Null Mutation in the Gene Encoding Plasma Membrane Ca2+-ATPase Isoform 2 Impairs Calcium Transport into Milk*

The means by which calcium is transported into the milk produced by mammary glands is a poorly understood process. One hypothesis is that it occurs during exocytosis of secretory products via the Golgi pathway, consistent with the observation that the SPCA1 Ca2+-ATPase, which is expressed in the Golgi, is induced in lactating mammary tissue. However, massive up-regulation of the PMCA2bw plasma membrane Ca2+-ATPase also occurs during lactation and is more strongly correlated with increases in milk calcium, suggesting that calcium may be secreted directly via this pump. To examine the physiological role of PMCA2bw in lactation we compared lactating PMCA2-null mice to heterozygous and wild-type mice. Relative expression levels of individual milk proteins were unaffected by genotype. However, milk from PMCA2-null mice had 60% less calcium than milk from heterozygous and wild-type mice, the total milk protein concentration was lower, and an indirect measure of milk production (litter weights) suggested that the PMCA2-null mice produce significantly less milk. In contrast, lactose was higher in milk from PMCA2-null mice during early lactation, but by day 12 of lactation there were no differences in milk lactose between the three genotypes. These data demonstrate that the activity of PMCA2bw is required for secretion of much of the calcium in milk. This major secretory function represents a novel biological role for the plasma membrane Ca2+-ATPases, which are generally regarded as premier regulators of intracellular Ca2+.

The mammary gland transports large amounts of Ca 2ϩ from the blood to the milk via the mammary secretory cells. These transcellular Ca 2ϩ fluxes must be rigorously controlled to prevent cytotoxicity (1). This is a formidable task, because mammary glands store 12-30 mol of Ca 2ϩ /g of tissue compared with less than 1 mol/g in non-mammary tissues (2,3). The plasma membrane Ca 2ϩ -ATPase (PMCA), 1 secretory pathway Ca 2ϩ -ATPase (SPCA), and sarcoplasmic-endoplasmic reticulum Ca 2ϩ -ATPase (SERCA) are up-regulated in lactating mammary glands (4 -6). However, PMCA2bw 2 is the most abundant Ca 2ϩ -ATPase expressed during lactation, and with an ϳ100fold up-regulation of PMCA2 protein during lactation its expression levels are closely correlated with milk production and calcium secretion (4,5).
PMCAs, along with SERCA and SPCA pumps, are responsible for establishing and maintaining appropriate intracellular Ca 2ϩ levels (7)(8)(9)(10)(11) with SERCAs and SPCAs sequestering Ca 2ϩ in intracellular organelles and PMCAs providing a high affinity system for extrusion of Ca 2ϩ from the cell. Mammalian PMCAs are encoded by a multigene family consisting of four members termed PMCA 1-4 (7,8). Cloning has revealed some 20 -30 PMCA proteins generated by alternative splicing of the primary gene transcripts at two sites identified as A and C. Alternative splicing of site C (COOH-terminal tail) has been shown to alter the regulatory properties of PMCA isoforms particularly with respect to phosphorylation and calmodulin stimulation (12)(13)(14) with splice variant b of PMCA2 exhibiting a very high affinity for calmodulin (15), thereby making it particularly effective for Ca 2ϩ extrusion. Alternative splicing at site A has also been shown to occur in PMCA2 (12,16), but its likely function was determined only recently when Chicka and Strehler (17) showed that the w splicing pattern results in apical membrane targeting of PMCA2 regardless of the splicing pattern at site C. Consistent with those results, the w splice form of PMCA2b, one of four splice variants involving site A, localizes to the apical membrane of mammary secretory cells (5), and a splice variant of bullfrog PMCA2 corresponding to the w variant of mammals is expressed in the sensory cilia located on the apical aspect of vestibular and cochlear hair cells (18).
Gene-targeting studies indicate that individual PMCA isoforms and splice variants control specific physiological functions (19 -21). For example, PMCA2 is abundantly expressed in the stereocilia of vestibular and cochlear hair cells (18), and mice lacking PMCA2 are deaf and exhibit significant difficulties with balance (20). As noted above, PMCA2bw is expressed in mammary glands and is massively up-regulated during lactation (4,5), suggesting that it may be involved in the secretion of Ca 2ϩ into the milk. This is of particular interest because it has been suggested that milk calcium is derived from calcium originally sequestered in vesicles of the secretory pathway and delivered to the milk by exocytosis (22)(23)(24)(25)(26). As far as we are aware, the alternative hypothesis, that Ca 2ϩ is secreted directly via the calmodulin-sensitive plasma membrane Ca 2ϩ pump, has not been considered previously. In the current study, we show that mice homozygous for the loss of PMCA2 produce milk with 60% less calcium than that of heterozygous or wild-type mice. These data show that PMCA2, and specifi-cally the PMCA2bw splice variant (the characteristics of which should allow the enzyme to function as a high efficiency Ca 2ϩ extrusion system on the apical membrane), is necessary to produce the high levels of Ca 2ϩ in milk.

EXPERIMENTAL PROCEDURES
Animal Procedures-All animal procedures were approved by the National Animal Disease Center's Animal Care and Use Committee. PMCA2 knock-out mice were prepared as described (20). Mice were housed individually in hanging basket cages with sawdust bedding. The mice were genotyped using the following primers in a single reaction. The primers are PMCA2-195 (TCC AGA TCG ACA GCG GAA GGA ACG), PMCA2-193 (GAT CCC GTC AAA GAC GTT GCG CTC), and PGK NEO3Ј (CTG ACT AGG GGA GGA GTA GAA GG). For genotyping, tail snips were collected from mice 14 -16 days old. Tail snips were digested according to the instructions of the Sigma REDExtract-N-Amp tissue PCR kit, and PCR was performed on the genomic DNA extract. The PCR products for wild-type and knock-out alleles were 149 and 278 bp, respectively. Female mice of all three genotypes were bred with wild-type males. Starting on day 6 of lactation milk was collected every other day until day 12 of lactation when the mice were anesthetized with a 50:50 mixture of CO 2 :O 2 followed by decapitation. Mammary and brain tissue was collected, flash frozen in liquid N 2 , and stored at Ϫ70°C until processed. Blood samples were taken for measurement of plasma calcium and 1,25(OH) 2 D 3 (27).
Tissue Microsomal Membrane Preparation-Mammary and brain tissue microsomes were prepared as described previously (28). Briefly, tissue was homogenized in 10 volumes of Buffer A, which contained Tris-HCl (10 mM), MgCl 2 (2 mM), phenylmethylsulfonyl fluoride (0.1 mM), EDTA (1 mM), 4 g/ml aprotinin, and 4 g/ml leupeptin at pH 7.5. The homogenate was mixed with an equal volume of Buffer B (Buffer A plus 0.3 M KCl) and centrifuged at 10,000 ϫ g for 10 min. The supernatant was collected, adjusted to 0.7 M KCl by the addition of solid KCl, and centrifuged at 100,000 ϫ g for 1 h. The supernatant was discarded and the pellets were resuspended in Buffer C (Buffer A plus 0.15 M KCl). Membrane preparations were stored at Ϫ70°C until assayed. Proteins were determined using the Bio-Rad Protein Assay kit using a bovine serum albumin standard.
Gel Electrophoresis and Western Blotting-The methods used were basically as described previously (29). Briefly, microsomes were incubated for 15 min at room temperature in a modified Laemmli buffer containing 150 mg/ml urea and 65 mM dithiothreitol. Samples were then electrophoresed for 1.5 h at 125 volts in a 6% Tris-glycine gel (Novex, San Diego, CA). Proteins were transferred to nitrocellulose membranes for 1 h at 25 volts in 0.7 M glycine, 0.025 M Tris at pH 7.4. Blots were developed using Pierce's Supersignal with the protocol provided by the manufacturer. Anti-PMCA2 and SPCA antibodies were described previously (5,6). Anti-SERCA2 antibody was purchased from Affinity BioReagents (Deerfield, IL).
Mouse Milking-Starting on day 6 of lactation mouse milk was collected as follows. Baby mice were removed from the mother for 4 h prior to milk collection. Lactating mice were then anesthetized with an IP injection of ketamine (100 mg/kg of body weight), xylazine (10 mg/kg of body weight), and acepromazine (3 mg/kg of body weight). After they were anesthetized, the mice were given 0.1 IU of oxytocin intraperitoneally, and 10 min was allowed to pass prior to the start of milking. Milk was collected using a device made of small tubing connected to a 1.5-ml tube and a low vacuum that was pulsed. Milk was collected into the tube. Milk calcium was determined by atomic absorption spectroscopy, and milk protein was determined using the Bio-Rad Protein Assay kit using a bovine serum albumin standard. Milk lactose was determined by high performance liquid chromatography. Briefly 50 l of milk was diluted 1/10 in deionized water and added to fructose, which served as an internal standard. The sample was centrifuged for 15 min at 16,000 ϫ g, and 110 l of supernatant was removed for assay. The diluted sample was then chromatographed on a calcium Supelcogel column (300 mm ϫ 7.8 mm) (Supelco, Bellefonte, PA) using water at 0.5 ml/min and a Waters high performance liquid chromatograph (Waters Corporation, Milford, MA) and quantitated with a Waters 410 Differential Refractometer. Lactose concentration was determined using an external calibration curve, and sample extraction efficiency was corrected using the internal fructose standard.
Milk Production-Milk production for the three genotypes was estimated indirectly by weighing the pups on the days of lactation indicated. These measurements were conducted on a separate group of animals that were not milked, and the litter size for all three genotypes had been normalized to 6 pups on day 1 of lactation.
Statistics-Statistics were conducted using the JMP statistical package (SAS Institute, Inc. Cary, NC).
Milk Calcium Is Sharply Reduced by Loss of PMCA2-Milk calcium was measured in mice of all three genotypes on days 6 -12 of lactation. Milk calcium concentrations in both Pmca2 ϩ/ϩ and Pmca2 ϩ/Ϫ mice were ϳ95 mM, whereas milk calcium in Pmca2 Ϫ/Ϫ mice was reduced 60% to ϳ38 mM (p Ͻ 0.01) during the sampling period ( Fig. 2A). Because the bulk of calcium secreted into milk is thought to arrive as a protein complex, the data for calcium were also corrected for milk protein secretion (Fig. 2B). By either measure, Pmca2 Ϫ/Ϫ mice secreted significantly less calcium into their milk than Pmca2 ϩ/ϩ or Pmca2 ϩ/Ϫ mice.
Effect of Mouse Genotype on Milk Protein and Lactose-Analysis of whey acid protein and general milk proteins (Fig.  3A) showed that the mouse genotype had no effect on the overall patterns of milk protein expression. Compared with Pmca2 ϩ/ϩ mice, total milk protein concentration was reduced 39% (p Ͻ 0.01) in Pmca2 Ϫ/Ϫ mice during early lactation but only 14% (p Ͻ 0.05) by day 12 (Fig. 3B). Milk protein concentrations in Pmca2 ϩ/Ϫ mice were more variable and tended to be reduced compared with those of Pmca2 ϩ/ϩ mice, and milk proteins were the same for both Pmca2 ϩ/Ϫ or Pmca2 Ϫ/Ϫ mice by day 12 of lactation.
Effect of Mouse Genotype on Milk Production-Because milk production by nursing mice cannot be measured directly, an indirect measurement of relative milk production by mice of all three genotypes was obtained by weighing their litters on days 2-12. Litter sizes were the same for all genotypes. Litters from Pmca2 Ϫ/Ϫ mice were significantly lighter by day 4 of lactation (Fig. 4B). The growth of Pmca2 Ϫ/Ϫ litters lagged behind those of Pmca2 ϩ/Ϫ or Pmca2 ϩ/ϩ litters throughout the sampling period.
General Calcium Homeostasis of the Genotypes- Table I shows blood calcium and 1,25(OH) 2 D 3 for the three genotypes on day 12 of lactation. All mice had normal blood values. The blood calcium of Pmca2 ϩ/ϩ or Pmca2 ϩ/Ϫ mice was lower (p Ͻ 0.05) and their blood 1,25(OH) 2 D 3 was higher (p Ͻ 0.05) than that of Pmca2 Ϫ/Ϫ mice. DISCUSSION There is considerable information about the structures, tissue distributions, and biochemical characteristics of PMCA isoforms. Only recently, however, with the development of gene-targeted animals, has it been possible to begin systematic studies of the in vivo physiological functions of individual isoforms and their splice variants (8, 10 -12, 19 -21, 30, 31). Analysis of the lactating mammary gland as a site of enormous calcium transport and storage offers the opportunity to gain insights regarding the specific physiological function of PMCA2bw, the most highly expressed Ca 2ϩ -ATPase in lactating mammary tissue (4,5). Therefore the objective of this study was to determine the physiological role of PMCA2bw in calcium transport into milk and its general effects on milk composition.
Our data show that PMCA2bw is the primary controller of milk calcium concentration as the absence of PMCA2bw (Fig.  1B) results in a ϳ60% decline in milk calcium concentration ( Fig. 2A). Interestingly, a 40% decline in PMCA2bw expression in Pmca2 ϩ/Ϫ mice compared with Pmca2 ϩ/ϩ mice had no effect on milk calcium concentrations ( Fig. 2A), suggesting either that apical PMCA2 activity of wild-type mice is not rate-limiting for calcium secretion or that the small increases in the expression of SPCA1 and SERCA2b seen in Pmca2 ϩ/Ϫ mice (Fig. 1, C and D) provide some compensation for the reduction in PMCA2 levels. However, much larger increases in the expression of SPCA1 and SERCA2b seen in Pmca2 Ϫ/Ϫ mice (Fig.  1, C and D) were unable to fully compensate for the complete loss of PMCA2bw in Pmca2 Ϫ/Ϫ mice with respect to milk calcium (Fig. 2, A and B). If the increased expression of SPCA1 and SERCA2b in Pmca2 Ϫ/Ϫ mammary glands does, in fact, provide partial compensation for the deficit in milk calcium, then it would imply that these intracellular pumps play a subsidiary role in the delivery of calcium to the milk, most likely via exocytosis. However, if such compensation is occurring in the knock-out mice, then the normal contribution of PMCA2 to the calcium composition of the milk is likely to be even greater than indicated by the ϳ60% reduction observed in the knock-out mice. Thus, much of the calcium in milk appears to be secreted directly across the apical membrane via PMCA2bw.
Intra-organelle calcium plays a critical role milk protein synthesis, casein phosphorylation, and to a lesser extent protein secretion (32,33). Endoplasmic reticulum calcium is required for milk protein synthesis, as depletion of calcium from the endoplasmic reticulum inhibits milk protein synthesis (33,34). The Golgi on the other hand contains the bulk of lactating mammary tissue calcium (35), and this calcium pool is required for normal casein phosphorylation as well as casein micelle formation (33,34). Despite the many roles for calcium in milk protein synthesis and processing, we found no effect of genotype on either the phosphorylation of milk proteins (data not shown) or the expression of individual milk proteins (Fig. 3A). However, total milk protein concentration was significantly reduced in Pmca2 Ϫ/Ϫ mice (Fig. 3B) and to a lesser extent in Pmca2 ϩ/Ϫ mice. This result suggests that PMCA2bw may affect intra-organelle calcium pools as has been seen for other PMCA pumps (36). Therefore, the loss of PMCA2bw may contribute to the reduction of total milk protein seen in Pmca2 Ϫ/Ϫ mice. However, milk lactose is the primary osmotic regulator in milk. Thus, increased lactose concentrations are associated with higher milk water content (37). The largest decrease in total milk protein (39%) in Pmca2 Ϫ/Ϫ mice occurs on day 6 of lactation when Pmca2 Ϫ/Ϫ mice have significantly higher milk lactose concentrations (Fig. 4A). The increased water transport associated with these higher milk lactose concentrations would lead to the observed decreased milk protein concentrations because of simple dilution. Once lactose concentrations are equal (days 10 -12) for all three genotypes, the reduction in milk protein in Pmca2 Ϫ/Ϫ and Pmca2 ϩ/Ϫ mice is only 14% compared with Pmca2 ϩ/ϩ mice, whereas milk calcium and milk calcium/mg of milk protein has reached its nadir at a 60% reduction in Pmca2 Ϫ/Ϫ mice.
Studies by others have suggested that all calcium (free and bound) arrives in milk via the secretory pathway (22)(23)(24)(25)(26) with little if any milk calcium arriving in milk via direct extrusion across the apical membrane. However, the data supporting these conclusions are indirect at best. The effects of the loss of PMCA2bw on milk calcium suggest that the primary function of PMCA2bw is apical transport of calcium into milk as PMCA2bw is concentrated on the apical membrane of lactating mammary tissue (5). It should be pointed out that PMCA2bw protein turns over rapidly in lactating tissue as the apical membrane is secreted into the milk as a part of milk fat secretion (5). This requires new synthesis of PMCA2bw to meet the needs of apical membrane renewal (5). Therefore, there is always newly synthesized PMCA2bw in transit to the apical membrane of lactating cells. The work of Taylor et al. (38) suggests that PMCAs in transit to the plasma membrane of a cell may contribute to Golgi calcium accumulation. By Western blotting their data showed some PMCA transiting the Golgi. However, their Golgi calcium transport studies did not address the role of the Golgi resident and thapsigargin-insensitive SPCA in liver Golgi. They concluded that thapsigargin-independent calcium uptake by rat liver Golgi was caused by PMCA. It is more likely that this activity is caused by SPCA. Because SPCA1 is the second most abundant Ca 2ϩ -ATPase in lactating tissue, it is likely the primary controller of calcium uptake in mammary Golgi (5). In the case of lactating mammary tissue, any potential contribution of PMCA2bw to intraorganelle calcium accumulation will only be known following direct measures of the relative contributions of each pump to Golgi and endoplasmic reticulum calcium uptake. Finally, the dogma that little or no calcium enters milk via apical membrane calcium transport needs to be reexamined. The large concentration of PMCA2bw present on the apical membrane along with data showing that the loss of PMCA2bw results in a massive reduction in milk calcium strongly argues for a major calcium transport pathway in the apical membrane of lactating mammary tissue.
Milk production (estimated indirectly by litter weight) is sig-  nificantly reduced in Pmca2 Ϫ/Ϫ mice compared with either Pmca2 ϩ/ϩ or Pmca2 ϩ/Ϫ mice (Fig. 4B). This may be caused by the reduced calcium available to support lactation in Pmca2 Ϫ/Ϫ mice but may also be caused by behavioral changes that result from the loss of PMCA2. These animals are deaf and have mobility problems because of loss of balance control that occurs with this mutation. Subjective observations of Pmca2 Ϫ/Ϫ mothers lead to the conclusion that they had reduced mothering skills. Pups separated from these mothers were not gathered up quickly for suckling as was the case for the other two genotypes. Some of this behavior could be reasonably attributed to their inability to hear their pups and/or the extra effort they had to put into feeding themselves as a result of their mobility problems. There was no abnormal alteration in general calcium metabolism in the Pmca2 Ϫ/Ϫ that could have contributed to their lower milk production (Table I.) They had higher blood calcium values and lower blood 1,25(OH) 2 D 3 values than either Pmca2 ϩ/ϩ or Pmca2 ϩ/Ϫ mice. But these values are in the normal range and just reflect the lower calcium requirements for lactation experienced by the Pmca2 Ϫ/Ϫ mothers.
In summary, these studies of lactation in Pmca2 Ϫ/Ϫ mice clearly show that the loss of the major Ca 2ϩ -ATPase expressed in lactating mammary tissue results in a significant reduction in milk calcium and to a lesser extent milk protein. It is concluded that PMCA2bw is the major regulator of calcium required for normal milk production by the mammary gland. The function of PMCA2bw in mammary glands is clearly not to provide fine regulation of cytosolic calcium levels, the role we normally assign to PMCAs. This major macrocalcium secretory function in support of lactation represents a novel biological role for the plasma membrane Ca 2ϩ -ATPases, which are generally regarded as premier regulators of intracellular Ca 2ϩ .