Schistosoma mansoni Ca2+-ATPase SMA2 Restores Viability to Yeast Ca2+-ATPase-deficient Strains and Functions in Calcineurin-mediated Ca2+Tolerance*

The sarco(endo)plasmic reticulum of animal cells contains an ATP-powered Ca2+ pump that belongs to the P-type family of membrane-bound cation-translocating enzymes. InSchistosoma mansoni, the sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA) is encoded by the SMA1 andSMA2 genes. A full-length SMA2 cDNA clone was isolated, sequenced, and expressed into a yeast Ca2+-ATPase-deficient strain requiring plasmid-borne rabbit SERCA1a for viability. The S. mansoniCa2+-ATPase supports growth of mutant cells lacking SERCA1a, indicating functional expression in yeast and a role in calcium sequestration. Subcellular fractionation showed that the SMA2 ATPase is localized in yeast internal membranes. SMA2 expression was found to be associated with thapsigargin-sensitive, Ca2+-dependent ATPase activity. The activity increased 2-fold upon calcineurin inactivation, which correlates within vivo stimulated contribution of SMA2 in calcium tolerance. These results suggest that calcineurin controls calcium homeostasis by inhibiting Ca2+-ATPase activity in an internal compartment.

Schistosoma mansoni is a human parasite trematode causing the chronic debilating disease schistosomiasis, which affects hundreds of millions of people over the world. Many different aspect of the parasite's life cycle, such as penetration through the host's skin, locomotion, feeding, and eggshell formation, are controlled by calcium (1)(2)(3).
In animal cells, calcium is stored principally within the sarco(endo)plasmic reticulum, which acts as a storage site for subsequent calcium release in response to environmental signals and as a sink for maintaining cytoplasmic calcium concentrations at submicromolar level (4). The sarco(endo)plasmic reticulum compartment is loaded by an ATP-powered Ca 2ϩ pump (5,6), a member of the P-ATPase family, which is characterized by the formation of a phosphoenzyme catalytic intermediate during ATP hydrolysis (7). In animal cells, the subfamily of sarco(endo)plasmic reticulum Ca 2ϩ -ATPases (SERCAs) 1 comprises six isoforms, which are encoded by three separate genes and each of which has a distinct pattern of tissue-specific and developmentally regulated expression (8,9).
In S. mansoni, there is evidence of at least two SERCA-like pumps, SMA1 and SMA2, in addition to a homolog of the yeast Golgi PMR1 Ca 2ϩ -ATPase (10) and the rat secretory pathway SPCA Ca 2ϩ -ATPase (11) as suggested by PCR analysis (12). Ca 2ϩ -stimulated, Mg 2ϩ -dependent ATPase activity has been found in S. mansoni tissue homogenates (13) and microsomal fractions (14), which is coupled with an active transport of calcium (15). However, a correlation between this ATPase activity and specific Ca 2ϩ -ATPase isoforms has not been established.
In this study, we determined the amino acid sequence of the S. mansoni SMA2 Ca 2ϩ -ATPase. We used a yeast expression system to determine whether Ca 2ϩ -dependent ATPase activity was associated with the expression of SMA2. The results indicate that the S. mansoni Ca 2ϩ pump is present in yeast intracellular membranes and shows kinetic properties and inhibitor sensitivities comparable to those of other SERCA isoforms. In Saccharomyces cerevisiae, the PMR1 and PMC1 genes encode for Ca 2ϩ -ATPases located in the Golgi system and the vacuole, respectively (16,17). A yeast strain lacking both Ca 2ϩ -ATPase functions is not viable (17). However, inactivation of calcineurin, a Ca 2ϩ /calmodulin-activated phosphoprotein phosphatase (18,19) restores viability to the Ca 2ϩ -ATPase-deficient strain, through post-transcriptional activation of the low affinity vacuolar H ϩ /Ca 2ϩ exchanger, VCX1/HUM1 (20,21). We found that the S. mansoni SMA2 substitutes for loss of both PMR1 and PMC1 Ca 2ϩ -ATPases and restores the calcineurindependent Ca 2ϩ tolerance of vacuolar pmc1 mutants. The physiological relevance of these results for the regulation of calcium homeostasis in yeast is discussed.

Strains, Media, and Growth Conditions-Adult worms of the Puerto
Rican strain of S. mansoni were recovered from infected hamsters by portal venous perfusion as described previously (22).
The S. cerevisiae strains used in this study are listed in Table I. The Ca 2ϩ -ATPase mutant strains K605 (pmc1⌬), and K616 (pmc1⌬ pmr1⌬ cnb1⌬) are isogenic to the parental strain W303-1A and were isolated by Cunningham and Fink (17). The MG10 pmc1⌬ pmr1⌬ double mutant expressing SMA2 from the LEU2-marked p315SMA2 plasmid (see below) is derived from the MGY10 -6B strain. 2 MGY10 -6B contains deletions of the PMR1 and PMC1 genes and requires rabbit SERCA1a * This work was supported by the Belgian National Funding for Scientific Research and the Interuniversity Poles of Attraction Program-Belgian State, Prime Minister's Office-Federal Office for Scientific, Technical and Cultural Affairs. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence (s)  on a Ura3-marked plasmid for viability. The MG10 strain containing only p315SMA2 was selected on 5Ј-FOA-containing medium that selects for loss of the plasmid-borne SERCA1a. The MGY8 -3D pmc1⌬ cnb1⌬ strain was selected as a Leu ϩ Trp ϩ His Ϫ meiotic segregant of a diploid strain that was derived from a standard genetic cross between K605 and K616. The vcx1::Kan R null allele was transformed in the MG10 and MGY8 -3D strains to give MM10DV and MGY8DV, respectively. Geneticin-resistant transformants were isolated and correct integration of the Kan R marker at the VCX1 target site was confirmed by PCR amplification, using two oligonucleotides (PNVCX1K1 and PNVCX1K2) that are complementary to the 5Ј-and 3Ј-flanking regions of VCX1 gene and oligonucleotide (PKAN) which is located within the Kan R disruption cassette.
For Ca 2ϩ tolerance tests, yeast cells were grown in minimal SD medium to late stationary phase and diluted in water to a final concentration of 40 ϫ 10 6 cells/ml. Serial 10-fold dilutions were spotted onto YPD and SD media supplemented with 200 and 400 mM CaCl 2 and 100 g/ml cyclosporin A when indicated. Plates were incubated for 3-5 days at 30°C.
Isolation of a Full-length SMA2 cDNA Clone from S. mansoni-An adult worm cDNA library constructed in gt10 was screened by hybridization to the 0.86-kb PCR product SmII (12). Five positive phage plaques were isolated and their cDNA insert subcloned as an EcoRI fragment into the pBluescript SK(ϩ) (Stratagene) plasmid. The nucleotide sequence of the insert from plasmids 2A1 and 1A1 was determined by the dideoxynucleotide method (24), using synthetic oligonucleotides and a set of enzymatically deleted cDNA fragments. Sequence analysis of the encoded polypeptide showed that an approximately 0.6-kb fragment of the complete SMA2 cDNA was lacking from the 5Ј-terminus. The cDNA library was therefore rescreened using the 1-kb EcoRI-BamHI fragment of the 2A1 clone as a radiolabeled probe. Candidates were analyzed by a PCR-based method (25) using the following oligonucleotides: 5041 (5Ј-TGGTCGCTTAGTTTTACCGTTTTC-3Ј), which is complementary to the 3Ј end of the gt10 left arm, 5042 (5Ј-GAGGTG-GCTTATGAGTATTTC-3Ј), which is complementary to the 5Ј end of the gt10 right arm, and 2A1P (5Ј-TTTTTAGCAGCCATTCT-3Ј), which is located 200 bp downstream of the 5Ј end of 2A1. Only one selected clone (3B1) contained a 3370-bp cDNA sequence identical to that of the 2A1 and 1A1 clones, except for a missing deoxythymidylic acid nucleotide at position 1606 in 3B1.
Plasmid Constructions-A full-length SMA2 cDNA (cSMA2) was reconstructed by standard cloning techniques. First, the BamHI-BamHI fragment of plasmid 2A1 (the latter site from vector sequence) and the BamHI-EcoRI fragment of plasmid 1A1 were ligated together into pSK(ϩ) as a BamHI-EcoRI fragment, yielding plasmid 2A1ABE. Then, a 1.2-kb fragment containing a 5Ј-engineered PstI restriction site close to the start AUG codon of SMA2 as well as an deoxyadenylic acid nucleotide at position Ϫ3 relative to the AUG codon was PCR-amplified using clone 3B1 as a template and oligonucleotides 5ATGPCR (5Ј-GCCACTGCAGCAAAATGGAAACAGCATTTTCG-3Ј) and 2A1OL1 (5Ј-AAGTGATGCACTTCTGG-3Ј). The PCR product was cut with PstI and HaeII and added to the HaeII-HindIII fragment of the 2A1ABE plasmid, and both fragments were cloned together into PstI-plus-HindIIIcut pSK(ϩ), yielding plasmid pSKSMA2. The 3.2-kb PstI-HindIII fragment of pSKSMA2 plasmid that contains cSMA2 was used to replace the plant PMA2 cDNA cloned under the control of the promoter of the PMA1 gene (P PMA1 ) in the low copy number (Cen) LEU2-marked cp(PMA1)pma2 plasmid (26). The resulting SMA2 expression plasmid was named p315SMA2. Plasmid p426SMA2 is the high copy number (2), URA3-marked pRS426 vector (27) containing the P PMA1 ::cSMA2 insert from p315SMA2. The p315SMA2 and p426SMA2 plasmids lacking cSMA2 are referred to as p315C and p426C, respectively.
To construct the vcx1::Kan R null allele, the Kan R gene from plasmid pFA6a-kanMX6 (28) was amplified by PCR, using two 60-mer oligonucleotides that correspond to the 5Ј-and 3Ј-flanking regions of the VCX1 gene, respectively (positions Ϫ49 and ϩ1223) (20).
Other DNA Manipulations-Yeast genomic DNA preparation and plasmid and phage DNA purification were carried out as described (29,30). PCR reactions were performed using the TaqI (Amersham Pharmacia Biotech) and PwoI (Boehringer Mannheim) DNA polymerases. All PCR products and DNA junctions obtained through successive cloning steps were sequenced to rule out unexpected mutations.
Northern Blot Analysis-Total RNA from yeast cells and S. mansoni adults worms was prepared following the procedures described in Refs. 31 and 32. Total RNA was resolved by formaldehyde-agarose gel electrophoresis, transferred to a N ϩ -Hybond (Amersham Pharmacia Biotech) nylon membrane by capillary transfer (32) and hybridized to cSMA2 that had been labeled by nick translation as described by the supplier (Life Technologies, Inc.). Equal amount of total RNA was loaded onto the gel as indicated by methylene blue staining of ribosomal RNAs.
Preparation of Yeast Membranes-The various membrane fractions analyzed in this study were prepared by differential centrifugation as described in Ref. 33 with some modifications. Yeast cells were grown to the middle exponential (40 -50 ϫ 10 6 cells/ml) phase in 1000 ml of minimal medium at 30°C. The cells (10 g) were washed twice with cold water and then resuspended in 15 ml of 250 mM sorbitol, 1 mM phenylmethylsulfonyl fluoride, 1 mM MgCl 2 , 10 mM imidazole, 5 mM dithiothreitol, and 2 g/ml protease inhibitors. After cell disruption with glass beads, lysates were centrifuged twice at 1500 ϫ g for 5 min to remove cell debris and then at 5000 ϫ g for 5 min to give the P 5/5 pellet. The S 5/5 supernatant fraction was then centrifuged at 15,000 ϫ g for 40 min to obtain a crude membrane fraction, called P 15/40 , and a supernatant fraction that was centrifuged at 100,000 ϫ g for 1 h to yield the high speed membrane fraction P 100/60 . In some cases, the 15,000 ϫ g/40-min centrifugation step was skipped to recover all membrane proteins, with only a slight decrease in specific SMA2 ATPase activity. The P tot pellet fraction, which was obtained after centrifugation of the S 5/5 supernatant at 100,000 ϫ g for 1 h, consisted of intracellular and plasma membranes as well as polysomes. Membrane pellets were resuspended in 1 mM MgCl 2 , 10 mM imidazole, pH 7.5, and stored at Ϫ70°C.
For subcellular fractionation experiments, P tot membranes (10 mg of protein) were loaded onto a linear 20 -60% sucrose gradient and then centrifuged at 100,000 ϫ g for 16 h at 4°C. Gradients were fractionated from the top to bottom, manually, in 1-ml aliquots (fractions 1-15) that were stored at Ϫ70°C.
Protein concentration was determined by the Lowry (34) or BCA (Sigma procedure, TPRO562) method, using bovine serum albumin as a standard.
ATPase Activity Assay-SMA2-dependent ATPase activity was assayed in 10 mM HEPES/PIPES, pH 7.2, 100 mM KCl, 2.5 mM MgCl 2 , 1 mM MgATP, 100 M CaCl 2 , 2 mM phosphoenolpyruvate, 10 g/ml pyruvate kinase, 50 mM KNO 3 and 2.5 mM NaN 3 . Samples were pre-incubated 5 min at 30°C before the addition of 5 l of 25 mM MgATP, pH 7.0. After 2, 4, 6, and 8 min at 30°C, 20-l aliquots were withdrawn and the reaction was stopped by 0.32% (w/v) NH 4 -molybdate in 1.14 N H 2 SO 4 . Inorganic phosphate was measured at 630 nm using the method described in Ref. 38. SMA2-dependent ATPase activity was calculated from the difference between the slopes obtained in the absence or presence of 40 M thapsigargin. Stock solution of thapsigargin was prepared in dimethyl sulfoxide, which at the highest concentrations used (4.8% (v/v) in the assay) had no effect on ATPase activity. For measurements of apparent K m for MgATP and V max , ATPase activity was determined in 50 mM MES/Tris, pH 7.0, using an ATP regenerating system and MgATP concentrations ranging from 0 to 3.5 mM. Free Mg 2ϩ concentrations were kept at 2.5 mM as described (39). In experiments involving inhibitors (vanadate, erythrosine B, cyclopiazonic acid), ATP concentration was kept constant at 1 mM MgATP. Measurements of pH optimum were made in 50 mM MES, 50 mM MOPS, 50 mM Tris, 1 mM MgATP, 2.5 mM sodium azide, 50 mM KNO 3 , adjusted to the indicated pH with HCl or KOH. To determine the effects of calcium concentrations on ATPase activity, the reaction mixture also contained EGTA. Free Ca 2ϩ concentrations was calculated from stability constants for CaEGTA and MgEGTA complexes (40,41).
The yeast plasma membrane H ϩ -ATPase activity was measured at pH 6.0 (33) after a pre-incubation of 5 min at 30°C in the absence of ATP. Mitochondrial ATPase activity was assayed in 25 mM Tris/HCl, pH 9.0, 500 M vanadate, 50 mM KNO 3 , in the absence or presence of 10 mM sodium azide.
Computer Analysis-The nucleotide sequence was edited using the DNASTAR Software (Lasergene). Homology searches were done using the BLAST program (42) implemented at NCBI (National Center for Biotechnology Information). Amino acid alignments and identity levels were performed using the Pile Up and Bestfit programs (GCG software version 8.0, Genetics Computer Group) available on the Belgian EMBnet Node (BEN).

RESULTS
Isolation of a Complete SMA2 cDNA Clone-A cDNA library of S. mansoni adult worms was screened by DNA hybridization to the PCR product SmII (12) which codes for an internal sequence of SMA2 (for S. mansoni ATPase 2). DNA sequencing analysis of independently isolated clones (clones 2A1 and 1A1 in Fig. 1A) revealed a 5Ј-truncated open reading frame highly homologous to rabbit SERCA isoforms. The 1-kb EcoRI-BamHI fragment of clone 2A1 was then used to reprobe the S. mansoni cDNA library (see "Experimental Procedures"), resulting in the isolation of clone 3B1 that contains the entire SMA2 coding region (Fig. 1A).
Northern Blot Analysis of S. mansoni SMA2-The full-length SMA2 cDNA contains an open reading frame of 3033 bp. The 5Ј-and 3Ј-untranslated regions are 126 and 989 bp long, respectively. No poly(A) addition site could be identified in the 3Ј-untranslated region. Northen blot analysis of total RNA extracted from S. mansoni adult worms revealed a unique 4.3-kb band (Fig. 1B, lane 1), consistent with the size predicted from the sequence of the isolated SMA2 cDNA clones. The SMA2 transcript seems not to be subject to alternative splicing control.
Sequence Analysis of the SMA2 ATPase-The SMA2 cDNA encodes a 111-kDa polypeptide which shares 70% overall identity with the S. mansoni SMA1 isoform (12). Most amino acid differences are located in the N-and C-terminal regions as well as the region flanking the fluorescein isothiocyanate binding site. Regions typically conserved in the P-type family of transport ATPases (8) are found in SMA2 (Fig. 2). These encompass the hydrophilic head protruding into the cytosol and containing the phosphorylation and ATP-binding domains (43) as well as the fluorescein isothiocyanate binding site (44), the smaller cytosolic loop located in the N-terminal region, the hinge motif (45), and the membrane-spanning segments M1-10 (46).
The M4 -6 and M8 transmembrane segments contain amino acids crucial for high affinity Ca 2ϩ binding (47) (double underlined residues in Fig. 2). A close examination of these regions revealed that SMA2 is more related to Ca 2ϩ -ATPases than other P-type ATPases. In particular, the Glu-771 residue in the M5 region of SERCA1a, which is only found in SERCA isoforms, is also conserved in SMA2. The M1 transmembrane segment of SMA2 contains the Lys-Ile-Leu-Leu-Met pentapeptide (residues 63-67), which is identical to the putative sarco/ endoplasmic reticulum retention signal (Lys/Arg-Ile-Leu-Leu-Leu) found in SERCA ATPases (48). This region has been shown to target chimeras of SERCA and plasma membrane Ca 2ϩ pumps to the endoplasmic reticulum of animal Cos cells (49). In common with the non-muscle SERCA3 isoform, SMA2 lacks the phospholamban-binding region (50). The C terminus of SMA2 also lacks the calmodulin-binding domain found in plasma membrane Ca 2ϩ -ATPases (51).
Functional Expression of S. mansoni SMA2 in Yeast-We used a yeast expression system to determine whether SMA2dependent ATPase activity could be detected. The SMA2 cDNA was expressed in S. cerevisiae from the strong and constitutive promoter of the plasma membrane H ϩ -ATPase (PMA1) gene (55,56), cloned in the high copy (2) p426SMA2 or low copy (Cen) p315SMA2 plasmids (see "Experimental Procedures"). Expression of SMA2 was tolerated in the YPH500 wild-type strain although the doubling time increased from 1.5 h (control cells transformed with an empty plasmid) to 2 h (cells with p315SMA2) and 3 h (cells with p426SMA2). Electron microscopic analysis of SMA2-expressing cells showed no accumulation of karmellae or other internal membranes (data not shown) in contrast to the overexpression of yeast hydroxymethylglutaryl-CoA reductase (57), plant plasma membrane H ϩ -ATPase (26), or rabbit SERCA1a. 2 In S. cerevisiae, the PMC1 and PMR1 Ca 2ϩ -ATPases function together in Ca 2ϩ sequestration. They transport Ca 2ϩ from the cytosol into the vacuole and the Golgi, respectively. Mutant strains lacking both Ca 2ϩ -ATPase functions are not viable on standard rich YPD medium, which contains low (0.18 mM) Ca 2ϩ (17). Synthetic lethality of pmr1⌬ pmc1⌬ double mutants is prevented by inactivation of calcineurin (17), a Ca 2ϩ /calmodulindependent protein phosphatase (18,19) that inhibits vacuolar VCX1-dependent H ϩ /Ca 2ϩ exchange. In the absence of calcineurin, VCX1 functions in Ca 2ϩ sequestration much more efficiently and substitutes for loss of PMR1 and PMC1 (Ref. 20, and strain K616 in Fig. 3).
We showed that expression of SMA2 from the p315SMA2 plasmid restores cell viability to the MG10 pmr1⌬ pmc1⌬ mutant strain (Fig. 3). In this functional assay, SMA2 was able to replace plasmid-borne rabbit SERCA1a which is required for growth of pmr1⌬ pmc1⌬ mutants. 2 In separate control experiments, SERCA1a could not be replaced by the empty p315C vector as indicated by the inability of the transformed cells to  (17); PMCA1, human plasma membrane Ca 2ϩ -ATPase isoform 1b from erythrocyte (54). The sequences are aligned over seven chosen regions, which encompass the conserved phosphorylation and ATP-binding regions, the hinge motif, and the M1, M4 -6, and M8 transmembrane segments. Amino acids that form the high affinity Ca 2ϩ binding site are double underlined. Dots indicate residues identical to those of SMA2. Gaps in the alignments are represented by dashes. Numbers on the right of the sequences refer to amino acid positions. Sequences were aligned using the Pileup program (GCG software version 8.0, Genetics Computer Group) followed by manual re-adjustment. grow on 5-FOA medium that selects against the URA3-marked, SERCA1a-containing plasmid (see "Experimental Procedures"). DNA sequencing of p315SMA2 recovered from the MG10 cells revealed no unexpected mutations in the SMA2 coding region and the PMA1 promoter. We can also rule out the possibility that viability of the MG10 strain resulted from activation of VCX1, through a mutation in calcineurin B regulatory subunit (CNB1). The MG10 strain, unlike the K616 pmr1⌬ pmc1⌬ cnb1⌬ triple mutant failed to grow in the presence of 200 mM CaCl 2 (Fig. 3). Moreover, we showed that deletion of VCX1 was not lethal in MG10DV cells expressing SMA2 (see Fig. 6).
Localization of SMA2 in Yeast Internal Membranes-Cell lysates were prepared from the wild-type YPH500 strain transformed with low copy p315SMA2 or high copy p426SMA2 plasmids. Fractionated membrane fractions were isolated by differential centrifugation and assayed for the presence of SMA2 protein by immunoblotting, using antibodies directed against rabbit SERCA1a Ca 2ϩ -ATPase (see "Experimental Procedures"). Cross-reaction was expected from the high sequence identity shared by the two Ca 2ϩ pumps. The antibodies detected a 111-kDa polypeptide, which was not detected in control wild-type membranes lacking SMA2 (Fig. 4A). The apparent mobility of SMA2 in SDS-polyacrylamide gel agrees well with the size predicted from its amino acid sequence. SMA2 is mainly (92%) found in the crude membrane fraction P 15/40 . However, small amounts of SMA2 were also found in the microsomal fraction P 100/60 . The abundance of SMA2 in P tot membranes was estimated to be 4.6 g/mg of total protein. This value is probably an underestimation as rabbit SERCA1a was used for the calibration curve. Higher levels of expression were obtained with the low copy p315SMA2 plasmid than with the high copy p426SMA2 (compare lanes 4 and 5 in Fig. 4A), consistent with Northern blot analysis of SMA2 transcript levels (Fig. 1B, lanes 3 and 2).
To determine the subcellular localization of SMA2, total membranes consisting of P 15/40 and P 100/60 membranes were prepared from YPH500 wild-type cells bearing p315SMA2 and fractionated by sucrose gradient density centrifugation. Individual fractions were assayed for the distribution of proteins typical of a particular organelle, by immunoblotting or by measuring marker enzyme activity (see "Experimental procedures"). Fig. 4 (B and C) shows that the marker enzyme for plasma membranes (PMA1) migrated to a density peaking at 45% sucrose and was clearly separated from mitochondrial ATPase and other internal membranes, peaking at sucrose concentrations of 28% and 42%, respectively. The SMA2 protein fractionates predominantly with the endoplasmic reticulum marker, SEC63 (Fig. 4C). The fractions were also assayed for SMA2-dependent thapsigargin-sensitive ATPase activity (see below). Nearly all thapsigargin-sensitive ATPase activity was found in SEC63-enriched internal membranes (fractions 7-9 in Fig. 4B). This ATPase activity was not detected in cells lacking the SMA2 expression plasmid. Subcellular localization of SMA2 was also analyzed by electron microscopy. Sections of SMA2-expressing cells were incubated in the anti-SERCA1a antiserum followed by incubation in gold-conjugated protein A. However, no specific immunolabeling was detected under these conditions (data not shown).
Thapsigargin-sensitive ATPase Activity Is Associated with SMA2 Expression-SMA2-dependent ATPase activity was measured in solubilized membrane fractions obtained after gel filtration chromatography instead of the sucrose gradient fractions, as the latter showed significant contamination by PMA1 (Fig. 4C).
Yeast P tot membranes (4 mg/ml) were solubilized with n-dodecyl-␤-D-maltoside in a ratio of 0.8 mg of detergent/mg of membrane protein. These conditions, which solubilized 50% of total membrane protein, enabled the complete recovery of SMA2 activity in the supernatant fraction (see "Experimental Procedures"). Addition of 0.1% (w/w) asolectin to the solubilization medium resulted in a 2-fold increase in ATPase activity and in a better SMA2 stabilization at either 4°C or 30°C. Among the other detergents tested, Triton X-100, CHAPS, and n-octyl-␤-D-glucopyranoside solubilized less than 20% of membrane-bound SMA2 activity, whereas octaethylene glycol mono-n-dodecyl ether (C 12 E 8 ), lysolecithin, and sodium deoxycholate inactivated the ATPase activity (data not shown).
Solubilized membranes were fractionated by gel filtration chromatography, and fractions were assayed for the presence of SMA2 by immunoblotting and measuring thapsigargin-sensitive ATPase activity (see "Experimental Procedures"). Thapsigargin is a plant sesquiterpene lactone, which is a specific inhibitor of SERCAs (58,59). Fig. 5 shows elution of SMA2 in the front of the bulk of the total protein (fractions 14 -20). There is a one-fraction shift between the fractions enriched by SMA2 (fractions [15][16][17] and the peak of thapsigargin-sensitive ATPase activity (fractions 14 and 15). This result would indicate partial SMA2 inactivation in those fractions where the detergent/protein ratio is no longer optimal, consistent with the observation that 75% of total SMA2 activity was lost after gel filtration chromatography. Gel filtration fractions 14 and 15 were combined to analyze kinetic properties of SMA2. An 8-fold purification and a yield of 11.5% of ATPase activity was achieved in these pooled fractions compared with total membranes. The amount of PMA1 protein in the pooled fractions represented less than 6% (data not shown).
Kinetic Properties of SMA2 ATPase Activity-Thapsigarginsensitive ATPase activity of solubilized membrane fractions enriched with SMA2 was measured in the presence of 100 M free Ca 2ϩ , 2.5 mM free Mg 2ϩ (shown to be required for maximal stimulation of the activity), an ATP regenerating system, and inhibitors of pyrophosphatases, mitochondrial, and vacuolar ATPases (see "Experimental Procedures"). Under these conditions, thapsigargin-sensitive ATPase activity represented up to 75% of total ATPase activity.
In Eadie-Hofstee plots, SMA2 activity shows an apparent K m for MgATP of 0.25 mM and a V max of 0.11 mol P i ⅐min Ϫ1 ⅐mg Ϫ1 (data not shown). An optimum pH of 7-7.4 was detected (data not shown), which is similar to the optimum pH of the SERCA3 isoform (60). For measurement of Ca 2ϩ dependence, free Ca 2ϩ was varied in the reaction mix using EGTA as described under "Experimental Procedures." An assay of ATPase activity as a function of calcium concentration revealed a maximum at 10 M free Ca 2ϩ , an apparent K 1/2 (Ca 2ϩ concentration yielding half-maximal activation) of 1 M free Ca 2ϩ (Fig. 5C). This result is very close to that of SERCA1a (60). In a Hill plot representation, the best linear fit of the data indicated a Hill coefficient of 1.29, consistent with the presence of two calcium binding sites showing positive cooperativity.
Inhibitor sensitivities of SMA2 expressed in solubilized yeast membranes are summarized in Table II. Vanadate is a characteristic inhibitor of P-type ATPases. SMA2 activity was completely inhibited at vanadate concentrations above 40 M. SERCA ATPases are specifically inhibited by nanomolar concentrations of thapsigargin (59) and cyclopiazonic acid (61,62). In comparison with other SERCA isoforms, the SMA2 enzyme appears less sensitive to inhibition by thapsigargin (IC 50 of 18 M) and cyclopiazonic acid (IC 50 of 8 M). This relative inhibitor insensitivity probably results from membrane solubilization as ATPase activity was completely inhibited by 5 M thapsigargin in crude yeast membranes. Erythrosine B has been shown to inhibit ATP binding of the yeast plasma membrane H ϩ -ATPase, PMA1 (63). We found that SMA2 was likewise sensitive to this inhibitor, with an IC 50 of 3 M. Praziquantel is currently the drug of choice for the treatment of human schistosomiasis (64), causing rapid damages to the parasite's tegument and a subsequent increase in the intracellular pool of free Ca 2ϩ (65)(66)(67). However, praziquantel had no effect on SMA2 activity at concentrations up to 300 M (data not shown), consistent with the observation that Ca 2ϩ -dependent ATPase activity in S. mansoni membrane fractions is not affected by praziquantel (68).
The growth defects of pmc1⌬ mutants in high Ca 2ϩ media were not suppressed by the expression of SMA2, consistent with the inability of SMA2 to support cell growth of the MG10 pmr1⌬ pmc1⌬ mutant on media containing 200 mM CaCl 2 (Figs. 3 and 6A, lines 1 and 2). Interestingly, the role of SMA2 in Ca 2ϩ tolerance was increased after calcineurin inactivation, through VCX1/HUM1 independent mechanisms. We first constructed the MGY8 -3D pmc1⌬ mutant strain which also con- FIG. 4. Localization of SMA2 ATPase in yeast internal membranes. A, distribution of SMA2 in membranes fractionated by differential centrifugation. Cell homogenates were prepared from YPH500 cells expressing SMA2 from low copy p315SMA2 (Cen-SMA2) or high copy p426SMA2 (2-SMA2). After centrifugation at 5000 ϫ g for 5 min, the P 5/5 pellet fraction was discarded and the S 5/5 supernatant was then centrifuged at 15,000 ϫ g for 40 min to yield the P 15/40 pellet fraction. The S 15/40 supernatant was withdrawn and centrifuged at 100,000 ϫ g for 60 min, yielding P 100/60 and S 100/60 fractions. Yeast fractions (10 g) and rabbit sarcoplasmic reticulum (SR) membranes (100 ng) were analyzed by immunoblotting using polyclonal antibodies against rabbit SERCA1a. Wild-type P 15/40 membranes lacking SMA2 are shown as a control in lane 1. B, fractionation of total yeast membranes by sucrose density gradient centrifugation. Total membranes (P tot fraction) from YPH500 cells bearing p315SMA2 were loaded on a 20 -60% sucrose gradient and fractions were collected from the top (fraction 1) to the bottom (fraction 15) of the gradient. The fractions were assayed for thapsigargin-sensitive ATPase activity (SMA2), plasma membrane H ϩ -ATPase activity (PMA1), and mitochondrial ATPase activity (F1/FO). The activity was expressed as a percentage of total activity summed over the gradient. C, Western blot analysis of SMA2, PMA1 (plasma membrane marker), and SEC63 (endoplasmic reticulum marker). tained a null mutation of the CNB1 gene encoding the regulatory subunit of calcineurin (19). As the contribution by VCX1/ HUM1 to Ca 2ϩ tolerance is high in calcineurin-deficient strains (Refs. 20 and 21; Fig. 6A, line 3), we also constructed the MGY8DV pmc1⌬ cnb1⌬ vcx1⌬ triple mutant, which is identical to MGY8 -3D, except for the deletion of VCX1/HUM1. MGY8DV cells carrying the SMA2 expression plasmid, p426SMA2, were able to grow at 200 mM CaCl 2 (line 5) in contrast to the cells lacking p426SMA2 (line 4).
In yeast, calcineurin activity is inhibited by the immunosup-pressive drug cyclosporin A, in association with the cyclophilin A homolog CPR1 (69). The addition of cyclosporin A prevented the inhibitory effects of high Ca 2ϩ on the growth of pmc1⌬ mutants (Fig. 6B, left panel), consistent with the activation of VCX1 after calcineurin inhibition (17). We also found that SMA2 was able to support growth of the MM10DV pmr1⌬ pmc1⌬ vcx1⌬ triple mutant on high Ca 2ϩ media when supplemented with cyclosporin A (Fig. 6B, right panel). Therefore, it seems likely that the function of the SMA2 Ca 2ϩ -ATPase in yeast is inhibited through a calcineurin-mediated regulatory event.
Calcineurin-dependent Stimulation of SMA2 Activity-To determine whether calcineurin regulates, either directly or indirectly, SMA2 activity, total membrane fractions were prepared from the pmc1⌬ (K605), pmc1⌬ cnb1⌬ (MGY8 -3D), and pmc1⌬ cnb1⌬ vcx1⌬ (MGY8DV) strains carrying the SMA2 expression plasmid and assayed for thapsigargin-sensitive ATPase activity (Fig. 7). The ATPase activity of MGY8DV membranes showed a 2-fold increase (21.6 nmol of P i ⅐min Ϫ1 ⅐mg Ϫ1 ) in comparison with that of K605 membranes (10.7 nmol of P i ⅐min Ϫ1 ⅐mg Ϫ1 ) and was significantly higher than the activity of MGY8 -3D membranes (17.5 nmol of P i ⅐min Ϫ1 ⅐mg Ϫ1 ). These findings indicate that inactivation of calcineurin results in stimulated SMA2 activity, consistent with increased contribution to Ca 2ϩ tolerance in vivo.

DISCUSSION
In this report, we have described the biochemical and functional properties of S. mansoni SMA2, a new sarco(endo)plasmic reticulum Ca 2ϩ -ATPase isoform. SMA2 was able to support growth of yeast Ca 2ϩ -ATPase-deficient strains. The expression of SMA2 was associated with thapsigargin-sensitive, Ca 2ϩ -dependent ATPase activity in intracellular membranes. This ATPase activity showed kinetic properties and inhibitor sensitivity comparable to that of mammalian SERCA isoforms. We also found that SMA2 contributes to Ca 2ϩ tolerance in yeast strains lacking calcineurin activity through mutations in the CNB1 regulatory subunit or inhibition by cyclosporin A. These results indicate that calcineurin inhibits the activity of Ca 2ϩ -ATPases localized in intracellular compartments to maintain Ca 2ϩ homeostasis.
Four classes of Ca 2ϩ -ATPases have been identified up to the present. The well characterized plasma membrane (PMCA) and SERCA Ca 2ϩ pumps differ from one another in their primary structure, mode of regulation, kinetic properties and cellular localization. The yeast PMR1 Ca 2ϩ -ATPase and its rat homologue SPCA1 belong to the third class of secretory pathway Ca 2ϩ pumps. It has recently been shown that the kinetic characteristics and inhibitor sensitivities of PMR1 are clearly distinct from both SERCA and PMCA pumps (70). The yeast vacuolar PMC1 ATPase belongs to the fourth class consisting of plasma membrane-like Ca 2ϩ -ATPases, which are localized in internal acidic compartments and lack the calmodulin-binding site of PMCA ATPases (70).
High sequence similarity to SERCA pumps as well as comparable inhibition of the ATPase activity by thapsigargin (58,59) and cyclopiazonic acid (61,62) clearly indicate that S. mansoni SMA2 belongs to the SERCA subfamily of Ca 2ϩ -AT-Pases. In common with the mammalian non-muscle SERCA3 isoform (71,72), SMA2 lacks the phospholamban-binding region (50), has an optimum pH of 7.0 -7.4, and is relatively insensitive to inhibition by vanadate (60).
SMA2 was found in intracellular yeast membranes enriched with SEC63, an integral membrane protein in the endoplasmic reticulum. The intracellular distribution of SMA2 differs from the plasma membrane localization of rabbit SERCA1b when expressed in yeast (73). One explanation could be the difference  in the yeast expression systems utilized and/or expression levels of the Ca 2ϩ ATPase isoforms. We used the constitutive PMA1 promoter to express SMA2 instead of the inducible GAL1 promoter (73). The low copy number SMA2 expression plasmid was better tolerated than the corresponding high copy number plasmid, which had adverse effects on cell growth and SMA2 expression levels. These results suggest that Ca 2ϩ -ATPase overproduction might be limited in yeast by cellular mechanisms that monitor the levels of membrane proteins and compensate for changes in these levels by inducing synthesis of novel internal membrane structures or karmellae (57). In a separate study, 2 the PMA1 promoter was also used to direct expression of rabbit SERCA1a. The expression level of SERCA1a was 4-fold higher than that of SMA2. In contrast with SMA2, expression of SERCA1a resulted in an accumulation of endoplasmic reticulum membranes which were found to affect fractionation of plasma membrane and endoplasmic reticulum markers in sucrose density gradients.
In the yeast S. cerevisiae, the PMR1 and PMC1 Ca 2ϩ -AT-Pases function together in Ca 2ϩ sequestration. Synthetic lethality of pmr1 pmc1 double mutants in low Ca 2ϩ conditions is prevented by expression of SMA2, indicating that the S. mansoni Ca 2ϩ pump substitutes for loss of both PMR1 and PMC1. The effects of pmc1 mutations on cell growth at high Ca 2ϩ concentrations are partially suppressed by the expression of SMA2 but only in strains deficient in calcineurin (and lacking the VCX1 Ca 2ϩ /H ϩ exchanger). It is unlikely that calcineurin regulates expression levels of SMA2 through the PMA1 promoter and the calcineurin-dependent transcription factor TCN1/CRZ1 (74,75) as calcineurin inactivation has no effect on the amount of PMA1 in plasma membranes (76). We also tried to determine whether SMA2 could suppress the growth defect of pmr1 mutants in calcium-depleted (0.81 M Ca 2ϩ ) growth media. However, we were not able to discriminate between the PMR1 wild-type and the pmr1 mutant strains under these conditions (data not shown).
Ca 2ϩ tolerance of pmc1 mutants is restored by increasing the copy number of the PMR1 or VCX1 genes (17). These three Ca 2ϩ transporters are therefore fully interchangeable in vivo despite a different subcellular localization. Their contributions to Ca 2ϩ tolerance, however, are differently regulated. In the wild-type strain, PMC1 provides the largest contribution be-causeVCX1functionisnegativelyregulatedthroughcalcineurindependent post-translational mechanisms. Expression of a PMC1 reporter gene increases dramatically in response to high Ca 2ϩ concentrations by calcineurin-dependent regulatory processes. In contrast, expression of a PMR1 reporter gene increases only 2-fold in a process which also seems to require Total membranes (P tot fractions) from YPH500 (wild-type), K605 (pmc1⌬), MGY8 -3D (pmc1⌬ cnb1⌬), and MGY8DV (pmc1⌬ cnb1⌬ vcx1⌬) cells expressing SMA2 from the low copy (p315SMA2) or high copy (p426SMA2) plasmids were assayed for thapsigargin-sensitive ATPase activity as described under "Experimental Procedures." ϩ, wild-type allele. The values shown are the mean plus standard deviation from two independent experiments. calcineurin activation (20). By using vacuolar ATPase mutants to analyze the role of calcineurin in intracellular Ca 2ϩ homeostasis, it had been suggested that PMR1 sequesters Ca 2ϩ into the Golgi under the negative control of calcineurin (77). Our results also indicate that SMA2 activity is inhibited by a calcineurin-dependent regulatory mechanism which remains to be identified.
The simplest explanation of these results is that calcineurin activation by Ca 2ϩ /calmodulin diminishes the contribution of PMR1 to Ca 2ϩ tolerance, through a direct or indirect inhibitory effect on the ATPase activity. As a result, Ca 2ϩ uptake into the Golgi is repressed and cytosolic free Ca 2ϩ concentration is increased, leading to further activation of calcineurin and amplification of the Ca 2ϩ signal. This would explain why PMR1, if not overexpressed, cannot replace PMC1 for growth under high Ca 2ϩ conditions. According to this explanation, the function of PMR1 in Ca 2ϩ transport would be derepressed in calcineurindeficient strains. However, the resulting activation of PMR1 would be not sufficient to compensate for loss of PMC1 at high Ca 2ϩ concentrations, as pmc1 cnb1 vcx1 mutants without the SMA2 expression plasmid failed to grow under these conditions. Consistently, the expression level of PMR1 is 2-fold less than that of PMC1 in the presence of 200 mM CaCl 2 (20). Finally, the possibility that the function of PMR1 is also regulated by VCX, through a calcineurin-independent mechanism, is suggested by the increased level of SMA2 ATPase activity in the pmc1 cnb1 vcx1 mutant compared with that seen in the pmc1 cnb1 mutant.
There is no evidence that calcineurin regulates intracellular ion homeostasis in S. mansoni. It is therefore difficult to assess the physiological relevance of our results to the regulation of S. mansoni Ca 2ϩ -ATPases. The mechanisms of calcineurin activation by Ca 2ϩ /calmodulin and inhibition by cyclosporin A are well conserved between yeast and animal cells (69,78). Moreover, the observation that yeast calcineurin inhibits heterologous intracellular Ca 2ϩ pumps is consistent with the effects of calcineurin overexpression on decreased levels of PMA1 activity in yeast (76) and with regulation of Na ϩ /K ϩ -ATPase activity by calcineurin in animal cells (79).
In conclusion, our results show that the yeast S. cerevisiae is an eminently suitable system for heterologous expression of Ca 2ϩ -ATPases from parasites. First, expression of S. mansoni SMA2 in yeast results in the synthesis of a fully functional Ca 2ϩ pump. The enzyme is correctly targeted to the endoplasmic reticulum and exhibits kinetic properties comparable to SERCA3 in mammalian non-muscle tissues. Second, we estimated the amount of SMA2 in yeast membranes as representing 0.5% of total protein. There is no significant background of unrelated Ca 2ϩ -dependent ATPase activity in yeast membranes, due to low levels of PMR1 and PMC1 expression. Moreover, the corresponding genes can be replaced by SMA2 with no effect on cell viability under standard growth conditions. Finally, the possibility to study the function of SMA2 in Ca 2ϩ tolerance and its regulation in vivo expands the usefulness of yeast for structure-function relationships studies on various S. mansoni Ca 2ϩ -ATPase isoforms.