Molecular cloning of an intracellular P-type ATPase from Dictyostelium that is up-regulated in calcium-adapted cells.

Results from a number of laboratories suggest that intracellular Ca2+ is involved in the regulation of Dictyostelium discoideum growth and development. To learn more about the regulation and function of intracellular Ca2+ in this organism, we have cloned and sequenced cDNAs that encode a putative P-type Ca2+ ATPase designated patA. The deduced protein product of this gene (PAT1) has a calculated molecular mass of 120,718 daltons. It exhibits about 46% amino acid identity with Ca2+ ATPases of the plasma membrane Ca2+ ATPase family and lower identity with sarco(endo)plasmic reticulum Ca2+ ATPase family members and monovalent cation pumps. However, PAT1 lacks the highly conserved calmodulin-binding domain present in the C-terminal region of most plasma membrane Ca2+ ATPase-type enzymes. When Dictyostelium amoebae are adapted to grow in the presence of 80 mM CaCl2, both the patA message and protein product are up-regulated substantially. These cells also exhibit an increase in the rate and magnitude of intracellular P-type Ca2+ uptake activity. Immunofluorescence analysis indicates that PAT1 colocalizes with bound calmodulin to intracellular membranes, probably components of the contractile vacuole complex. The presence of PAT1 on the contractile vacuole suggests that in Dictyostelium this organelle might function in Ca2+ homeostasis as well as in water regulation.

Intracellular calcium functions in the regulation of a wide variety of cellular processes in both higher and lower eukaryotic cells (1,2). Recently, considerable progress has been made in understanding Ca 2ϩ -signaling mechanisms in mammalian cells (3); however, much less is known about the role of this ion in lower organisms.
Calcium seems to be involved in development of the lower eukaryote Dictyostelium discoideum, as suggested by a number of observations. When starved of nutrients, amoebae of this organism synthesize and secrete cAMP in a periodic fashion. The cells respond to the waves of cAMP by differentiating, aggregating into multicellular structures, and eventually forming fruiting bodies comprised of two major cell types, stalk cells and spores (4,5). During early development, treatment of the amoebae with EGTA, a Ca 2ϩ chelator (6), La 3ϩ , a Ca 2ϩ -channel blocker (7), or putative intracellular Ca 2ϩ or calmodulin antagonists (8 -10) inhibits cell differentiation and/or aggregation. In addition, depletion of intracellular Ca 2ϩ stores with EGTA and A23187 reduces the stability and secretion of the enzyme cyclic nucleotide phosphodiesterase and its specific inhibitor (11,12). During late development, Ca 2ϩ accumulates preferentially in prestalk regions of migrating aggregates and influences stalk/ spore cell differentiation (13,14).
These observations indicate that changes in the distribution of intracellular Ca 2ϩ might be important in regulating Dictyostelium development. When the amoebae are shaken in buffer, they accumulate and extrude Ca 2ϩ ions. The rate of Ca 2ϩ uptake is increased dramatically by the activation of cell surface folate and cAMP receptors (7,15) through a process that might be G-protein independent (16). Receptor activation also results in the inositol 1,4,5-trisphosphate-induced release of Ca 2ϩ from a small intracellular Ca 2ϩ pool (17,18). Using a 45 Ca 2ϩ uptake assay and isolated cell fractions, a major inositol 1,4,5-trisphosphate-insensitive intracellular Ca 2ϩ pool has been identified (19,20). This pool is associated with "acidosomal" membranes (21,22), a component of the contractile vacuole complex (23,24). Transport of Ca 2ϩ into this pool occurs via a vanadate-sensitive (19), thapsigargin-insensitive (22) Ca 2ϩ ATPase, and ion movement is reported to be facilitated by a high intravesicular proton concentration (21,22). Using a Ca 2ϩ -sensitive electrode and filipin-permeabilized cells, a small intracellular inositol 1,4,5-trisphosphate-sensitive Ca 2ϩ pool has also been detected. Ca 2ϩ uptake into this pool is vanadate-resistant and might involve a H ϩ /Ca 2ϩ antiport (25). In addition, a Ca 2ϩ -stimulated ATPase associated with plasma membrane fractions has been reported (26); this enzyme might function in Ca 2ϩ efflux. At present, however, little is known about the movement of Ca 2ϩ ions between intracellular pools or the extrusion of Ca 2ϩ from the cells.
One way to dissect complex systems such as this is to clone genes encoding the various pumps/transporters and to use these clones to disrupt the endogenous sequences by homologous recombination or to reduce their expression by antisense RNA strategies. This approach has been very effective at identifying certain Ca 2ϩ -signaling pathways in Saccharomyces cerevisiae (27)(28)(29).
In all eukaryotic cells, Ca 2ϩ -translocating ATPases play a major role in Ca 2ϩ homeostasis (30). These pumps are members of a large family of P-type cation transport ATPases (so named because they form a phosphoenzyme intermediate during the catalytic reaction). All ATPases of this family possess regions of high amino acid sequence homology; however, the different members can be subclassified according to their cellular localization and function, e.g. plasma membrane Ca 2ϩ ATPase (PMCA) 1 and sarco(endo)plasmic reticulum Ca 2ϩ ATPase (SERCA) (31,32). In the present work, we have used the PCR to clone cDNAs of a D. discoideum gene patA (P-type ATPase A) that encode a putative Ca 2ϩ ATPase. Indirect immunofluorescence analysis suggests that the product of this gene (PAT1) is associated with an intracellular organelle, probably the contractile vacuole.

EXPERIMENTAL PROCEDURES
Strains and Culture Conditions-Strain AX2, an axenic derivative of D. discoideum (33), was used throughout this study. In most experiments, amoebae were grown in HL-5 medium (34) supplemented with 200 g of dihydrostreptomycin/ml and shaken at 250 rpm on a New Brunswick gyratory shaker at room temperature until the cell density reached 6 -10 ϫ 10 6 cells/ml. In a few experiments, the amoebae were grown in HL-5 medium buffered at pH 6.6 with 6.1 mM MES-NaOH (MES/HL-5) in place of phosphate and supplemented with 80 mM CaCl 2 or 80 mM MgCl 2 . The cells were adapted to grow in the presence of high calcium by increasing the CaCl 2 concentration of the medium stepwise over a period of 3-4 weeks. In both cases, the cells were harvested by centrifugation at 700 ϫ g for 2 min and washed twice in salt solution (35) before use.
All procedures involving plasmid and phage amplification (36) were performed using Escherichia coli strains DH5␣ and LE392, respectively.
RNA Isolation and Northern Blot Analysis-Total RNA was extracted from dry ice/ethanol frozen cell pellets (2 ϫ 10 7 cells) by the guanidine thiocyanate method as described (37). The samples were enriched for mRNA using the Poly(A)Ttract mRNA Isolation System IV (Promega). Poly(A) ϩ RNA samples (approximately 3 g) were fractionated on 1% agarose, 0.66 M formaldehyde gels and transferred to GeneScreen Plus (Dupont NEN) according to the manufacturer's instructions. The membranes were probed with random primed cDNA fragments. Hybridization and washing conditions were as described (37).
Cloning and Sequencing of patA cDNAs-DNA sequences corresponding to P-type ATPases were identified by "touchdown" PCR amplification (38) of D. discoideum genomic DNA, isolated as described (39) using degenerate oligodeoxyribonucleotides provided by Dr. Kyle Cunningham (The Johns Hopkins University). Primer A1 (5Ј-CGG-GATCCGTNATNTGYWSNGAYAARACNGGNAC-3Ј) was based on the amino acid sequence of the highly conserved phosphorylation site of these enzymes and possessed a BamHI restriction site, while primer B2 (5Ј-CGGAATTCGSRTCRTTNRYNCCRTCNCCNG-3Ј) contained an EcoRI site and corresponded to sequences involved in ATP binding (28,40). Amplification conditions were as described previously (41). PCR products of four distinct sizes were obtained. Fragments of each size were cloned into BamHI/EcoRI-digested pBluescript II KSϩ (Stratagene) and partially sequenced. DNA and deduced amino acid sequence homology searches on the PCR products using the EMBL/GenBank data base revealed one fragment (0.95 kb) with an open reading frame encoding an amino acid sequence with appreciable identity to Ca 2ϩ ATPases. This fragment was used as a probe to screen a 3-h D. discoideum cDNA library constructed in gt11 (a gift of Dr. Peter Devreotes, The Johns Hopkins University School of Medicine).
The cDNA insert in the first clone isolated (cDNA1) was completely sequenced using a combination of exonuclease III/mung bean nucleasegenerated deletions (36) and synthetic oligodeoxyribonucleotides. Double-stranded DNA sequencing was performed using a Sequenase Version 2.0 kit (U. S. Biochemical Corp./Amersham). The deduced open reading frame of cDNA1 encoded a putative protein with significant identity to Ca 2ϩ ATPases, but it lacked both the 5Ј-initiation and 3Ј-termination sequences. A ϳ0.45-kb fragment spanning the 5Ј-end of cDNA1 to a unique internal EcoRV restriction site was used to rescreen the cDNA library, and several additional clones were obtained. Two of the new cDNAs (cDNA2 and cDNA3) were found to encode the missing 3Ј-and 5Ј-ends, respectively. The combined overlapping sequences of cDNAs 1-3 encode the full-length patA cDNA sequence submitted to the GenBank TM /EMBL data bank.
Antiserum Production and Affinity Purification of Antibodies-After drawing preimmune serum, a female New Zealand White rabbit was immunized intramuscularly with approximately 100 g of GST-PAT1 fusion protein every 6 weeks for 6 months. Samples of blood (20 -30 ml) were collected 4 weeks after each injection (43). Aliquoted serum was stored at Ϫ80°C.
A 50-g sample of GST-PAT1 was transferred to nitrocellulose after SDS-PAGE and used as an affinity matrix to purify antibodies specific for the fusion protein (anti-PAT1) as described (36).
Immunostaining Conditions-Vegetative AX2 cells were immunostained with affinity-purified anti-PAT1 and with mouse monoclonal anti-Dictyostelium calmodulin antibodies (2D1, a gift of Dr. Margaret Clarke, Oklahoma Medical Research Foundation). Anti-PAT1 was diluted 1:100, and the 2D1 antibodies were diluted 1:2000 in PBS-2 (130 mM NaCl, 6 mM Na 2 HPO 4 /KH 2 PO 4 , pH 6.9). Secondary antibodies were fluorescein isothiocyanate isomer I-conjugated goat anti-rabbit IgG (1: 20, Caltag) and Texas Red-conjugated goat anti-mouse IgG (1:20, Caltag), both diluted in PBS-2. For double-staining experiments, cells growing in HL-5 were harvested at a density of 2 ϫ 10 6 cells/ml by centrifugation (22°C) at 700 ϫ g for 2 min, washed twice in PBS-2, and then resuspended in PBS-2 at 2 ϫ 10 5 cells/ml. Approximately 2 ϫ 10 5 cells were permitted to attach to glass slides at room temperature (45 min), and then they were rinsed in PBS-2, fixed for 10 min in methanol, 1% formaldehyde, 0.05% Triton X-100 at Ϫ20°C as described (45), except the agar overlay was omitted. After rinsing with PBS-2, the slides were blocked by incubation for 30 min in 3% skim milk in PBS-2 and then stained sequentially for 30 min at room temperature with the 2D1 antibodies, Texas Red-conjugated goat anti-mouse IgG, anti-PAT1, and fluorescein isothiocyanate isomer I-conjugated goat anti-rabbit, washing with PBS-2 for 10 min after each incubation. Some control cells were treated with preimmune primary serum instead of anti-PAT1 or stained singly with anti-PAT1. For other control cells, the primary antibody was omitted.
Immunostained cells were examined using a Bio-Rad MRC-600 confocal microscope with xenon-argon laser attached to a Nikon Optiphot microscope with a 60ϫ objective lens (NA ϭ 1.4) and simultaneous dual channel (split-screen) imaging. For each cell, the background gray levels and gain were adjusted so that the gray scale levels did not saturate the detector (i.e. fell within the range 1-255), and a series of optical sections about 0.7 m apart were recorded. Confocal images were printed using a Mitsubishi thermal video printer; phase-contrast images were photographed directly through the microscope using a 100ϫ NA 1.3 phase contrast objective and a Nikon F3 camera.

RESULTS
Cloning a Dictyostelium PMCA Homolog-Ca 2ϩ ATPases are members of a family of P-type ion pumps that contain ten conserved domains (40). Degenerate oligodeoxyribonucleotides (primers A1 and B2) corresponding to two of these regions, a phosphorylation site and a site involved in ATP binding (28,40), were used in a PCR experiment to amplify specific sequences from D. discoideum and S. cerevisiae genomic DNA. PCR products of 1.2, 0.95, and 0.8 kb were obtained with the S. cerevisiae DNA; these fragments are known to contain sequences from at least five different P-type ion pumps (27,28,46,47). Under the same conditions, Dictyostelium genomic DNA gave four PCR products with sizes of approximately 1.1, 0.95, 0.8, and 0.6 kb. The smallest fragment was shown subsequently to correspond to the Dictyostelium 16S rRNA gene. The three remaining PCR products were similar in size to the S. cerevisiae products, suggesting that the two organisms might possess a similar number of P-type ion pumps.
Partial sequencing of the D. discoideum PCR products identified four unique sequences with relatively high amino acid sequence identity to known P-type ion pumps (Table I). One of these PCR products (0.95 kb) showed appreciable identity to Ca 2ϩ ATPases; therefore, this sequence was used to isolate corresponding cDNAs. Analysis of these cDNAs revealed a complete open reading frame (patA). The proposed ATG start codon in the patA cDNA is in a sequence (AAAATGA), which agrees well with the consensus translation initiation sequence (AXAATGG) of D. discoideum (48). Moreover, this ATG is preceded by an in-frame TAA stop codon 12 bp upstream. The open reading frame terminates after coding for 1115 amino acids, corresponding to a protein with a calculated molecular mass of 120,718 daltons.
The putative protein product of patA (PAT1, Fig. 1) contains the conserved phosphorylation and ATP-binding domains present in all P-type ATPases. It shows highest amino acid sequence identity to PMCA family members and lower identity to the SERCA family of Ca 2ϩ ATPases and monovalent cation pumps (Table II). Hydropathy analysis of the deduced amino acid sequence (not shown) reveals a profile very similar to those of PMCA pumps, except for a deletion in the PAT1 sequence between the second and third transmembrane domains. This is a phospholipid-binding regulatory region that is subject to alternative splicing and is not well conserved among the known PMCA isoforms (32). The vacuolar Ca 2ϩ ATPase from S. cerevisiae (PMC1 and its protein product Pmc1p, Ref. 28), a PMCA family member, also has a large deletion as well as an insertion in this region. Like Pmc1p, PAT1 lacks the identically conserved amino acid sequence associated with calmodulin-binding subdomain A. This sequence, essential for calmodulin binding, is present near the C terminus of all mammalian PMCA isoforms (49) but is absent in the yeast isoform, which is truncated in this region (28) and in PAT1 where sequence conservation is lost (Fig. 2). Calmodulin-binding subdomain B, which is thought to influence the affinity of calmodulin binding to subdomain A (49), also appears to be absent in PAT1. The   A Putative Intracellular Ca 2ϩ ATPase from Dictyostelium amino acid sequence of this subdomain is variable in the different PMCA isoforms, although some conservation is maintained. PROSITE analysis of PAT1 reveals a potential cAMPdependent protein kinase phosphorylation site (Fig. 1, amino  acids 1031-1034), which is also present in certain PMCA isoforms (e.g. human PMCA1a, Ref. 50).
Expression of patA-Northern blot analysis of total RNA revealed that patA mRNA is expressed constitutively at very low levels throughout D. discoideum development (data not shown). When the RNA is enriched for polyadenylated message, one transcript is observed at ϳ4 kb (Fig. 3A, lane 1). Interestingly, when the amoebae are adapted to grow in MES/ HL-5 media supplemented with 80 mM CaCl 2 , the level of the transcript is increased substantially (Fig. 3A, lane 2). In the experiment shown, analysis of the major band on a Packard Instant Imager (Canberra Packard) indicated that the patA transcript is ϳ10-fold higher in the Ca 2ϩ -grown cells. The same concentration of MgCl 2 in the growth medium has no affect on the level of patA mRNA (Fig. 3A, lane 3). For this analysis, the amount of RNA loaded in each lane was normalized by reprobing the membrane with vatP cDNA (Fig. 3B); vatP expression is not affected appreciably by the growth conditions used in this experiment. 2 To detect the PAT1, antibodies were raised against the Cterminal 176 amino acids of the protein fused to GST and affinity purified as described under "Experimental Procedures." Total membrane samples prepared from the cells used in Fig. 3 were subjected to Western analysis using the affinitypurified anti-PAT1. The antibodies detected a single band of approximately 120 kDa, the predicted molecular mass of PAT1 (Fig. 4, lane 1). Membranes from cells grown in the presence of CaCl 2 possess a significantly higher level of PAT1 (Fig. 4, lane  2). This elevated level of PAT1 correlates well with the increased abundance of patA message observed in the Ca 2ϩgrown cells. In contrast, membranes from Mg 2ϩ -grown cells contain levels of PAT1 comparable to cells grown in unsupplemented MES/HL-5 medium (Fig. 4, lane 3).
Ca 2ϩ Accumulation by Filipin-permeabilized Amoebae-To determine if Ca 2ϩ -adapted cells show a corresponding increase in Ca 2ϩ -pumping activity, ATP-dependent Ca 2ϩ uptake was assayed in filipin-permeabilized cells cultured in the different media. As illustrated in Fig. 5, amoebae grown in unsupplemented MES/HL-5 or in MES/HL-5 containing 80 mM MgCl 2 , take up Ca 2ϩ linearly for ϳ6 min, accumulating 6 -7 nmol/mg protein. The kinetics and magnitude of Ca 2ϩ uptake by these cells is very similar to results obtained with AX2 cells grown in regular HL-5 medium (20). 3 In contrast, cells grown in the presence of 80 mM CaCl 2 exhibit a ϳ75% increase in the rate and a ϳ90% increase in the amount of Ca 2ϩ accumulated, when 2 J. Moniakis and Y. Xie, unpublished observation. 3 1 and 4) or in the same medium supplemented with 80 mM CaCl 2 (lanes 2 and 5) or 80 mM MgCl 2 (lanes 3 and 6) were subjected to SDS-PAGE on 7.5% gels and transferred to nitrocellulose membranes. Lanes 1-3 were probed with affinity-purified anti-PAT1 antibodies while lanes 4 -6 were probed with preimmune serum. assayed under the same conditions. Ca 2ϩ uptake by both Ca 2ϩgrown cells and cells cultured in regular HL-5 is inhibited more than 93% by 100 M vanadate (data not shown). Therefore, growth in the presence of Ca 2ϩ appears to increase the activity of one or more intracellular P-type Ca 2ϩ ATPases in these cells.
Localization of PAT1-Indirect immunofluorescence microscopy was used to determine the cellular location of PAT1. Initial staining experiments with affinity-purified anti-PAT1 resulted in patterns very reminiscent of those reported by Zhu and colleagues (51) in their studies on the contractile vacuole complex of D. discoideum using antibodies against calmodulin. These workers had shown earlier (52) that insoluble calmodulin in these cells is highly concentrated on the membranes of this organelle. From viewing doubly stained specimens, using affinity-purified anti-PAT1 and a monoclonal anti-calmodulin antibody (2D1, Ref. 53), we conclude that in all cells examined the two antibodies gave very similar vesicular and punctate staining patterns, as illustrated in Fig. 6. The staining patterns observed were due to the primary antibodies because no staining was seen with preimmune serum or with PBS substituted for the primary antibody. The colocalization of the two antibodies was not due to artefactual "bleed-through" from one channel to the other because singly stained specimens were observed only in one channel. Thus, these results suggest that PAT1 colocalizes with bound calmodulin to the contractile vacuolar membranes.

DISCUSSION
Like other eukaryotic cells, amoebae of D. discoideum should possess a variety of P-type pumps to regulate the distribution and concentration of intracellular cations during growth and development. Using the PCR, we have identified four genes from Dictyostelium that exhibit appreciable identity to P-type pumps from other organisms (Table I). In addition, a number of other cDNAs from this organism have been cloned, which, by virtue of sequence similarity, seem to be related to these ATPases. 4 Therefore, D. discoideum appears to possess at least as many P-type ion pumps as S. cerevisiae, where five have been identified to date (27,28,46,47). In the present study, we have focused our work on one Dictyostelium gene, patA.
Several lines of evidence suggest that patA encodes an intracellular P-type Ca 2ϩ ATPase. First, the amino acid sequence deduced from patA cDNAs contains phosphorylation and ATPbinding motifs conserved in all P-type ATPases as well as a putative cAMP-dependent protein kinase phosphorylation site present in certain PMCA isoforms (Fig. 1). Moreover, amino acid sequence alignment (Table II) and hydropathy analysis of PAT1 suggest that this protein is a member of the PMCA family of Ca 2ϩ ATPases. Second, both patA mRNA (Fig. 3) and PAT1 (Fig. 4) are overexpressed in Ca 2ϩ -adapted cells. To grow amoebae in the presence of relatively high concentrations of Ca 2ϩ , the Ca 2ϩ content of the growth medium must be increased gradually over a period of several weeks. Thus, growing the cells under conditions of Ca 2ϩ stress appears to select variants that overexpress patA. A similar up-regulation of a SERCA Ca 2ϩ ATPase has been observed in Chinese hamster lung fibroblast DC-3F cells during growth in the presence of thapsigargin, an inhibitor of the SERCA family of ATPases (54). Ca 2ϩ -grown Dictyostelium cells, permeabilized with filipin, also exhibit an increase in the rate and magnitude of non-mitochondrial, vanadate-sensitive Ca 2ϩ uptake (Fig. 5). The enhanced ability of these cells to sequester Ca 2ϩ is probably a consequence of the elevated levels of PAT1. Interestingly, Ca 2ϩ accumulation continued after 6 min in Ca 2ϩ -grown cells but leveled off in cells not adapted to 80 mM Ca 2ϩ . This may be due to an enhanced Ca 2ϩ storage capacity in the Ca 2ϩgrown cells. Alternatively, the possibility exists that other intracellular P-type Ca 2ϩ ATPases are also up-regulated in Ca 2ϩstressed cells. Third, cell localization studies on PAT1 using indirect immunofluorescence indicate that this ATPase resides on membranes of the contractile vacuole complex (Fig. 6).
In mammalian cells, PMCA-type Ca 2ϩ ATPases are situated on the plasma membrane, whereas in Dictyostelium, PAT1, a PMCA homolog, appears to be a component of the contractile vacuole. This finding is consistent with the biochemical evidence for intracellular PMCA activity in this organism (19 -22). Recently, intracellular PMCA enzymes have also been identified in other organisms. For example, a gene encoding a vacuolar PMCA-type pump (PMC1) has been cloned from S. cerevisiae (28), and intracellular PMCA-like activities have been 4 C.-H. Siu, personal communication. characterized in plants (55,56), although in these cases the enzymes seem to be associated with the endoplasmic reticulum.
Although PAT1 is a PMCA homolog and it colocalizes in the cells with calmodulin, normally a regulator of PMCA activity, there is no evidence that PAT1 activity is regulated by calmodulin. Sequence alignment analysis of PAT1 with the highly conserved calmodulin-binding domains of plasma membrane PMCA-type ATPases reveals that PAT1 (like Pmc1p of S. cerevisiae) lacks this domain (Fig. 2). It also lacks putative amphiphilic helices in the C-terminal region that have been implicated in calmodulin binding (57). In agreement with the sequence analysis, biochemical characterization of intracellular PMCA-type activity in Dictyostelium cells indicates that the activity is unaffected by calmodulin supplementation or by the addition of calmodulin antagonists (19,20,22). In mammalian cells, however, the activity of PMCA enzymes can often be elevated by limited proteolysis of the C-terminal regulatory domain to levels comparable to Ca 2ϩ /calmodulin activation (32), and this process is thought to be a significant regulatory mechanism in vivo (58,59). If PAT1 is regulated by such a process, it might obscure the calmodulin sensitivity of the enzyme in biochemical assays. Therefore, at this time, we cannot rule out the possibility that calmodulin regulates PAT1 activity by interacting with as yet unidentified sequences on the enzyme.
Contractile vacuoles are morphologically complex organelles found in many freshwater protozoa and amoebas where they are thought to function in osmoregulation (24,51,60). These structures accumulate water and ions by poorly understood mechanisms and discharge their contents outside the cell by fusion with the plasma membrane (23). Based on the properties and intracellular localization of a Ca 2ϩ ATPase in Dictyostelium, Milne and Coukell (20) proposed that extrusion of excess Ca 2ϩ from these cells might be facilitated by the fusion of Ca 2ϩ -sequestering vesicles with the cell membrane. Subsequent studies revealed that the high affinity Ca 2ϩ ATPase associated with these vesicles resides in a buoyant membrane fraction (21,22), probably a component of the contractile vacuole system (23,24). In the present study, we show that the putative Ca 2ϩ ATPase, PAT1, colocalizes with bound calmodulin to membranes of the contractile vacuole and that Ca 2ϩadapted cells overexpress PAT1 and possess elevated intracellular Ca 2ϩ uptake activity. Together, these observations suggest that in D. discoideum the contractile vacuole complex might play an important role in Ca 2ϩ homeostasis as well as in water regulation.