerevisiae.

A recent study examined a large collection of knockout strains from the Saccharomyces Genome Deletion Project for defects in homotypic vacuolar fusion (Seeley et al.,2002). Surprisingly, 714 out of 4828 deletion strains examined(∼15%) exhibited alterations in vacuole morphology. After excluding a large number of genes thought to influence vacuole morphology by indirect means, 137 genes (∼3%) were chosen as candidate VAM genes that were thought to play a direct role in vacuolar morphology. Among these were a variety of genes that encoded proteins previously related to homotypic vacuolar fusion, including fusion catalysts, enzymes of lipid metabolism,SNARES, GTPases and their effectors, protein kinases, phosphatases, and cytoskeletal proteins. Another group of genes encoded proteins involved in cation transport (including the PMR1 gene). Because vacuolar fragmentation was associated with the deletion of these genes, it was reasoned that the loss of these gene products caused defects in vacuolar fusion,resulting in the vacuolar fragmentation phenotype. However, on the basis of the strong correlation between cellular Ca²⁺ stress and vacuolar fragmentation we observed in both WT and mutant yeast strains, we propose that an elevation of cytosolic Ca²⁺ might lead to vacuolar fragmentation as a regulatory response to aid in Ca²⁺ sequestration in WT yeast,rather than simply being the result of a defect in vacuolar fusion associated with high cytosolic Ca²⁺. On a per unit volume basis, multiple small vacuoles provide a greater surface area than fewer, larger vacuoles. This increased surface/volume ratio could aid in accommodating the function of the increased number of Pmc1p transporters in the vacuolar membrane that are induced by calcineurin activation, thus allowing vacuolar Ca²⁺sequestration to proceed more efficiently.

In both the current study and a prior study of genes that influence homotypic vacuolar fusion (Seeley et al.,2002), it was shown that a pmr1Δ strain exhibits a vacuolar fragmentation phenotype. By contrast, another study found that a pmr1Δ mutation reversed the vacuolar fragmentation phenotype associated with oxidative stress in a sod1Δ strain(Corson et al., 1999). These results led to the conclusion that oxidative stress associated with the sod1Δ mutation altered cellular iron homeostasis, leading to oxidative damage that somehow led to vacuolar fragmentation. It was proposed that the pmr1Δ mutation suppressed this vacuolar fragmentation phenotype by raising the concentration of Mn²⁺ in the cytosol,which acted to scavenge free radicals and reduce oxidative stress. Thus, pmr1Δ mutations have been associated with both the induction and reversal of vacuolar fragmentation. These markedly different effects suggest the existence of a complex regulatory mechanism that allows the cytosolic levels of different divalent cations to influence vacuolar morphology in distinct ways. Further studies are needed to determine how this complex physiological adaptation is carried out.

The authors thank Lianwu Fu for fruitful discussions, Albert Tousson for providing assistance in digital imaging, and Kyle Cunningham, Kelly Tatchell and Aaron Straight for generously providing strains and other reagents. This work was supported by American Heart Association (Southeast Affiliate) grant 0255121B (to D.M.B.) and Hungarian National Foundation grant OTKA-T038144 (to A.M.).