Plasma Membrane Ca2+ Pump Isoforms 2a and 2b Are Unusually Responsive to Calmodulin and Ca2+ *

The full-length a and b variants of the rat plasma membrane calcium pump, isoform 2 (rPMCA2a and rPMCA2b), were constructed and expressed in COS-7 cells. To characterize these isoforms, calcium transport was determined in a microsomal fraction. Both rPMCA2a and rPMCA2b had a much higher affinity for calmodulin than the corresponding forms of hPMCA4, and rPMCA2b had the highest affinity among the isoforms that have been tested so far. When analyzed at a relatively high calmodulin concentration, rPMCA2b and, to a lesser extent, rPMCA2a showed higher apparent calcium affinity;i.e. they were more active at lower Ca2+concentrations than hPMCA4b. This indicates that these two variants of rat isoform 2 will tend to maintain a lower free cytosolic Ca2+ level in cells where they are expressed. Both variants also showed a higher level of basal activity (in the complete absence of calmodulin) than hPMCA4b, a property which would reinforce their ability to maintain a low free cytosolic Ca2+concentration. Experiments designed to determine the source of the higher apparent Ca2+ affinity of rPMCA2b showed that it came from the properties of the carboxyl terminus, rather than from any difference in the catalytic core.

The full-length a and b variants of the rat plasma membrane calcium pump, isoform 2 (rPMCA2a and rPMCA2b), were constructed and expressed in COS-7 cells. To characterize these isoforms, calcium transport was determined in a microsomal fraction. Both rPMCA2a and rPMCA2b had a much higher affinity for calmodulin than the corresponding forms of hPMCA4, and rPMCA2b had the highest affinity among the isoforms that have been tested so far. When analyzed at a relatively high calmodulin concentration, rPMCA2b and, to a lesser extent, rPMCA2a showed higher apparent calcium affinity; i.e. they were more active at lower Ca 2؉ concentrations than hPMCA4b. This indicates that these two variants of rat isoform 2 will tend to maintain a lower free cytosolic Ca 2؉ level in cells where they are expressed. Both variants also showed a higher level of basal activity (in the complete absence of calmodulin) than hPMCA4b, a property which would reinforce their ability to maintain a low free cytosolic Ca 2؉ concentration. Experiments designed to determine the source of the higher apparent Ca 2؉ affinity of rPMCA2b showed that it came from the properties of the carboxyl terminus, rather than from any difference in the catalytic core.
The plasma membrane Ca 2ϩ pump plays a key role in controlling the intracellular Ca 2ϩ concentration. This P-type ATPase is regulated by calmodulin and is responsible for the ATP powered removal of Ca 2ϩ from eukaryotic cells (1). The plasma membrane Ca 2ϩ pump (PMCA) 1 has a low level of activity in the absence of calmodulin. Calmodulin binds to an autoinhibitory domain (the C domain), and increases both the maximum velocity of the pump and the apparent Ca 2ϩ affinity.
To date, at least four different genes have been found which encode for PMCA (2). Additional variability is obtained by alternate splices occurring at two sites in the pump (3)(4)(5)(6)(7). In each of the four genes, the alternative splice sites (8) are located in the middle of the cytosolic loop between transmembrane domains 2 and 3 (splice site A) (9) and downstream of the last transmembrane domain, in the middle of the calmodulinbinding domain (splice site C) (9 -11). The first 18 amino acids of the calmodulin-binding domain are conserved for all PMCA isoforms, but the presence of the alternative RNA splice site in the middle of this region (at splice site C) changes the remainder of the calmodulin-binding domain as well as the carboxyl terminus (10). The isoforms whose mRNA contains a spliced-in exon are called "a," while those isoforms lacking the additional exon are called "b." 2 The a variants of the isoforms have a less basic calmodulin-binding domain as well as a different carboxyl terminus than the b variants. When synthetic peptides corresponding to representative a and b forms of the calmodulinbinding domain were compared, the b form of the peptide showed a 10-fold higher affinity for calmodulin than the a form of the peptide (12). Additionally, full-length isoforms hPMCA4a and hPMCA4b were overexpressed in COS-1 cells and the calmodulin-response curves were analyzed. As expected, the hPMCA4b isoform had a higher affinity for calmodulin than the hPMCA4a isoform (13).
In this study, we have compared isoforms 2a and 2b of the plasma membrane Ca 2ϩ pump of the rat with the most widely studied isoform to date, isoform 4b from human tissues. The exact isoforms utilized were rPMCA2az and rPMCA2bz. In this nomenclature, the last letter (z in this case) refers to a splice at site A which changes only a small region of the enzyme because it does not involve a frameshift (5). We did not compare different alternatively spliced products in this region, and so we will not discuss its properties further. In the remainder of this paper we will omit the z since both enzymes studied were of this form. We utilized the rat message instead of human partly because the DNAs were more easily available, but also because we anticipate that many future studies in different laboratories will be carried out with rat tissues. Therefore, utilizing the rat enzymes will allow the development of antibodies and other reagents which will contribute to a coordinated attack on discovering the properties of the different isoforms of the pump in rat. In the case of the present study, the rat enzymes are a good substitute for the human ones since their amino acid sequences are nearly the same. rPMCA2a and hPMCA2a are 97.9% identical when aligned, and rPMCA2b and hPMCA2b are 98.1% identical. Thus, almost all of the differences in properties can be attributed to the differences in the two genes which are being compared, rather than to the differences in species. This is evident when one observes that rPMCA2a is only 75.7% identical with hPMCA4b and that rPMCA2b is 75.4% identical with hPMCA4b. Because of these relationships, we will generally not mention the species differences in our discussion, but will focus on the different gene products and the alternative splice in the downstream region.
Because rPMCA2a and rPMCA2b come from a different gene than hPMCA4, some of the differences between 2 and 4 are scattered throughout the molecule. However, large stretches of the molecule are very highly conserved and the biggest sequence differences are concentrated in three regions: the amino terminus, the carboxyl terminus, and a region near the upstream alternative splice.
A recent paper (14) described the expression of hPMCA2b in insect cells and the effect of an alternate splice at site A. Although the authors found substantial differences between gene products 2 and 4 they did not find any difference in the characteristics of the isoforms of hPMCA2 produced by this alternate splice. In the present paper, we report the overexpression of full-length rPMCA2b and rPMCA2a isoforms from rat in COS-7 cells. rPMCA2a is produced by an alternate splice at site C which changes the calmodulin-binding domain and the rest of the carboxyl terminus of rPMCA2 and is expected to have specific regulatory characteristics. Unlike the case for the alterations at site A, we find substantial differences in properties caused by the alternate splice at site C. This report will focus on the functional properties of these two isoforms and how they compare with the most widely studied isoform to date, isoform 4b from human.

MATERIALS AND METHODS
Chemicals-45 CaCl 2 was purchased from NEN Life Sciences Products. LipofectAMINE, Opti-MEM, and restriction enzymes were obtained from Life Technologies, Inc. Calmodulin was purchased from Sigma. All other chemicals used for this study were of reagent grade.
Construction of the Full-length rPMCA2b-The full-length rPMCA2b isoform in the pBR322 vector was a gift from Dr. G. Shull (University of Cincinnati). The full-length rPMCA2b gene was excised from the pBR322 vector with ApaI and ligated into the Bluescript SK ϩ vector (Stratagene) at the ApaI site. The full-length DNA was excised from Bluescript SK ϩ vector with SalI and KpnI, cloned into the expression vector pMM 2 and sequenced using the Applied Biosystems Automatic Sequencer.
Collection of Rat Brain-A male Harlan Sprague-Dawley rat was given a lethal injection of sodium pentobarbital, then decapitated. The entire brain was rapidly dissected out and immediately placed in liquid nitrogen, where it was kept until use.
Isolation of mRNA-The isolation of mRNA from the brain was performed using the FastTrack mRNA isolation kit (Invitrogen), following the standard protocol.
Reverse Transcription-Reverse transcription was performed using the GeneAmp RNA PCR kit (Perkin-Elmer). Random hexamers were used as the reverse transcriptase primers, with the mRNA isolated from rat brain being used as the template. The incubation times and temperatures for the reverse transcription were 10 min at room temperature, 45 min at 42°C, 5 min at 99°C, and 5 min at 4°C.
Polymerase Chain Reaction-PCR amplification was done to produce the carboxyl terminus of the rPMCA2a isoform, using the GeneAmp RNA PCR kit (Perkin-Elmer), a total volume of 100 l was used. The reaction (in a Perkin-Elmer 9600 thermal cycler) was initiated with a 2-min melting step at 94°C, followed by 35 cycles of 94°C for 1 min, 52°C for 1 min, and 72°C for 1 min, with a final 7-min extension step at 72°C. The 5Ј forward PCR primer was CCTGAATCGGATCCAGA-CACAG, which contained a BamHI cut site. The 3Ј reverse PCR primer was CGACGCGGTACCCTTTAAATTTCACTC, which contained a KpnI cut site. The expected PCR product was 242 bp. After amplification, 15 l of each of the PCR samples were size fractionated by electrophoresis in a 1.5% agarose gel stained with ethidium bromide. A 123-bp DNA ladder (Life Technologies, Inc.) was run along with the PCR samples.
Construction of the Full-length rPMCA2a-We used the entire rPMCA2b cDNA up to the calmodulin-binding domain at which point the 2b sequence was replaced by a PCR product fragment which provided the a form of rPMCA2. RPMCA2b was digested with XhoI and run in a 1% agarose gel. The desired 752-bp DNA fragment (which included the calmodulin-binding domain) was excised from the gel, purified with GeneClean (BIO 101, Inc.) and cloned into the pSP72 vector (Promega) at the XhoI site. Both the 752-bp fragment in the pSP72 vector and the 242-bp PCR fragment were digested with BamHI and KpnI and run in a 1% agarose gel. The 2.9-kilobase band from the digested 752-bp fragment in the pSP72 vector and the ϳ226-bp digested PCR product were excised from the gel, cleaned using the GeneClean and Mermaid kits (BIO 101, Inc.), respectively, and ligated together using standard protocol. This ligation formed the rPMCA2a carboxyl terminus which included the new calmodulin-binding domain. The rPMCA2a carboxyl terminus in pSP72 and the full-length rPMCA2b in the pMM 2 vector were each digested with XhoI and KpnI. The desired bands (the rPMCA2b minus the calmodulin-binding domain in the pMM2 vector and the rPMCA2a carboxyl terminus from pSP72) were ligated together and sequenced.
Transfection-Transfection was carried out using Lipofectamine (Life Technologies, Inc.). 150-cm 2 flasks were seeded with 25 ϫ 10 5 COS cells. Transfection was initiated when the cells were 70 -80% confluent. DNA-LipofectAMINE complex for each flask was prepared by incubating 8 g of DNA and 100 l of LipofectAMINE in 3.6 ml of serum-free Opti-MEM for 30 min at room temperature. The cells were incubated with the DNA-LipofectAMINE complex in 14.5 ml of serum-free Opti-MEM for 5 h at 37°C, then supplemented with serum and the incubation continued for a total of 24 h. The DNA-LipofectAMINE-containing medium was then replaced with fresh tissue culture medium with 10% serum and the cells were incubated at 37°C for an additional 24 h.
Isolation of Microsomes from COS Cells-Crude microsomal membranes were prepared as described by Verma et al. (15).
Ca 2ϩ Transport Assay-Ca 2ϩ uptake by microsomal vesicles was measured for 5 min at 37°C by filtration through Millipore membrane filters (0.45 m, HA) as described previously (16). The desired free Ca 2ϩ concentrations were obtained by varying the concentration of EGTA. The microsomes were preincubated at 37°C for 3 min in the appropriate concentration of calmodulin before Ca 2ϩ uptake was initiated by the addition of ATP (6 mM). The Ca 2ϩ uptake by control membrane vesicles isolated from pMM2-transfected cells was subtracted from each data point.
Immunoblotting-1 g of microsomal membrane proteins was dissolved in electrophoresis sample buffer containing 62.5 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 5 mM EDTA, 125 mg/ml urea, and 100 mM dithiothreitol. The samples were then loaded on a 7.5% acrylamide gel following Laemmli's procedure (17). The samples were blotted after electrophoresis to polyvinylidine difluoride membrane (Bio-Rad) using 25 mM Tris, 0.7 M glycine as a transfer solution. The blots were immunostained with 5F10 monoclonal anti-Ca 2ϩ pump antibody (18) or with a monoclonal anti-calmodulin mixture (Sigma) of three antibodies as needed.

RESULTS
Expression of rPMCA2a and rPMCA2b in COS Cells-To study the ATP-dependent Ca 2ϩ transport activity of rPMCA2a and rPMCA2b isoforms, the full-length clones were overexpressed in COS cells. These isoforms are identical up to residue 1095, but downstream of this residue the alternative splice in the calmodulin-binding domain changes the carboxyl terminus for each isoform. The carboxyl terminus of rPMCA2a is shorter, so that the protein produced has 1154 amino acids, while the rPMCA2b protein has 1198 amino acids (10).
Crude membranes from COS cells transfected with cDNA encoding hPMCA4a, hPMCA4b, rPMCA2a, and rPMCA2b were prepared, solubilized, and their protein (1 g) separated by a 7.5% SDS-polyacrylamide gel. Fig. 1 shows a Western blot using the monoclonal antibody 5F10 to visualize rPMCA2a, rPMCA2b, hPMCA4a, and hPMCA4b. The estimated molecu- lar weight for isoforms rPMCA2a and rPMCA2b corresponded to the expected molecular mass for each isoform based on their protein sequences. The level of expression was nearly the same for all four isoforms.
Determination of Ca 2ϩ Transport Activities in Microsomes from rPMCA2a, rPMCA2b, and hPMCA4b-transfected COS Cells-Microsomal membranes isolated from rPMCA2a, rPMCA2b, and hPMCA4b transfected cells were assayed in the presence of thapsigargin and oligomycin to inhibit the activity of the endogenous endoplasmic reticulum Ca 2ϩ pump and the mitochondrial ATPase as described (16). Table I shows the activity of Ca 2ϩ uptake by the microsomal membranes from rPMCA2a, rPMCA2b, and hPMCA4b transfected COS cells in the absence and presence of calmodulin (540 nM). Both rPMCA2a and rPMCA2b had higher basal activity than hPMCA4b when tested in the absence of calmodulin. The activity when measured at saturating calmodulin and Ca 2ϩ concentrations was nearly the same, with only small variations due to variations in the expression level.
The calmodulin dependence of each isoform was observed by measuring its Ca 2ϩ transport activities at a fixed Ca 2ϩ concentration as a function of calmodulin concentration. The calmodulin response curves of the isoforms are compared in Fig. 2. RPMCA2b showed the highest sensitivity to calmodulin; this isoform was over four times more sensitive to calmodulin than hPMCA4b. These results were consistent with those of Hilfiker et al. (14) when they studied these pumps expressed in insect cells and measured the Ca 2ϩ -dependent ATPase activity. Our study also included the calmodulin response curve of isoform rPMCA2a. Although rPMCA2a had somewhat lower affinity for calmodulin than rPMCA2b, it still showed much higher affinity than the corresponding 4 isoform, hPMCA4a (see Enyedi et al. (13)).
A crucial element in the behavior of the plasma membrane Ca 2ϩ pump is the affinity it has for calcium. Consequently, to study the characteristics of the isoforms, the Ca 2ϩ transport activities were tested by analyzing the dependence of Ca 2ϩ uptake on free Ca 2ϩ in the presence or in the absence of calmodulin. In Fig. 3, the Ca 2ϩ transport activities of rPMCA2a and rPMCA2b at 540 nM calmodulin are compared with those of hPMCA4b. Although 540 nM calmodulin gave nearly full activation at saturating Ca 2ϩ for each isoform, the apparent Ca 2ϩ affinities of the isoforms were different. RPMCA2b, and to a lesser extent rPMCA2a, were more responsive to Ca 2ϩ stimulation than hPMCA4b (Fig. 3). Both rPMCA2a and rPMCA2b showed higher activities at lower Ca 2ϩ concentrations.
To determine whether the higher apparent Ca 2ϩ affinity of rPMCA2b came from a difference in the catalytic core of the enzyme, we investigated whether the Ca 2ϩ responsiveness of hPMCA4b could be brought up to the level displayed by rPMCA2b. We did this by comparing ct120, a constitutively activated form of hPMCA4b, with rPMCA2b. Ct120 is made from hPMCA4b by removal of the regulatory carboxyl terminus, and is fully activated without calmodulin (16). Fig. 4A compares the Ca 2ϩ responses of rPMCA2b and hPMCA4b, in the presence of enough calmodulin to saturate at high Ca 2ϩ , with that of ct120, for which calmodulin was not added. This figure shows that the Ca 2ϩ response of rPMCA2b is essentially identical to that of ct120. Fig. 4B shows that the level of calmodulin (540 nM) used in Fig. 4A is enough to give full activation of rPMCA2b, even at the lowest free Ca 2ϩ concentration used (61 nM).
Additionally, each isoform's characteristics were compared by measuring the dependence of Ca 2ϩ uptake on free Ca 2ϩ in the absence of calmodulin. The activities measured in the absence of calmodulin (Fig. 5A) are graphed as a percentage of the maximum activity for each isoform that was determined in the presence of saturating calmodulin and Ca 2ϩ . When measured  2. rPMCA2b is more sensitive to calmodulin stimulation than hPMCA4b. The calmodulin dependence of Ca 2ϩ uptake by microsomal vesicles isolated from COS-7 cells transfected with hPMCA4b (open triangles), rPMCA2a (diamonds), or rPMCA2b (circles) is shown. Ca 2ϩ uptake by vesicles made from pMM2-transfected cells has been subtracted from all data points. Membrane vesicles were preincubated at 37°C and Ca 2ϩ uptake was initiated by the addition of ATP. The free Ca 2ϩ concentration was 10 M. Ca 2ϩ uptake activity was measured as in the absence of calmodulin both rPMCA2a and rPMCA2b were 3-4 times more responsive to Ca 2ϩ stimulation than hPMCA4b. In addition, in the absence of calmodulin, the V max for rPMCA2b was much higher than for rPMCA2a or hPMCA4b. It was 71% of the maximum activity whereas in the case of rPMCA2a and hPMCA4b it was 46 and 23%, respectively.
Table I also shows that the basal activity of rPMCA2b was the highest among the isoforms. In the presence of calmodulin hPMCA4b had a maximum velocity over four times above its basal level, while rPMCA2a and rPMCA2b had maximum velocities only 2 and 1.4 times, respectively, over the basal level (Fig. 5, Table I). Hilfiker et al. (14) also had similar results with the human PMCA2 isoform, which they discussed in connection with the possibility of incomplete removal of the endogenous calmodulin.
To address this concern, we did additional experiments to determine whether the endogenous calmodulin was removed during the preparation of the microsomal membranes. In Fig.  5B, the dependence of Ca 2ϩ uptake on free Ca 2ϩ by rPMCA2b in the absence of calmodulin was measured again. This time the membrane was preincubated with or without EGTA for 10 min at 37°C prior to the start of the Ca 2ϩ uptake assay without calmodulin. The slight decrease in the activity of rPMCA2b indicates that only a small amount of calmodulin was removed by the EGTA treatment. The second experiment addressing the high basal level of rPMCA2b was done using synthetic peptide C28R2. This synthetic peptide corresponds to the 28 residues that make up the calmodulin-binding domain in isoform rPMCA2b, and binds calmodulin very tightly, with a K d of about 0.1 nM (12). Fig. 6 shows that the addition of up to 3 M C28R2 at 10 M Ca 2ϩ did not inhibit the Ca 2ϩ uptake of rPMCA2b in the absence of calmodulin. As a positive control, the inhibition of rPMCA2b by C28R2 was also tested when the enzyme was activated by exogenous calmodulin. As Fig. 6 shows, C28R2 inhibited the calmodulin-activated portion of the activity easily. These results indicated that the Ca 2ϩ dependent activity measured was not due to the presence of endogenous calmodulin.
The final experiment of this set involved the use of an anticalmodulin antibody to test for the presence of calmodulin in the microsomal membrane. Equal amounts of hPMCA4b and rPMCA2b microsomal fractions were immunoblotted and stained with an anti-calmodulin antibody following SDS-gel electrophoresis. No calmodulin was detected in either sample (results not shown). Unfortunately, the significance of this result was limited by the insensitivity of the anti-calmodulin FIG. 4. A, a fully activated construct from hPMCA4b is equally sensitive to Ca 2ϩ as is rPMCA2b. The activity of rPMCA2b and hPMCA4b with calmodulin and of the construct ct120 are shown. Measurements and symbols are as in Fig. 3, except that inverted triangles represent ct120. Data points are from assays using three to four different membrane preparations. B, calmodulin dependence of rPMCA2b at 61 nM free Ca 2ϩ . The conditions were like those used for Fig. 2, except for the much lower free Ca 2ϩ . antibody. The results obtained in the experiments using EGTA incubation, C28R2, and the anti-calmodulin antibody indicated that we had been successful in removing endogenous calmodulin and that the high basal level in rPMCA2b is an intrinsic property of the enzyme. DISCUSSION This paper analyzed the characteristics of isoforms rPMCA2b and rPMCA2a and compared them to those of the more widely studied form, hPMCA4b. A recent study (14) has expressed, in insect cells, hPMCA2b (which they called PMCAIIA) and variants of gene 2 at splice site A and studied their Ca 2ϩ -ATPase activity. When they compared hPMCA2b to hPMCA4b, they found hPMCA2b had a much higher affinity for calmodulin. The work reported here expressed rPMCA2b and also another splicing variant of gene 2, rPMCA2a, in COS cells and showed active Ca 2ϩ transport activity. The two isoforms, 2b and 2a, differ only in their carboxyl terminus. They are produced by an alternate splice which changes the structure of the carboxyl terminus starting from the middle of the calmodulin-binding domain. As a result of a similar alternate splice, the affinities of hPMCA4a and 4b for calmodulin were very different; hPMCA4b had a much higher affinity for calmodulin than hPMCA4a (13). Our data measuring active Ca 2ϩ transport (Fig. 2) agreed with that of Hilfiker et al. (14), that isoform 2b had higher affinity for calmodulin than hPMCA4b. An additional finding of this paper was that, although rPMCA2a showed a lower affinity for calmodulin than rPMCA2b, its affinity was still much higher than that of hPMCA4a (see Enyedi et al. (13) for comparison). Since the calmodulin-binding domains of isoforms 2 and 4 are not very different from one another, it seemed probable that the difference in the calmodulin affinity of these isoforms originates in another region of rPMCA2.
We observed that the apparent Ca 2ϩ affinity of the 3 isoforms decreased in the order rPMCA2b, rPMCA2a, and hPMCA4b, when tested at a relatively high concentration of calmodulin. The higher apparent Ca 2ϩ affinity of rPMCA2b could be due to a difference in the regulatory carboxyl terminus, or to a difference in affinity of the catalytic core of the molecule for Ca 2ϩ . We tested for a difference in the catalytic core by comparing the Ca 2ϩ response curve of ct120 (the truncated mutant of hPMCA4b lacking the regulatory carboxyl terminus) with the Ca 2ϩ response curve of rPMCA2b which had been saturated with calmodulin. The result, shown in Fig.  4A, showed that the catalytic core of hPMCA4b was capable of a Ca 2ϩ affinity just as high as that of rPMCA2b. This indicated that the difference in the Ca 2ϩ affinity comes from the difference in the carboxyl terminus of these isoforms. Hilfiker et al. (14) were not able to detect any difference between the Ca 2ϩ stimulation of rPMCA2b and hPMCA4b in the presence of calmodulin, perhaps because of the different expression and assay system they used. Also, this difference in the activity is seen only at low Ca 2ϩ concentrations and could be overlooked.
In the absence of calmodulin, rPMCA2b showed much higher activity than hPMCA4b at each Ca 2ϩ concentration. These results agreed with Hilfiker et al. (14), who discussed the higher activity of rPMCA2b in the context of the tight binding of calmodulin to the enzyme. Using several methods, our results indicated that no calmodulin remained bound to the membranes containing rPMCA2b. Recent experiments indicated that in addition to the calmodulin-binding domain, hPMCA4b has a downstream inhibitory region which is responsible for the very low activity of this isoform in the absence of calmodulin (19,20). Since this region, between residues 1113 and 1134 of hPMCA4b, is quite different from the corresponding region of rPMCA2b, it is possible that, unlike hPMCA4b, rPMCA2b does not have an extra downstream inhibitory region.
RPMCA2a also showed higher activity in the absence of calmodulin than hPMCA4b; it resembled hPMCA4a in this respect (compare Fig. 5A with Fig. 5 from Verma et al. (15)). Recent experiments (15) have shown that hPMCA4a has a much longer calmodulin-binding domain than hPMCA4b and that the whole inhibitory region (which appears to be less effective in self-inhibition than the one in hPMCA4b) is included within this domain. The structure of the regulatory region of the rPMCA2 isoforms remains to be determined but the data on rPMCA2a indicate that the inhibitory regions of rPMCA2a and hPMCA4a might be similar.
Tissue distribution of the PMCA isoforms in human and rat has been examined by S1 nuclease protection, polymerase chain reaction, in situ hybridization, and at the protein level by Western blot analysis (8,10,21). These studies have shown that PMCA1 and 4 are broadly distributed, leading to the suggestion that they represent the "housekeeping" isoforms. On the other hand, PMCA2 and 3 were only detected in specialized tissues and cells: RPMCA2b mRNA has been localized to brain, heart, liver, skeletal muscle, spleen, and testes whereas rPMCA2a is found only in brain and heart. The data presented in this paper suggest that rPMCA2b will have very different properties from hPMCA4b under physiological conditions. Intracellular calmodulin concentrations are usually very high (2-5 M in most cells, in brain about 50 M); Figs. 3 and 4B show that rPMCA2b remains activated by moderate levels of calmodulin even below 0.1 M intracellular Ca 2ϩ concentration. This indicates that rPMCA2b is a form of the PMCA pump which is extremely effective in removing Ca 2ϩ from the cytosol, a property which may make an important contribution to the physiology of cells where it is expressed.
FIG. 6. C28R2 does not inhibit rPMCA2b in the absence of added calmodulin, but does inhibit the portion of the activity stimulated by calmodulin. Ca 2ϩ uptake was measured as in Fig. 3, except that here it was measured as a function of C28R2 concentration at a constant Ca 2ϩ concentration of 10 M. The calmodulin concentration was 50 nM in the experiment shown in the lower panel, enough to fully activate the enzyme.