Plasma Membrane Ca²⁺-ATPase Overexpression Reduces Ca²⁺ Oscillations and Increases Insulin Release Induced by Glucose in Insulin-Secreting BRIN-BD11 Cells

RPMI-1640, l-glutamine, penicillin, steptomycin, G418, FCS, Ca²⁺, and Mg²⁺-free Hank’s balanced salt solution (HBSS) as well as cell dissociation buffer, PBS, NotI, and AlwNI were from Life Technologies (Merelbeke, Belgium). Pluronic F-127, furaptra, and fura-2 were from Molecular Probes (Eugene, OR). SgrAI and KpnI were from Boehringer Mannheim (Brussels). Oligomycin, thapsigargin, and calmodulin were from Calbiochem (Bierges, Belgium). Digitonin was from Sigma Aldrich (Bornem, Belgium). Ouabain was from ICN Biomedicals (Aurora, OH).

Rat pancreatic islets isolation, total RNA preparation, RT, PCR, cloning, and sequencing of PCR products has been described elsewhere (20). To construct the full-length PMCA2wb cDNA, three different fragments of 1,140, 1,669, and 1,562 bp were amplified by PCR using different primer pairs (Table 1) based on rat PMCA2 cDNA sequence (21). The first fragment (I), containing the 135 bp of site A, was amplified using primer pair 1-2. Primer 1 contains the ATG starting codon, which is underlined in Table 1. The second fragment (II) was amplified using primer pair 3-4 and also contains the 135 bp of site A. The third fragment (III) was amplified using primer pair 5-6; primer 6 contains the TAG stop codon, which is underlined in Table 1, and does not contain site C. The PCR fragments I-III were subcloned into the vector pCR-Blunt (Invitrogen, San Diego, CA). DNA sequencing of selected clones was determined using cycle sequencing with AmpliTaq polymerase (Perkin-Elmer). The rat full-length cDNA clone PMCA2wb was constructed by digestion and ligation of fragments I–III. The 5′-end 1,087-bp NotI-AlwNI fragment containing the alternatively spliced region site A, the 1,107-bp AlwNI-SgrAI central portion fragment, and the 1,542-bp SgrAI-KpnI 3′-end fragment were ligated to each other (Table 2). The final full-length PMCA2wb was subcloned into the multicloning site of the pcDNA3.1(−) mammalian expression vector (Invitrogen) and digested by NotI and KpnI. The positive clones were verified by restriction enzyme mapping and sequencing.

Cells were grown in RPMI-1640 medium supplemented with 10% FCS, 2 mmol/l l-glutamine, 100 IU/ml penicillin, and 100 μ g/ml streptomycin in an incubator with 5% CO₂ and 95% air at 37°C. The glucose concentration in the culture medium was 11.1 mmol/l, except where otherwise stated. Cells were gently washed with Ca²⁺- and Mg²⁺-free HBSS for 30 s at room temperature, detached from culture flask by incubation with cell dissociation buffer for 1–2 min, and immediately used for further applications. To establish a cell line that stably expresses rat PMCA2wb, BRIN-BD11 cells were transfected with 5 μg PMCA2wb recombinant pcDNA3.1(−) vector by the LipofectAMINE method (Life Technologies). BRIN-BD11 cells were also transfected with the pcDNA3.1(−) vector carrying the neomycin-resistance gene but lacking the PMCA2wb cDNA. The neomycin-resistant colonies were selected with 250 μg/ml G418 in RPMI culture medium. Colonies were then picked and grown as individual clones in the presence of 250 μg/ml G418. The transcription and expression levels of PMCA2wb protein in the transfected cells were determined by RT-PCR, Western blot analysis, immunofluorescence, and PMCA activity.

To determine the transcription pattern of the PMCA2wb isoform, a quantitative RT-PCR method was used and total RNA was isolated from control and PMCA2wb-transfected BRIN-BD11 cells as previously described (21). The quantitative analysis of PMCA isoforms was performed using 2 μg total RNA in RT reaction. The cDNA was diluted in a final volume of 50 μl, and 3 μl were amplified by PCR using two different primer pairs (Table 3) and the Pwo DNA polymerase kit (Boehringer Mannheim). We used primer pair 7-8 (30 pmol each) for PMCA2 and primer pair 9-10 (10 pmol each) for β-actin (22) as an internal control. After amplification, the samples were analyzed on 1.2% agarose gel stained with ethidium bromide and the cDNA bands were quantified by scanning densitometry.

Plasmamembranes were prepared from control BRIN-BD11 cells as well as transfected cells as previously described (23), with some modifications. Cells were washed twice with chilled PBS after detachment and suspended in hypotonic 5 mmol/l Tris-HCl buffer (pH 8.0) (5 × 10⁷ cell per ml buffer) containing proteinase inhibitors (1 mmol/l PMSF, 1 mmol/l iodoacetate, 1 mmol/l benzamidine, 2 mmol/l DTT, 1 mmol/l EDTA, 0.1 mg/ml soybean trypsin inhibitor, and 10 μmol/l leupeptin), left for 40 min on ice, and then homogenized in a glass homogenizer. The homogenate was centrifuged at 900g for 10 min at 4°C to remove nuclei, and the supernatant was centrifuged for 40 min at 120,000g. The final pellet was resuspended at a concentration of 2–6 mg/ml protein in 20 mmol/l ice-cold Mops buffer (pH 7.4) containing the proteinase inhibitors described above. The membrane preparations were aliquoted, frozen immediately in liquid nitrogen, and stored at − 70°C. Protein concentration was determined using Bio Rad Dc Protein Assay (Bio-Rad, Nazareth, Belgium) with BSA as standard.

Membrane proteins were separated by electrophoresis on 1-mm-thick 7.5% SDS-PAGE according to Laemmli (24). Proteins were transferred electrophoretically onto Hybond ECL nitrocellulose membrane (Amersham) using an IMM-2 Semi-Dry Blotting Device System (WEP) with a continuous buffer system at 200 mA (10 V) for 2 h at room temperature, as recommended by the manufacturer. Kaleidoscope Prestained Standards (Bio-Rad) were used for molecular mass determination of protein bands. Nitrocellulose membrane (Amersham) was analyzed using a BM Chemiluminescence Western Blotting Kit (Mouse/ Rabbit) (Boehringer Mannheim). Membranes were incubated with rabbit polyclonal antibodies raised against PMCA2 (rabbit anti-PMCA2; SWant, Bellinzona, Switzerland) diluted 1:1,000 in blocking solution for 12–16 h at 4°C. The blots were visualized and quantified by scanning densitometry using Chemi Doc (Bio-Rad).

Control and PMCA2wb-transfected BRIN-BD11 cells were plated on coverslips and analyzed by indirect immunofluorescence microscopy 48–60 h after plating. The cells were washed with Tris-buffered saline (TBS; 20 mmol/l Tris and 137 mmol/l NaCl, pH 7.2), fixed for 20 min in 4% formol at pH 7.4 (4°C), and washed with TBS. The cells were then permeabilized with a solution containing 0.01% Triton X-100, 197 μmol/l MgCl₂, 19.5 μmol/l DTT, and 10% glycerol (pH 7.4, 4°C), washed twice with TBS, and incubated in a blocking buffer containing 1% horse serum (Vector Laboratories, Burlingame, CA) in TBS for 20 min. The coverslips were exposed to the primary antibody (rabbit anti-PMCA2) and diluted 1/1,000 in TBS 1% BSA buffer for 1 h. The control cells were incubated in TBS 1% BSA buffer without antibody and washed three times with TBS. The cells were then treated with secondary antibody (Alexa Fluor 594 goat anti-rabbit IgG [H + L] conjugate; Molecular Probes), diluted 1/400 in TBS 1% BSA for 45 min, and washed four times with TBS. The cells were incubated with 300 nmol/l 4′, 6-diamidino-2-phenyoindole (DAPI) solution (Molecular Probes) and washed twice with TBS. The coverslips were mounted on slides using Vectashield (Vector Laboratories) and examined using an Axioplan microscope (Zeiss) equipped with an HBO 100 W or XBO 100 W illuminator and a ×100 objective and photographed with a Dual-mode cooled CCD Camera C4880 (Hamamatsu).

Crude microsomal membranes were prepared following a modified version of a procedure previously described (25). Control and PMCA2wb-transfected cells (5–10 × 10⁶/ml) were suspended in hypotonic solution containing 10 mmol/l Tris-HCl (pH 7.5), 1 mmol/l MgCl₂, 0.1 mmol/l PMSF, 4 μg/ml aprotinin, 1 μg/ml leupeptin, and 2 mmol/l DTT. The cells were swollen for 15 min on ice and then homogenized with 40 strokes in a Dounce homogenizer. The homogenate was diluted with an equal volume of a solution containing 0.5 mol/l sucrose, 0.3 mol/l KCl, 2 mmol/l DTT, and 10 mmol/l Tris-HCl (pH 7.5), homogenized again with 20 strokes, and centrifuged at 5,000g for 15 min at 4°C. The supernatant was made in 0.6 mol/l KCl and, to remove calmodulin, an excess of EDTA (1.5 mmol/l) was also added. The suspension was centrifuged at 110,000g for 1 h at 4°C to pellet the microsomal fraction. The final pellet was resuspended in a solution containing 0.25 mol/l sucrose, 0.15 mol/l KCl, 10 mmol/l Tris-HCl (pH 7.5), 2 mmol/l DTT, and 20 μmol/l CaCl₂ at a protein concentration of 1–3 mg/ml, aliquoted, frozen immediately in liquid nitrogen, and stored at −70°C. Protein concentration was determined using Bio Rad Dc Protein Assay (Bio-Rad) with BSA as standard.

Ca²⁺ influx into microsomal vesicles was measured at 37°C by rapid filtration through filters as previously described, with some modifications (25). The filters were presoaked in a 150 mmol/l KCl/1 mmol/l CaCl₂ solution to reduce background. A total of 20 μg microsomal proteins were resuspended in the uptake buffer containing 100 mmol/l KCl, 25 mmol/l TES-triethanolamine (pH 7.2), 5 mmol/l NaN₃, 4 μg/ml oligomycin, 0.5 mmol/l ouabain, 7 mmol/l MgCl₂, 40 mmol/l KH₂PO₄, 1 mmol/l EGTA, and sufficient CaCl₂ (including ⁴⁵Ca²⁺) to obtain the desired free [ Ca²⁺] using the Max Chelator program (Stanford University). Thapsigargin (200 nmol/l) was also included to inhibit the activity of the sarco(endoplasmic) reticulum Ca²⁺-ATPases (SERCAs). Immediately after the addition of ⁴⁵Ca²⁺ (2–3 10⁶ cpm/ml), the reaction was started by addition of 6 mmol/l ATP. The incubation time ranged from 0 to 10 min, and the transport activity was determined in the presence or absence of 240 mmol/l calmodulin. At given times, 100 μl of the mixture were rapidly filtered through 0.45 μmol/l nitrocellulose filters (Millipore, Brussels), which were washed three times with a cold 150-mmol/l KCl/1 mmol/l CaCl₂ solution. The filters were then counted with a Beckman scintillation counter.

The media used to incubate the BRIN-BD11 cells was a Krebs-Ringer solution (pH 7.4, 37°C) containing 115 mmol/l NaCl, 1 mmol/l CaCl₂, 5 mmol/l KCl, 1 mmol/l MgCl₂, and 10 mmol/l HEPES/NaOH. The media were gassed with ambient air. The different media contained 1.1, 11.1, or 16.7 mmol/l glucose. The method used for the measurement of ⁴⁵Ca²⁺ uptake was performed as previously described (26). In brief, BRIN-BD11 cells were preincubated in 1 ml of a nonradioactive solution for 30 min and then incubated for 90 min in 1 ml of the same medium, which also contained 10 μCi/ml ⁴⁵Ca²⁺. At the end of the incubation, the cells were separated from the incubation medium using a combined lanthanum and oil technique (27).

A drop of the cell suspension containing 50,000 cells was aliquoted onto round glass coverslips placed in Petri dishes and incubated at 37°C in a 5% CO₂/95% O₂ incubator. After 2 h, fresh culture medium was added to the Petri dishes. The cells were further incubated for 24 h at 37°C before use. The medium used to incubate or perfuse the BRIN-BD11 cells consisted of a Krebs-Ringer bicarbonate-buffered solution (pH 7.4, 37°C) containing 115 mmol/l NaCl, 5 mmol/l KCl, 1 mmol/l CaCl₂, 1 mmol/l MgCl₂, 24 mmol/l NaHCO₃, and 1.1 mmol/l glucose. The medium was equilibrated against a mixture of O₂ (95%) and CO₂ (5%) at pH 7.4. The different media also contained, when required, 1.1, 11.1, or 16.7 mmol/l glucose, 50 mmol/l KCl, 10 μmol/l verapamil, and 2 μmol/l thapsigargin. The cells were incubated with fura-2 acetoxymethyl ester (final concentration 2 μ mol/l) during 60 min at 37°C in a Krebs-Ringer bicarbonate-buffered solution containing 2.8 mmol/l glucose. The coverslips were then transferred to a tissue chamber mounted on an inverted fluorescence microscope (Diaphot TDM; Nikon, Tokyo) for epifluorescence. Fura-2 fluorescence of single cells was measured by dual-excitation fluorimetry using a camera-based image analysis system (Magical; Applied Imaging, Sunderland, U.K.). The excitation and emission wavelengths were set at 340/380 and 510 nm, respectively, and a pair of ratioable images (at the excitations of 340 and 380 nm, 30-ms interval) were taken every 2.5 s. [Ca²⁺]_i was calculated from the ratios of the 340- and 380-nm signals as previously described (28).

Furaptra loading, cell permeabilization, and measurements of intracellular Ca²⁺ stores were performed as described (29), with some modifications. Briefly, cells were aliquoted onto round glass coverslips as described above and loaded with furaptra acetomethyl ester (4 μmol/l) in an extracellular buffer (125 mmol/l NaCl, 5.9 mmol/l KCl, 1.2 mmol/l MgCl₂, 1.3 mmol/l CaCl₂, and 25 mmol/l HEPES; pH 7.4) containing 0.05% (wt/vol) Pluronic F-127 for 60 min at 37°C. After rinsing, the coverslips were then transferred to a tissue chamber and mounted on an inverted fluorescence microscope. Permeabilization of the plasma membrane was carried out with 4 μmol/l digitonin in intracelllular buffer (140 mmol/l KCl, 3 mmol/l Na₂ATP, 2.1 mmol/l MgCl₂, 2 mmol/l EGTA, 10 mmol/l HEPES, and 0.418 mmol/l CaCl₂). Furaptra fluorescence of single-loaded cells was measured by use of dual-excitation microfluorimetry with the Spex photometric system (Optilas, Alphen aan den Rijn, Holland). The excitation wavelengths (340 and 380 nm) were alternated at the frequency of 1 Hz. The emission wavelength was set at 510 nm. After the sudden drop of fluorescence, the detergent was removed and intracellular buffer was added.

Insulin release from BRIN-BD11 cells in tissue culture was determined using cell monolayers. The BRIN-BD11 cells were harvested and seeded into 24-multiwell plates at a density of 5 × 10⁵ per well and allowed to attach overnight at 37°C. Standard RPMI-1640 tissue culture medium containing 11.1 mmol/l glucose, which was routinely used to culture BRIN-BD11 cells, was removed from each well and replaced with culture media containing either 1.1 or 16.7 mmol/l glucose. The cells were incubated for 24 h at 37°C, and aliquots of the culture media were removed from each separate well and stored at −20°C for subsequent insulin measurement by radioimmunoassay (30).

Insulin release from BRIN-BD11 cells monolayers was also measured over 90 min. After 24 h culture at 37°C, the culture medium was removed and replaced by 1 ml of the Krebs-Ringer bicarbonate-buffered solution used for [Ca²⁺]_i measurement, containing 1.1 mmol/l d-glucose and 0.05% BSA. After 30 min preincubation, the cells were further incubated for 90 min at 37°C in the same medium containing 1.1, 11.1, or 16.7 mmol/l glucose. Aliquots of the culture media were removed from each separate well and stored at −20°C for subsequent insulin measurement by radioimmunoassay (30).

Dynamic insulin release was studied using a multiwell approach as described elsewhere (31). BRIN-BD11 cells were seeded into 48-multiwell plates at a density of 2.5 × 10⁵ per well and allowed to attach overnight at 37°C. The cells were washed and preincubated for 30 min in Krebs-Ringer solution buffered with 10 mmol/l HEPES/NaOH (pH 7.4, 37°C) and supplemented with 0.05% BSA. The preincubation medium was then replaced at 30-s intervals with the same medium containing 11.1 or 16.7 mmol/l glucose to produce a time course of insulin secretion. Thus, the first well was exposed to glucose for 24 min and the last for 30 s. Insulin was simultaneously collected from all the wells by inverting the test plate into a second 48-well plate. The plates were stored at −20°C for subsequent insulin measurement by an enzyme immunoassay, Mercodia Rat Insulin ELISA (Mercodia, Uppsala, Sweden), as described by the manufacturer.

Insulin output was expressed in terms of the numbers of cells in each well at the end of the experiment.

The methods used to measure the production ³HOH from D[ 5-³H] glucose and the generation of ¹⁴CO₂ from d-[U¹⁴C] glucose have been previously described (32–34). Briefly, groups of 10⁵ cells were incubated for 90 min in 80 μl of a bicarbonate-buffered medium. The reaction was stopped by addition of 20 μl of metabolic inhibitors (5 mmol/l KCN, 10 μmol/l Antimycin A, and 10 μmol/l rotenone) in a citrate NaOH buffer (400 mmol/l, pH 4.9).

The results are expressed as means ± SE. The statistical significance of differences between data were assessed by ANOVA followed by Tukey’s post test.

The full-length clone for the PMCA2wb isoform, containing the 135-bp exon at splicing site A but without the 227-bp exon at site C, was constructed from PCR cDNA and stably transfected in BRIN-BD11 cells. Two cell clones expressing high levels of PMCA activity (clones 4 and 5) were used for studies described here. However, similar results were observed with a third clone (clone 2) showing a lower level of overexpression (see below). In addition, three clones transfected with the expression vector lacking the PMCA2wb construct were generated.

The transcription and expression levels of PMCA2wb in transfected cells were determined by RT-PCR method, Western blot analysis, immunofluorescence, and measurement of PMCA activity (ATP-dependent ⁴⁵Ca²⁺ uptake) in membrane preparations of control and PMCA2wb-overexpressing cells.

In both control and overexpressing cells, quantitative RT-PCR performed using primer pairs 7–8 and 9–10 for PMCA2 and β-actin (Table 3) yielded two bands of 300 and 589 bp, respectively (Fig. 1A). Quantitative PCR analysis indicated a four to five times higher transcription level of PMCA2wb in clones 4 and 5 than in control (nontransfected and vector-only transfected) cells.

Membrane proteins from BRIN-BD11 cells and positive clones were analyzed using antibodies specific for PMCA2. In control (nontransfected and vector-only transfected) cells, a faint band at the expected molecular mass range (∼135 kDa) was observed (Fig. 1B, lanes 1 and 2). Immunoblot of PMCA2wb-overexpressing cells showed a markedly increased amount of the 135-kDa protein, with the amount detected for clones 2, 4, and 5 being about two, four, and seven times higher, respectively, than in control cells (Fig. 1B, lanes 3–5). To test the stability of the expression, the clones were maintained in tissue culture with selection antibiotics for up to 9 months (70 passages) without showing any decrease in PMCA2wb expression level.

The overexpression of the PMCA2wb protein at the level of the plasma membrane of the transfected cells was visualized by immunofluorescence using a rabbit anti-PMCA2 antibody. Overexpressing cells (Fig. 2H–J) showed an intense staining predominantly located along the cytoplasmic perimeter of the cells, whereas control (nontransfected and vector-only transfected) cells displayed only a faint fluorescence at their periphery (Fig. 2F and G). The results support the conclusion that the largest fraction of the overexpressed PMCA2wb pump was correctly targeted to the plasma membrane.

Additional support for correct targeting of PMCA2wb came from the assay of PMCA activity using isolated microsomes. An excess concentration of thapsigargin was included in the uptake medium to ensure complete inhibition of SERCA pump activity without affecting the calmodulin-dependent Ca²⁺ uptake mediated by the PMCA (35). Figure 3A illustrates the time courses of ATP-dependent ⁴⁵Ca uptake in control nontransfected and PMCA2wb-overexpressing cells (clone 5). Both cell types displayed a time-dependent increase in ⁴⁵Ca uptake. However, the increase in ⁴⁵Ca uptake was more rapid and marked in overexpressing than control cells, and the rate of ATP-dependent Ca²⁺ uptake by the membrane preparation from the PMCA-transfected cells was about three to four times higher (clones 4 and 5, respectively) than that of control BRIN-BD11 cells (P < 0.001).

A critical parameter for correct functioning of the PMCA is its activation by Ca²⁺ and calmodulin. Therefore, Ca²⁺ transport in microsomes isolated from overexpressing cells was tested as a function of the free Ca²⁺ concentration and the presence of calmodulin. As shown in Fig. 3B, the activity of PMCA2wb in membrane prepared from overexpressing cells (clone 5) was dependent on the Ca²⁺ concentration in the assay medium within the appropriate range (0.01–10 μmol/l). Addition of calmodulin greatly enhanced the activity without changing the sensitivity to ambient Ca²⁺ (Fig. 3B). Compared with control cells, the uptake of ⁴⁵Ca by microsomes from PMCA2wb-overexpressing cells was about four to five times higher in both the presence and absence of calmodulin (P < 0.001).

Taken as a whole, the data presented thus far indicate that PMCA2wb was successfully overexpressed in BRIN-BD11 cells, that the protein was adequately targeted to the plasma membrane, that it was functional, and that its activity showed an appropriate responsiveness to Ca²⁺ and calmodulin.

To evaluate the role played by PMCA on Ca²⁺ homeostasis, we examined the effect of PMCA2wb overexpression on changes in [ Ca²⁺]_i induced by membrane depolarization. Figure 4 illustrates the effect of membrane depolarization induced by K⁺ (50 mmol/l) on [Ca²⁺]_i. First, in this series of experiments, as in another (Fig. 5), basal [Ca²⁺]_i did not differ between control and overexpressing cells and averaged 124 ± 6 nmol/l (n = 350). Second, whether in control (Fig. 5A) or overexpressing cells (clone 4, Fig. 5B; clone 5, Fig. 5C), K⁺ induced a biphasic increase in [Ca²⁺]_i consisting in an initial peak followed by a plateau phase. The increase in [Ca²⁺]_i was rapidly reversible upon removal of K⁺ from the solution. Cells transfected with the vector only showed a K⁺-induced increase in [Ca²⁺]_i similar to that of parental nontransfected cells (P > 0.20) (Fig. 4A). Third, in PMCA2wb-overexpressing cells, the magnitude of the increase in [Ca²⁺] _i was markedly reduced. Thus, a two- and sixfold reduction of the rise in [Ca²⁺]_i was observed in clones 4 and 5, respectively, compared with vector-only transfected cells. Indeed, the increase in [Ca²⁺]_i, as measured by the area under the curve during the time of K⁺ stimulation, averaged 12.3 ± 0.5, 6.5 ± 0.3, and 2.1 ± 0.1 μmol · l⁻¹ · h⁻¹ in vector-only transfected cells, clone 4 cells, and clone 5 cells, respectively (P < 0.0001). In view of the inverse correlation between the level of protein expression and the increase in [Ca²⁺]_i, the effect of K⁺ on clone 2, which shows a lower overexpression, was also tested (Fig. 4D). The result obtained was intermediate between control cells and clone 4 and 5 cells (8.7 ± 0.5 μmol · l⁻¹ · h⁻¹), indicating the existence of an inverse correlation between the level of protein expression and the K⁺-induced increase in [ Ca²⁺]_i, (Fig. 6A). The reduction in the rise in [Ca²⁺]_i can be understood as the result of an increased outward Ca²⁺ transport in PMCA2wb-overexpressing compared with control cells. The results also indicate that there was no difference in PMCA activity between nontransfected and vector-only transfected cells.

To further examine the effect of PMCA2wb overexpression on Ca²⁺ handling in BRIN-BD11 cells, [Ca²⁺]_i changes evoked by 11.1 mmol/l glucose were investigated. This concentration of glucose elicits characteristic cyclic oscillations of the membrane potential and [Ca²⁺] _i in pancreatic β-cells and BRIN-BD11 cells (3,17,36). Figure 5 illustrates typical individual [Ca²⁺]_i responses induced by glucose in control and PMCA2wb-overexpressing cells. In response to 11.1 mmol/l glucose, 40 and 50% of control cells (nontransfected and vector-only transfected cells, respectively) displayed large amplitude [Ca²⁺]_i oscillations (Fig. 5A). In contrast, overexpressing cells never showed such large [Ca²⁺]_i oscillations but instead displayed medium size (clone 4, Fig. 5B) or small size oscillations (clone 5, Fig. 5C). An intermediate situation was observed in clone 2 (data not shown).

A similar picture was observed in response to a higher glucose concentration (16.7 mmol/l). Indeed, in control nontransfected (Fig. 7A) and vector-only transfected cells (Fig. 7B), large amplitude [Ca²⁺] _i oscillations were observed in 46 and 44% of the cells, respectively, while transfected cells did not show such large amplitude oscillations (clone 4, Fig. 7C; clone 5, Fig. 7D). The oscillations induced in control cells by 16.7 mmol/l glucose were less individualized than those evoked by 11.1 mmol/l glucose and tended to occur in bursts (compare Figs. 5 and 7).

A similar difference is seen with respect to β-cell electrical activity; an increase in duration of the bursts of electrical activity has been observed when the concentration of glucose is increased from 11.1 to 16.7 mmol/l (3,17). The data obtained in glucose-stimulated cells are consistent with those obtained in response to K⁺.

⁴⁵Ca uptake was measured over 90 min of incubation in the presence of glucose (1.1 and 11.1 mmol/l, Fig. 8A and 16.7 mmol/l, Fig. 8B). In nontransfected and vector-only transfected cells, 11.1 mmol/l glucose induced a 17 and 21% increase in ⁴⁵Ca uptake, respectively (P < 0.001). In overexpressing cells, basal ⁴⁵Ca uptake (at 1.1 mmol/l glucose) was reduced (by ∼34, 51, and 36% in clones 2, 4, and 5, respectively; P < 0.001) compared with vector-only transfected cells, and a similar effect was observed at 11.1 mmol/l glucose (−34, 45, and 38%, respectively; P < 0.001) (Fig. 8A). Similar results were obtained in response to 16.7 mmol/l glucose (P < 0.001) (Fig. 8B). Thus, overexpression of PMCA2wb leads to a reduction in cellular Ca²⁺ content and rises in [Ca²⁺] _i.

To evaluate the depletion of the intracellular Ca²⁺ stores, we examined the effect of the SERCA inhibitor thapsigargin on [Ca²⁺]_i (Fig. 9). Because thapsigargin releases Ca²⁺ from endoplasmic reticulum (ER) stores, the rise in [Ca²⁺]_i induced by the drug can be taken as an estimate of ER Ca²⁺ content. In control cells (nontransfected or vector-only transfected cells) and at 1.1 mmol/l glucose (Fig. 9A), thapsigargin provoked a transient increase in [Ca²⁺]_i. In overexpressing cells, the rise in [ Ca²⁺]_i was reduced by ∼82% (clone 2, Fig. 9B; clone 4, Fig. 9D) and 100% (clone 5, P < 0.001; Fig. 9C) compared with vector-only transfected cells, indicating a depletion of the ER Ca²⁺ stores.

A similar picture was observed when the experiments were carried out in the presence of 16.7 mmol/l glucose and 10 μmol/l verapamil; the latter drug was added to offset the effect of glucose-induced membrane depolarization (Fig. 10).

The low-affinity Ca²⁺ indicator furaptra was used to monitor free Ca²⁺ in the ER of individual BRIN cells after controlled permeabilization of the plasma membrane, as previously described (29,37).

After recording the fluorescence obtained by excitation at 340 and 380 nm, cells were permeabilized in intracellular medium containing 4 μ mol/l digitonin. Upon the sudden drop in fluorescence caused by the loss of cytoplasmic furaptra, the detergent was immediately removed but the measurement of the fluorescence at both wavelengths was continued. The loss of cytoplasmic furaptra was associated with an inversion of the 340/ 380-nm fluorescence excitation ratio, indicating that the remaining indicator was exposed to higher concentrations of free Ca²⁺ prevailing in intracellular stores (29,37). In a previous study (37), we observed that thapsigargin, an inhibitor of ER Ca²⁺-ATPase, released ∼50% of the Ca²⁺ pool sensed by furaptra under the present condition. Therefore, thapsigargin was used to estimate Ca²⁺ stores in intracellular organelles and more specifically in the ER. In control cells (nontransfected, Fig. 11A and vector-only transfected cells, Fig. 11B), thapsigargin induced a drop of ∼50 and 43% in furaptra 340/380-nm fluorescence excitation ratio, respectively, and the effect of thapsigargin was reduced by 58 ± 7% and 47 ± 5% in PMCA2wb-overexpressing cells (clone 4, Fig. 11C; clone 5, Fig. 11D; P < 0,001) compared with vector-only transfected cells. This indicates that the Ca²⁺ concentration and content of ER stores were markedly reduced in PMCA2wb-overexpressing cells. Incidentally, it is conceivable that part of the difference seen in control and overexpressing cells may result from a reduced capacity of the ER to take up Ca²⁺. Indeed, in view of the presence of ATP in the medium, the ratio reached after permeabilization may indicate ATP-driven uptake by the ER. However, thapsigargin was added immediately after permeabilization; therefore, the difference also results from a decreased accumulation of Ca²⁺ before permeabilization. In a previous study, PMCA overexpression was also shown to reduce ER Ca²⁺ concentration (38).

Insulin release was initially measured over 24 h of static incubation in RPMI-1640 tissue culture medium in the presence of 1.1 and 16.7 mmol/l glucose. The latter concentration of glucose elicits near-maximal stimulation of insulin secretion from BRIN-BD11 cells (18). In control cells (nontransfected or vector-only transfected cells), 16.7 mmol/l glucose induced a two- to threefold increase in insulin release (P < 0.05) (Fig. 12A). In PMCA2wb-overexpressing cells (clones 4 and 5), basal insulin release, measured at 1.1 mmol/l glucose, was increased by ∼80 and 19% in clones 4 and 5, respectively (P < 0.05), and the stimulation in insulin release induced by 16.7 mmol/l glucose was increased by ∼88 and 76% (clones 4 and 5, respectively, compared with vector-only transfected cells; P < 0.05). Similar data were observed in response to 11.1 mmol/l glucose (data not shown). Because the increase in insulin release was of lower magnitude in clone 5 than in clone 4, insulin secretion was also measured in clone 2. In the latter clone (Fig. 12A), insulin release, whether basal or stimulated, was increased by ∼180 and 260% compared with vector-only transfected cells (P < 0.01). The release of insulin obtained in control and transfected cells in response to 16.7 mmol/l glucose is illustrated in Fig. 6B. Thus, all overexpressing cells (clones 2, 4, and 5) showed a larger increase in glucose-induced insulin release than control cells. However, within the overexpressing cells, there was an inverse relationship between PMCA overexpression and insulin release (Fig. 6B).

To relate insulin release and Ca²⁺ measurements, insulin release was measured over a shorter period of time (90 min) (Fig. 12B). The data obtained were similar to those obtained over 24 h. Thus, compared with vector-only transfected cells, basal insulin release was increased by 36, 15, and 8% in clones 2, 4, and 5, respectively (P < 0.05). Likewise, glucose-stimulated insulin release was increased by 60, 23, and 10%, respectively (P < 0.05). Again, there was an inverse relationship between PMCA overexpression and insulin release. To understand such a relationship, glucose metabolism was measured.

Table 4 illustrates glucose utilization and oxidation in control and overexpressing cells. At 1.1 mmol/l glucose, no difference in glucose utilization or oxidation (P > 0.05) was observed between control and overexpressing cells, except that glucose utilization was lower in vector-only transfected cells than in nontransfected cells (P < 0.01). Glucose (16.7 mmol/l) increased glucose utilization (P < 0.001) but failed to increase glucose oxidation (P > 0.06). In overexpressing cells (clones 2, 4, and 5), at 16.7 mmol/l glucose, there was an inverse relationship between PMCA overexpression and glucose utilization (Fig. 6C) that was parallel to that relating PMCA overexpression and insulin release (Fig. 6B).

Last, to better relate cytosolic Ca²⁺ and insulin release oscillations, a multiwell approach was used to measure the time course of insulin release (31). Nontransfected cells were grown in 48-well plates and stimulated by 11.1 and 16.7 mmol/l glucose one well at a time in 30-s intervals (Fig. 13A and C, respectively). The secretion patterns from the wells were linearized by smoothing the data with a three-point moving average, fitting it to a straight line (31), and then determining the positive or negative differences between the data and the straight line (residuals). Plotting the secretion as residuals more clearly shows the oscillatory pattern of insulin release (Figs. 13B and D).

Figure 14 shows insulin release oscillations in response to 11.1 mmol/l glucose (plotted as residuals) in vector-only transfected cells and in the three clones overexpressing PMCA. Vector-only transfected cells showed clear oscillations over 4-min periods (Fig. 14A). Clones 4 and 5 also showed oscillations, but the period of the oscillations was increased to 6 min (P < 0.0001). Surprisingly, clone 2 did not show such oscillations but instead showed an initial peak followed by a plateau phase. Exactly the same picture was observed when the cells were stimulated by 16.7 mmol/l glucose. Thus, vector-only transfected cells showed oscillations over 4-min periods, clones 4 and 5 oscillated over 6-min periods (P < 0.05), and clone 2 showed an initial peak followed by a plateau phase (Fig. 15).

The PMCA belongs to the P-type family of transport ATPases that form a phosphorylated intermediate during the reaction cycle. Four different genes corresponding to four isoforms PMCA1, -2, -3, and -4 have been evidenced. Genes 1 and 4 are transcribed in most tissues, whereas genes 2 and 3 are transcribed in specialized tissues. PMCA1 and -4 may thus represent housekeeping isoforms, while PMCA2 and -3 may be required for specialized functions in a limited number of tissues. Diversity among the ATPases is also generated by alternative splicing of the primary transcripts that may involve two major sites, termed A and C (for review see 19,39,40). In a previous work, we identified the PMCA transcripts expressed in the β-cell and characterized them at alternative splicing sites A and C (20). The identified isoforms were PMCA1xb, -2yb, -2wb, -3za, -3zc, and -4xb. Interestingly, our work was the first to demonstrate the presence, at the protein level, of substantial amounts of PMCA2 and -A3 in non-neuronal tissue (20), a finding that attests the special needs of β-cells in terms of Ca²⁺ homeostasis. Therefore, PMCA2(wb) was chosen to be cloned and overexpressed in BRIN-BD11 cells. The overexpression could be assessed at both the mRNA and protein level. The overexpressed protein was appropriately targeted to the plasma membrane, as shown by immunofluorescence and the increase in ATP-dependent ⁴⁵Ca uptake observed in plasma membrane microsomes. As expected for PMCA-mediated Ca²⁺ transport, ⁴⁵Ca uptake was regulated by the Ca²⁺ concentration in the assay medium within the submicromolar range and by calmodulin without changing the sensitivity to ambient Ca²⁺ (41).

Overexpression of PMCA2wb had substantial effects on [ Ca²⁺]_i and Ca²⁺ homeostasis in BRIN cells. Indeed, the rise in [Ca²⁺] _i induced by extracellular K⁺ was markedly blunted in overexpressing cells, and a positive and inverse correlation was observed between the level of PMCA expression and the area under the curve of [Ca²⁺]_i during K⁺ stimulation of the three clones examined (Fig. 6A). This reduction of K⁺-induced [Ca²⁺]_i rise can best be understood as the result of an increased Ca²⁺ outflow of the cell mediated by the overexpressed PMCA. In agreement with such a view, basal ⁴⁵Ca uptake was reduced and the ER stores appeared to be depleted as assessed by the SERCA inhibitor thapsigargin-induced Ca²⁺ release in fura-2 or furaptra-loaded cells. In addition, control cells transfected with the vector alone (without the PMCA2wb construct) did not show any difference in K⁺-induced [Ca²⁺]_i changes compared with nontransfected parental cells (Fig. 4A). Despite such major changes in cell Ca²⁺ homeostasis recorded in PMCA2wb-overexpressing cells, basal [ Ca²⁺]_i was not reduced. In two other studies of PMCA(1a and -4) overexpression, no change in basal [Ca²⁺]_i was observed (38,41), although a third study evidenced a 20–30% drop in basal [ Ca²⁺]_i in PMCA(4b)-overexpressing myogenic cells (42). Taken as a whole, the present data indicate that PMCA2wb operates within the range of the changes in [Ca²⁺]_i induced by K⁺ (100–1,200 nmol/l) but that although being overexpressed, the PMCA remained regulated and did not decrease [ Ca²⁺]_i below basal levels. Incidentally, the effect of PMCA overexpression on K⁺-induced [ Ca²⁺]_i increase was strikingly different from that of Na/Ca exchanger overexpression, although this protein also contributes to Ca²⁺ outflow from the β-cell (43). Indeed, in Na/Ca exchanger-overexpressing cells, the increase in [Ca²⁺]_i induced by K⁺ was accelerated, as was its decrease on K⁺ removal, but the global rise in [Ca²⁺]_i was rather comparable or even slightly higher than that in control cells (43). This points to the more complex actions exerted by the Na/Ca exchanger, which participates in both Ca²⁺ inflow and outflow and also generates a current due to the stoichiometry of the exchange (3 Na⁺ for 1 Ca²⁺); comparatively, the PMCA only mediates Ca²⁺ outflow (43).

The major observation of the present study was that PMCA overexpression markedly damped glucose-induced Ca²⁺ oscillations while simultaneously increasing the effect of glucose on insulin release. In PMCA-overexpressing cells, as in control cells, stimulation by glucose leads to a progressive increase in basal Ca²⁺, but without the development of large and distinct oscillations seen in control cells. Again, there was a positive and inverse correlation between the level of PMCA expression seen in the different clones and the amplitude of the [ Ca²⁺]_i oscillations induced by glucose. As in the case of K⁺, the decrease in glucose-induced [ Ca²⁺]_i oscillations can best be understood as resulting from an increased Ca²⁺ outflow from PMCA-overexpressing cells. Despite this reduction in glucose-induced [ Ca²⁺]_i oscillations, insulin release was not reduced but instead was increased. Such a phenomenon was observed whether the cells were exposed to 11.1 or 16.7 mmol/l glucose during during short (90 min) or long periods of time (24 h). Our data indicate that [Ca²⁺]_i oscillations are unnecessary for insulin release, even at an intermediate glucose concentration (11.1 or 16.7 mmol/l), and that their presence may even lead to a less efficient insulin secretory process compared with that induced by a stable increase in [Ca²⁺]_i.

It has been proposed that the rapidly descending phases of mouse β -cell [Ca²⁺]_i oscillations result from the outward transport of Ca²⁺ from the cell, a finding that is compatible with the view that glucose stimulates such an outward transport (7). Consequently, it can be expected that enhanced removal of Ca²⁺ from the cytoplasm would prevent the establishment of the critical [ Ca²⁺]_i required for a feedback inhibition of glucose-stimulated entry of the ion (7). Such a view is consistent with the observation that the acceleration of the outward transport of Ca²⁺ by protein kinase activation suppresses the slow [Ca²⁺]_i oscillations seen in mice (7). Thus, overexpression of PMCA, by preventing the establishment of the critical [Ca²⁺]_i required for a feedback inhibition of glucose-stimulated entry of the ion, would suppress periods during which [Ca²⁺] _i drops to or below the basal level and would hence lead to an improved insulin release. In fact, PMCA overexpression may simply offset part of the consequences resulting from the use of a high extracellular Ca²⁺ concentration. Therefore, the oscillations in electrical activity and [Ca²⁺]_i seen in mouse at high [Ca²⁺]_o may not represent a physiological model for the study of the process of insulin release. In this respect, it is interesting to notice that Gilon et al. (5), in order to study the synchronicity between [Ca²⁺]_i and insulin release oscillations, had to use an extremely elevated concentration of extracellular Ca²⁺ (10 mmol/l). Recently, Ravier et al. (44) studied the effect of “imposed” or “forced” oscillations of cytoplasmic Ca²⁺ or metabolism on insulin secretion in mouse islets. However, the protocol used in the latter study corresponds to intermittent stimulations of Ca²⁺ rises and not to real or spontaneous Ca²⁺ oscillations, as could be explored in the present study using PMCA2wb-overexpressing cells.

Insulin release was increased in all three clones overexpressing PMCA compared with control cells. However, within the three clones, there was an inverse relationship between insulin release and PMCA level of expression. This inverse relationship was paralleled by an inverse relationship between PMCA expression and glucose utilization at 16.7 mmol/l, indicating that the difference in insulin release resulted from a difference in glucose utilization among the three clones. In a recent study, Jung et al. (45) correlated the temporal changes of glucose level, oxygen level, and [Ca²⁺]_i in single islets. They observed that although rises in [Ca²⁺]_i preceded pulses in oxygen consumption during oscillations, [Ca²⁺] _i increase was associated with decreases in glucose consumption through glycolysis (45). Our present data showing that PMCA-overexpressing cells, which display inhibited rises in [Ca²⁺]_i and a higher rate of glucose utilization, are in agreement with the latter data. They are also in agreement with those of Detimary et al. (46), who showed that a rise in [Ca²⁺]_i causes a decrease in ATP/ADP ratio, a phenomenon that probably results from the activation of ATPases involved in the maintenance of Ca²⁺ and Na⁺ homeostasis in β-cells or insulin exocytosis. Incidentally, in contrast with the situation found in normal islets (47), a rise in glucose concentration did not cause any increase in glucose oxidation in BRIN-BD11 cells, a situation reminiscent of that found in RINm5F, where glucose even causes, through a Crabtree effect, inhibition of glucose oxidation (48). BRIN-BD11 cells result from the fusion of normal β-cells with RINm5F cells, and the metabolic situation found in BRIN-BD11 cells, which show no change in glucose oxidation, is intermediate between that of normal β-cells and RINm5F cells (49).

What is not clear from the present study, and warrants further study, is the reason why there was an inverse relationship between PMCA expression level and glucose utilization in overexpressing cells. Although increases in [Ca²⁺]_i may be associated with decreases in glucose consumption, it is conceivable that a decrease in [Ca²⁺]_i that is too large may have the opposite effect, perhaps through direct Ca²⁺ effects on glycolytic enzymes (50, 51).

The present data confirm the observation of Deeney et al. (31) showing that clonal β-cells grown in separate wells display oscillations in insulin release in response to glucose. This indicates that the synchrony that must exist between β -cells to lead to oscillations in insulin release is apparently not mediated, or not exclusively mediated, by either electrical coupling or a putative soluble synchronizing factor that is possibly secreted by some or all cells of the population (31). It is conceivable that the synchrony of the cells in separate wells results from the similar metabolic effects of glucose on the different groups of cells (31). Interestingly, oscillations were in 4-min periods, which is similar to the findings of Deeney et al. (31) in HIT cells.

Another major observation of the present study was that insulin release oscillations induced by glucose (11.1 and 16.7 mmol/l) persited in PMCA-overexpressing cells (clones 4 and 5) despite the absence, in the latter cells, of [Ca²⁺]_i oscillations. This indicates that even at an intermediate glucose concentration, [ Ca²⁺]_i oscillations are not necessary for insulin release oscillations, which are apparently directly driven by oscillations of glucose metabolism (12–15). Interestingly, the periodicity of the oscillations was changed from 4 to 6 min in overexpressing cells, which may indicate that [Ca²⁺]_i oscillations and/or intracellular Ca²⁺, although not directly driving insulin release oscillations, play a regulatory role of such oscillations. Interestingly, no oscillations in insulin release was observed in the clone showing the largest insulin response to both 11.1 and 16.7 mmol/l glucose (clone 2). It is conceivable that a change in intracellular Ca²⁺ homeostasis, as induced by PMCA overexpression, may not only regulate the periodicity of insulin release oscillations but may also lead to their disappearance, with a subsequent increase in total insulin release.

In conclusion, our data show that overexpression of PMCA2wb inhibits glucose-induced oscillations of [Ca²⁺] _i, a phenomenon that is accompanied by an increase in glucose-induced insulin release. Hence, [Ca²⁺] _i oscillations appear unnecessary for glucose-induced insulin release and may be less favorable than a stable increase in [ Ca²⁺]_i for optimal hormone secretion, and large increases in [Ca²⁺]_i, occurring during [Ca²⁺]_i oscillations, lead to inhibition of both β-cell glucose metabolism and insulin release. In addition, oscillations of [Ca²⁺]_i do not directly drive insulin release oscillations but may nevertheless intervene in the fine regulation of such oscillations. It would be of interest to confirm such observations in normal pancreatic β-cells.

This work was supported by the Belgian Fund for Scientific Research (FRSM 3.4562.00), of which F.V.E. is a Postdoctoral Researcher.

The authors thank P. Lebrun for help in insulin release measurements, A. Sener (Laboratory of Experimental Medicine, Brussels) for help in the measurement of glucose metabolism, and C. Pastiels and A. Van Praet for technical help.