(Received for publication, March 31, 1997, and in revised form, May 13, 1997)
From the Department of Biochemistry and Molecular
Biology, Mayo Clinic/Foundation, Rochester, Minnesota 55905 and the
§ National Institute of Haematology and Immunology, Daroczi
ut 24, 1113 Budapest, Hungary
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 plasma membrane Ca2+ pump plays a key role in controlling the intracellular Ca2+ concentration. This P-type ATPase is regulated by calmodulin and is responsible for the ATP powered removal of Ca2+ from eukaryotic cells (1). The plasma membrane Ca2+ 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 Ca2+ 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 (3-7). 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 calmodulin-binding 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 calmodulin-binding 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 Ca2+ 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.
45CaCl2 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 rPMCA2bThe 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+ vector (Stratagene) at the ApaI site. The full-length DNA was excised from Bluescript SK+ vector with SalI and KpnI, cloned into the expression vector pMM2 and sequenced using the Applied Biosystems Automatic Sequencer.
Collection of Rat BrainA 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 mRNAThe isolation of mRNA from the brain was performed using the FastTrack mRNA isolation kit (Invitrogen), following the standard protocol.
Reverse TranscriptionReverse 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 ReactionPCR 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 µl 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 5 forward PCR primer was
CCTGAATCGGATCCAGACACAG, which contained a BamHI cut site.
The 3
reverse PCR primer was CGACGCGGTACCCTTTAAATTTCACTC, which
contained a KpnI cut site. The expected PCR product was 242 bp. After amplification, 15 µl 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.
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 ~226-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 pMM2 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.
TransfectionTransfection was carried out using Lipofectamine (Life Technologies, Inc.). 150-cm2 flasks were seeded with 25 × 105 COS cells. Transfection was initiated when the cells were 70-80% confluent. DNA-LipofectAMINE complex for each flask was prepared by incubating 8 µg of DNA and 100 µl 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 CellsCrude microsomal membranes were prepared as described by Verma et al. (15).
Ca2+ Transport AssayCa2+ uptake by microsomal vesicles was measured for 5 min at 37 °C by filtration through Millipore membrane filters (0.45 µm, HA) as described previously (16). The desired free Ca2+ 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 Ca2+ uptake was initiated by the addition of ATP (6 mM). The Ca2+ uptake by control membrane vesicles isolated from pMM2-transfected cells was subtracted from each data point.
Immunoblotting1 µg 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-Ca2+ pump antibody (18) or with a monoclonal anti-calmodulin mixture (Sigma) of three antibodies as needed.
To study the ATP-dependent Ca2+ 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 µg) 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 molecular 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 Ca2+ 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 Ca2+ pump and the mitochondrial ATPase as described (16). Table I shows the activity of Ca2+ 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 Ca2+ 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
Ca2+ transport activities at a fixed Ca2+
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 Ca2+-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
Ca2+ pump is the affinity it has for calcium. Consequently,
to study the characteristics of the isoforms, the Ca2+
transport activities were tested by analyzing the dependence of
Ca2+ uptake on free Ca2+ in the presence or in
the absence of calmodulin. In Fig. 3, the Ca2+ 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
Ca2+ for each isoform, the apparent Ca2+
affinities of the isoforms were different. RPMCA2b, and to a lesser
extent rPMCA2a, were more responsive to Ca2+ stimulation
than hPMCA4b (Fig. 3). Both rPMCA2a and rPMCA2b showed higher
activities at lower Ca2+ concentrations.
To determine whether the higher apparent Ca2+ affinity of
rPMCA2b came from a difference in the catalytic core of the enzyme, we
investigated whether the Ca2+ 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 Ca2+ responses of
rPMCA2b and hPMCA4b, in the presence of enough calmodulin to saturate
at high Ca2+, with that of ct120, for which calmodulin was
not added. This figure shows that the Ca2+ 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 Ca2+ concentration used (61 nM).
Additionally, each isoform's characteristics were compared by
measuring the dependence of Ca2+ uptake on free
Ca2+ 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
Ca2+. When measured in the absence of calmodulin both
rPMCA2a and rPMCA2b were 3-4 times more responsive to Ca2+
stimulation than hPMCA4b. In addition, in the absence of calmodulin, the Vmax 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
Ca2+ uptake on free Ca2+ 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 Ca2+ 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 Kd of
about 0.1 nM (12). Fig. 6 shows that the
addition of up to 3 µM C28R2 at 10 µM
Ca2+ did not inhibit the Ca2+ 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 Ca2+ dependent activity measured was not
due to the presence of endogenous calmodulin.
The final experiment of this set involved the use of an anti-calmodulin 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 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.
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 Ca2+-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 Ca2+ 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 Ca2+ 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 Ca2+ 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 Ca2+ 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 Ca2+. We tested for a difference in the catalytic core by comparing the Ca2+ response curve of ct120 (the truncated mutant of hPMCA4b lacking the regulatory carboxyl terminus) with the Ca2+ 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 Ca2+ affinity just as high as that of rPMCA2b. This indicated that the difference in the Ca2+ 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 Ca2+ 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 Ca2+ concentrations and could be overlooked.
In the absence of calmodulin, rPMCA2b showed much higher activity than hPMCA4b at each Ca2+ 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 µM in most cells, in brain about 50 µM); Figs. 3 and 4B show that rPMCA2b remains activated by moderate levels of calmodulin even below 0.1 µM intracellular Ca2+ concentration. This indicates that rPMCA2b is a form of the PMCA pump which is extremely effective in removing Ca2+ from the cytosol, a property which may make an important contribution to the physiology of cells where it is expressed.