Theoretical models and indirect experimental observations predict that Ca2+ concentrations at the inner surface of the plasma membrane may reach, upon stimulation, values much higher than those of the bulk cytosol. In the past few years, we have shown that the Ca2+‐sensitive photoprotein aequorin can be intracellularly targeted and utilized for specifically monitoring the [Ca2+] of various organelles. In this work, we extend this approach to the study of the cytoplasmic rim beneath the plasma membrane. We have constructed a new aequorin chimera by fusing the photoprotein with SNAP‐25, a neuronal protein which is recruited to the plasma membrane after the post‐translational addition of a lipid anchor. The SNAP‐25–aequorin chimera, expressed in the rat aortic smooth muscle cell line A7r5, appears correctly sorted as revealed by immunocytochemistry. Using this probe, we demonstrate that the mean [Ca2+] of this cytoplasmic region ([Ca2+]pm) can reach values >10‐fold higher than those of the bulk cytosol ([Ca2+]c) upon activation of Ca2+ influx through plasma membrane channels. In unstimulated cells, the mean [Ca2+]pm appears also to be higher than the bulk cytosol, presumably reflecting the existence of microdomains of high [Ca2+].
The rim of cytoplasm immediately beneath the plasma membrane is an important site of action of intracellular Ca2+ ions. In this restricted environment, a number of Ca2+‐dependent processes are activated, such as secretion, ion channel activity, production of second messengers, etc. Based, however, on the proximity to the Ca2+ channels, the Ca2+ changes under the plasma membrane are expected to be quite different from those measured in the rest of the cytoplasm. Pioneering work using the low Ca2+ affinity n‐aequorin‐J has shown that microdomains of [Ca2+] in the range 200–300 μM occur in the presynaptic terminal of the giant squid synapse during transmitter release (Llinás et al., 1992, 1995; Silver et al., 1994). This [Ca2+] appears to be necessary to trigger the exocytotic machinery efficiently (Heidelberger et al., 1994). Apart from these models, however, direct evidence for a correlation between local Ca2+ changes and effects on cell function has been very difficult to obtain. In fact, while the Ca2+ dependence of various processes occurring at the cell surface has been investigated thoroughly in a large number of cell models, the measurement of Ca2+ concentration in this restricted cytosolic space is a still unsolved experimental task.
We have recently described the targeting of the Ca2+‐sensitive photoprotein aequorin to the lumen of various cellular organelles, i.e. nucleus (Brini et al. 1993), mitochondria (Rizzuto et al., 1992, 1993, 1994) and endoplasmic reticulum (Montero et al., 1995), and the use of these targeted Ca2+ probes for a subcellular analysis of Ca2+ homeostasis. The same approach was extended here to the study of the subplasmalemmal space. Indeed, compared with lipid‐bound fluorescent indicators (Etter et al., 1994, 1996), a targeted photoprotein can be expected to exhibit a more selective localization and, by virtue of its low Ca2+ affinity (∼10 and ∼100 μM for wild‐type aequorin and the Asp119→Ala aequorin mutant, respectively), a minor Ca2+‐buffering capacity.
Since the minimal sequence requirements for the sorting to the plasma membrane are still not completely known, the targeting of aequorin to the subplasmalemmal space was based on the construction of a fusion protein including aequorin and SNAP‐25, a protein which is synthesized on free ribosomes and recruited to the inner surface of the plasma membrane after the palmitoylation of specific cysteine residues (Hess et al., 1992). Moreover, as the [Ca2+] of the subplasmalemmal region ([Ca2+]pm) could be expected to be, at least transiently, very high, we have constructed two SNAP‐25–aequorin chimeras, differing only in the aequorin moiety: wild‐type aequorin in the chimera designated pmAEQwt, the Asp119→Ala mutant, endowed with lower Ca2+ affinity, in the chimera named pmAEQmut, allowing measurements of [Ca2+] up to 150 μM. By expressing these chimeras in the A7r5 line, we were able to monitor selectively [Ca2+]pm and demonstrate that, in unstimulated cells, the average [Ca2+] is apparently >10‐fold higher than in the bulk cytosol and can reach, upon Ca2+ influx through plasma membrane channels, values as high as 100 μM.
Construction and expression of the pmAEQ chimera
The start points for the construction were the wild‐type SNAP‐25 cDNA and two different HA1‐tagged aequorin cDNAs: the wild‐type aequorin (Brini et al., 1985; Inouye et al., 1995) and the Asp119→Ala mutant (Montero et al., 1995), endowed with lower Ca2+ affinity. Since a ClaI site is present at the 5′ end of the aequorin cDNAs, through a PCR step a ClaI site was inserted immediately upstream of the SNAP‐25 stop codon; the PCR‐modified SNAP‐25 cDNA was then fused to either the wild‐type aequorin cDNA (to generate the chimera designated pmAEQwt) or the low‐affinity mutant (generating the chimera designated pmAEQmut). Figure 1 shows a schematic map of the cDNA construct, while the details of the construction are discussed in Materials and methods. The chimeric cDNAs were subcloned in the expression vector pcDNAI (InVitrogen), and used for transfecting the cell line A7r5 (Kimes and Brandt, 1976). The intracellular distribution of transiently expressed pmAEQ and, for comparison, cytosolic aequorin, cytAEQ (Brini et al., 1995), was verified by immunostaining. Figure 2 shows confocal analysis of transfected cells. Three confocal sections of pmAEQ (Figure 2A–C) and cytAEQ (Figure 2E–G) are presented. In Figure 2A and B, a clear peripheral distribution of pmAEQ can be observed, with fine digitations appearing most strongly labelled; in Figure 2C, a diffuse labelling of the top surface of the cell can be appreciated. Conversely, the immunostaining of cytAEQ (Figure 2E–G) appears diffused to the whole cytoplasm, with partial nuclear exclusion. To further confirm the intracellular localization of the probes, confocal microscopy was performed on cells transiently expressing pmAEQ and cytAEQ after permeabilization with 100 μM digitonin for 2 min prior to fixation. In Figure 2D, a clear staining of the plasma membrane is still evident in cells expressing pmAEQ while, under the same conditions (Figure 2H), no staining was observed in cells expressing cytAEQ.
Native aequorin is composed of an apoprotein and a covalently bound co‐enzyme, coelenterazine. The reconstitution of a functional photoprotein is accomplished by incubating the cell expressing recombinant aequorin with the membrane‐permeant prosthetic group coelenterazine (Rizzuto et al., 1995). The efficiency of this procedure depends on the relative rates of aequorin reconstitution and consumption. In cell compartments endowed, at rest, with low Ca2+ concentrations (such as the cytoplasm, the nucleus or the mitochondria), reconstitution of recombinant aequorin is highly effective (Rizzuto et al., 1992, 1994; Brini et al., 1993, 1995), while in an organelle with high [Ca2+], the endoplasmic reticulum, a good reconstitution can be obtained only upon drastic reduction of the lumenal Ca2+ concentration (Montero et al., 1995). Table I shows the efficiency of reconstitution of pmAEQ and cytAEQ in media with or without extracellular Ca2+ (expressed as a percentage of maximal). In contrast to cytAEQ, the reconstitution of the pmAEQ chimeras was reduced drastically in media containing physiological Ca2+ concentrations (1 mM). This effect was more pronounced with pmAEQwt (11 ± 1%) than with pmAEQmut (21 ± 4%), suggesting, albeit indirectly, that the [Ca2+]pm is, in unstimulated cells, much higher than in the rest of the cytoplasm.
Ca2+ measurements with aequorin
Effect of Ca2+ readdition. In all experiments, measurements of [Ca2+]c and [Ca2+]pm were performed in monolayers of cells transiently transfected (Rizzuto et al., 1995) with cytAEQ or pmAEQ, after a 45 min reconstitution with coelenterazine in Ca2+‐free Krebs–Ringer buffer (KRB). As shown in Figure 3A, in cells expressing cytAEQ, light emission, initially close to background values, underwent a minor increase upon addition of 1 mM CaCl2; >95% of the total aequorin pool was discharged only by lysing the cells at the end of the experiment. The corresponding calibrated signal (Figure 3B) showed that [Ca2+]c was 0.2 μM in Ca2+‐ free medium and marginally increased upon Ca2+ readdition. Figure 3C shows that, with pmAEQwt, aequorin light emission was also close to background values in Ca2+ ‐free medium, but underwent a very large transient rise when the perfusing buffer was supplemented with 1 mM CaCl2 (>90% aequorin was consumed within ∼2 min). The calibration of the aequorin signal into Ca2+ values (Figure 3D) showed a large transient increase (peak >8 μM), which rapidly declined towards basal. Figure 3E shows that also with pmAEQmut, a major fraction of aequorin (90% in 2 min) was consumed upon Ca2+ addition. Indeed, the calibration of the aequorin signal (Figure 3F) showed a transient [Ca2+]pm increase which peaked at 120 μM. It is noteworthy that the rates of photon emission with pmAEQwt and pmAEQmut are quite similar, suggesting that, under these conditions, the increase in [Ca2+] of the subplasmalemmal region is so high as to saturate both probes. Accordingly, the values of [Ca2+]pm would be largely artefactual. In order to get a more quantitative measurement of the changes in [Ca2+]pm, the cells were thus challenged with a much lower extracellular Ca2+ concentration.
Figure 4A and C shows the increase in aequorin light emission induced by the addition of 100 μM CaCl2 to cells transiently expressing pmAEQwt and pmAEQmut, maintained in Ca2+‐free medium. Figure 4A and C shows that the increase in light emission induced by Ca2+ readdition (which induces a negligible change of [Ca2+]c, not shown) was quite different with the two pmAEQ chimeras: 60–70% of total aequorin was consumed in 2 min with pmAEQwt (Figure 4A) compared with 10–20% with pmAEQmut (Figure 4C). The corresponding calibrated [Ca2+]pm (Figure 4B and D) was similar with the two probes, in terms of both kinetics and peak amplitudes, i.e. 8 ± 2 μM (n = 5) with pmAEQwt (Figure 4B) and 10 ± 1.5 μM (n = 5; Figure 4D) with pmAEQmut. The steady‐state [Ca2+]pm (2.0 ± 0.5 μM), as measured with pmAEQmut, was reached in 2 min after Ca2+ readdition. This value was significantly higher than basal [Ca2+]c and lasted until Ca2+ was removed from the medium.
Effect of Ca2+ store depletion. While Ca2+ readdition to cells incubated in Ca2+‐free medium causes minor changes in [Ca2+]c, much larger [Ca2+]c increases can be elicited by the activation of store‐operated Ca2+ influx (Byron and Taylor, 1995). We have thus investigated the effect of Ca2+ store depletion on the [Ca2+]pm. The corresponding [Ca2+]c data are shown in Figure 5A. Transfected cells were first treated with the inositol 1,4,5 trisphosphate (IP3)‐generating agonist vasopressin, thus evoking a transient [Ca2+]c rise (peak 1.4 μM); Ca2+ influx was then triggered by adding 100 μM CaCl2 to the perfusing medium. Ca2+ readdition caused a second increase of [Ca2+]c at 0.8 μM and a sustained plateau at ∼0.4 μM; finally, upon addition of EGTA, [Ca2+]c returned to basal levels. With regard to the [Ca2+]pm, the release of stored Ca2+ caused a [Ca2+]pm rise quite similar in amplitude and kinetics to that of the cytosol (demonstrating that the Ca2+ affinity of the pmAEQ is not affected by the fusion of the photoprotein with SNAP‐25 and/or by the local microenvironment). Similar results were obtained by releasing the internal stores with caffeine (data not shown). Conversely, a very large and rapid rise was observed upon readdition of Ca2+, both with pmAEQwt (not shown) and with pmAEQmut (Figure 5B). The peak value, while close to the saturation of the probe with pmAEQwt, could be estimated to be 40 ± 9 μM (n = 5) using pmAEQmut.
These conditions (readdition of Ca2+ to depleted cells) were aimed at magnifying store‐operated Ca2+ influx. When Ca2+ influx was activated under more physiological conditions, i.e. the IP3‐generating agonist was administered in the presence of external Ca2+, the results were quite different. Cells transfected with either cytAEQ or pmAEQ were first exposed to 100 μM Ca2+ until a steady‐state was reached and were then challenged with vasopressin. Figure 5C shows that, after a marginal increase upon Ca2+ readdition, stimulation with vasopressin induced a transient [Ca2+]c increase, which peaked at 1.2 μM. Figure 5D shows that [Ca2+]pm, monitored with pmAEQmut, after the large overshoot upon Ca2+ readdition (peak ∼12 μM) underwent, upon stimulation with vasopressin, a rise from the plateau value of 2 μM to ∼5 μM. This increase, larger than that of [Ca2+]c, was, however, >10‐fold smaller than that evoked by Ca2+ readdition to Ca2+‐depleted cells (Figure 3), probably reflecting the known inhibitory effect of Ca2+ on store‐dependent Ca2+ current (Hoth and Penner, 1993; Zweifach and Lewis, 1995; Skutella and Ruegg, 1996). Finally, the addition of EGTA caused the return of [Ca2+]pm to a level indistinguishable from that of [Ca2+]c.
The possibility of measuring the [Ca2+] in the subplasmalemmal space, a cytosolic region which plays a key role in the control of Ca2+ homeostasis and in a number of physiological processes (secretion, ion and metabolite transport, etc.), is a major goal in the study of cell physiology. However, the methodologies now available to achieve this goal are hampered by technical and theoretical limitations. Indeed, given the spatial resolution of optical microscopes (under ideal conditions that of incident light, i.e. 300–500 nm), at best a cytosolic fluorescent Ca2+ indicator, such as fura‐2 (Grynkiewicz et al., 1985), will only measure the average of a very steep gradient. In an attempt to overcome this limitation, fluorescent Ca2+ indicators have been anchored to the plasma membrane via the addition of a hydrophobic fatty acid tail (Etter et al., 1994, 1996). The kinetics of [Ca2+] increase measured by lipid‐bound Ca2+ dyes upon opening of plasma membrane channels were found to be much faster than those recorded with the same indicator loaded in the cytosol, though, surprisingly, the peak increases were not very different. Here we present the development and the use of a specifically targeted aequorin chimera. Moreover, through the use of aequorin moieties with different Ca2+ affinities, a wide range of [Ca2+] can be measured accurately.
With this selective probe, the study of the [Ca2+] dynamics in the rim of cytoplasm immediately beneath the plasma membrane revealed an interesting, and in part unexpected, scenario. In particular, the mean [Ca2+]pm appears much higher than that of the bulk cytosol both when Ca2+ influx is activated and in unstimulated cells. This latter conclusion is supported by both the reduced reconstitution of the pmAEQ chimeras in Ca2+‐containing media (Table I) and by the direct monitoring of the steady‐state [Ca2+]pm (see Figure 4). However, the maintenance of a gradient between [Ca2+]pm and bulk [Ca2+]c would imply a continuous flux of Ca2+ into the cytosol, whereas in resting cells there is no net flux of Ca2+ across the plasma membrane. Two non‐mutually exclusive possibilities could thus account for the high [Ca2+]pm values. [Ca2+]pm could be heterogeneous also under resting conditions. In particular, hotspots could be generated transiently in the proximity of Ca2+ channels during spontaneous activity. The local [Ca2+] would lead to activation of Ca2+ extrusion, and, on average, the net flux would be zero. The pmAEQ signal would average the [Ca2+]pm values but, since the rate of photon emission of aequorin is proportional to the 2nd–3rd power of [Ca2+], the mean signal would be dominated by the regions at high Ca2+. Alternatively, the negative surface potential of the inner leaflet of the plasma membrane, which was estimated in other cell systems to be ∼20–30 mV (Hille et al., 1975), could account for a steady‐state difference in the resting [Ca2+]c and [Ca2+]pm levels. However, this interpretation appears unlikely to us for at least two reasons: (i) while transient microdomains might have escaped the spatial resolution of previous studies using lipid‐bound fura‐2, a homogeneous [Ca2+]pm of 1–2 μM should have been readily revealed as a rim of high [Ca2+] at the cell periphery; and (ii) the electrophysiological characterization of the basal activity of Ca2+‐dependent channels argues against this model. For instance, Ca2+‐dependent K+ channels present in A7r5 cells are activated by a [Ca2+] ranging from 0.1 to 1 μM (Van Rentherghem and Lazdunski, 1992) and, therefore, would be fully activated by a mean [Ca2+] of 1–2 μM underneath the plasma membrane. Such an activation is not found in resting cells.
Even more evident is the difference between [Ca2+]c and [Ca2+]pm upon activation of Ca2+ influx. The simple manoeuvre of shifting the cells from Ca2+‐free media to media containing 1 mM Ca2+ induces such a dramatic change in [Ca2+]pm that, due to the rapid consumption of the aequorin pool, only a lower limit (∼120 μM) of the real peak value can be obtained. Only when a 10‐fold lower [Ca2+] is added to the cells can the amplitude (∼10 μM) and the kinetic behaviour of the [Ca2+]pm transient be measured accurately. Under these latter conditions, it is also apparent that the [Ca2+]pm rise caused by Ca2+ readdition is ∼5‐fold increased if the Ca2+ influx pathway dependent on depletion of IP3‐sensitive stores (Hoth and Penner, 1992; Fasolato et al., 1994; Clapham, 1995) is activated previously. Interestingly, despite the low [Ca2+] in the extracellular medium, the peak values of the [Ca2+]pm increases appear much higher than the predicted values at the mouth of store‐dependent Ca2+ channels (Zweifach and Lewis, 1995) and approach those estimated for the active zones of presynaptic terminals (Heidelberger et al., 1994). In conclusion, the present data indicate that the activity of Ca2+ channels induces very high [Ca2+] increases (>100 μM) not only in specialized domains but also in wide regions under the plasma membrane. The described experimental approach, which may in the future also allow the targeting of aequorin chimeras to the mouth of a Ca2+ channel, now offers the possibility of quantitatively investigating the role of these [Ca2+] changes in modulating the biological events occurring at the inner surface of the plasma membrane.
Materials and methods
Construction of the SNAP‐25–aequorin chimeras
The chimeric cDNAs was constructed as follows. A PCR run was performed using the SNAP‐25 cDNA as a template and the following primers: forward, GAATTCTACCATGGCCGAGGACGCAGACA‐TGCGTAAT; reverse, ATCGATACCACTTCCCAGCATCTTTCT. The forward primer specifies an EcoRI site immediately in front of nucleotide −4 of the SNAP‐25 cDNA, whereas at the 5′ end of the reverse primer a ClaI site substitutes the native stop codon of the cDNA. After 30 PCR cycles (95°C 1 min, 62°C 2 min, 72°C 1 min), a single PCR product was identified, isolated and subcloned according to standard procedures into the SmaI site of PBS+ (Stratagene). The PCR‐modified SNAP‐25 cDNA (pSNAP‐25) was then excised from the vector by a ClaI–KpnI digestion and inserted into appropriately cut pBSK+ plasmids, containing the aequorin cDNAs. For the construction of pmAEQwt, pSNAP‐25 was subcloned upstream of a HindIII–EcoRI fragment encoding HA1‐tagged aequorin (Brini et al., 1995). For the construction of pmAEQmut, the pSNAP‐25 was subcloned upstream of a HindIII–EcoRI fragment encoding the Asp119→Ala aequorin mutant (Montero et al., 1995). In both cases, fusion occurred in‐frame, as verified by DNA sequencing. The entire inserts (pmAEQwt and pmAEQmut) were then excised via EcoRI digestion and inserted into the expression vector pcDNAI (Stratagene). The recombinant plasmids (pmAEQwt/pcDNAI and pmAEQmut/pcDNAI) were utilized for the transfections as previously described (Rizzuto et al., 1995).
Cell culture and transfection
Rat aortic smooth muscle cells of the A7r5 cell line (Kimes and Brandt, 1976) (passages 15–25) were grown in Dulbecco's modified Eagle's medium, supplemented with 10% fetal calf serum, 2 mM glutamine, penicillin (200 U/ml) and streptomycin (0.2 mg/ml) in 25 cm2 Falcon flasks. For transient expression experiments, the cells were seeded onto 13 mm coverslips and allowed to grow to 75% confluence. At this stage, transfection with the recombinant plasmids cytAEQ, pmAEQwt and pmAEQmut (4 μg DNA/coverslip) was carried out as previously described (Rizzuto et al., 1995).
For aequorin reconstitution, as indicated in the text, the cells were transferred to modified KRB (125 mM NaCl, 5 mM KCl, 1 mM Na3PO4, 1 mM MgSO4, 5.5 mM glucose, 20 mM HEPES, pH 7.4, 37°C), supplemented with 0.1 mM EGTA (in some experiments EGTA was omitted, with no significant difference in the count yield) and 5 μM coelenterazine. After 45 min, the coverslip was transferred to the luminometer chamber, and perfused with KRB containing 0.1 mM EGTA. All experiments were performed in KRB medium, supplemented, where indicated, with 0.1 mM EGTA, 100 μM or 1 mM CaCl2. The experiments were terminated by lysing the cells with 100 μM digitonin in a hypotonic Ca2+‐rich solution (10 mM CaCl2 in H2O), thus discharging the remaining aequorin pool. The light signal was collected and calibrated into [Ca2+] values as previously described (Brini et al., 1995; Rizzuto et al., 1995). In brief, the transfected cells were placed in a perfused, thermostatted chamber placed in the close proximity of a low‐noise photomultiplier, with built‐in amplifier–discriminator. The output of the discriminator was captured by a Thorn‐EMI photon counting board, and stored in an IBM‐compatible computer for further analyses. The aequorin luminescence data were calibrated off‐line into [Ca2+] values, using a computer algorithm based on the Ca2+ response curve of wild‐type and mutant aequorins, as previously described (Brini et al., 1995; Montero et al., 1995).
Immunolocalization of the HA1‐tagged recombinant aequorin
At 48 h after transfection, A7r5 cells were processed for immunofluorescence as follows. The cells were fixed with 3.7% formaldehyde in phosphate‐buffered saline (PBS) for 30 min, washed three times with PBS and then incubated for 30 min in PBS supplemented with 50 mM NH4Cl. Permeabilization was achieved with a 5 min incubation with 0.1% Triton X‐100 in PBS, followed by a 1 h wash with 1% gelatin (type IV from calf skin) in PBS. The cells were then incubated for 1 h at 37°C in a wet chamber with a 1:100 dilution (in 1% gelatin in PBS) of the monoclonal antibody 12CA5 (initially a kind gift from Dr Jacques Pouysségur, Nice, France; then obtained from BAbCo, Berkeley, CA), which recognizes the HA1 tag (Field et al., 1988). The cells were then incubated for 30 min in a wet chamber with a 1:50 dilution (in 1% gelatin in PBS) of the tetramethylrhodamine (TRITC)‐labelled anti‐mouse IgG secondary antibody. After each antibody incubation, the cells were washed four times with PBS. Fluorescence was then analysed with a 100× objective on a RCM8000 real‐time Nikon confocal microscope using a krypton–argon ion laser.
We thank G.Ronconi and M.Santato for technical assistance, A.Negro for kindly providing the SNAP‐25 cDNA and P.Pacaud and C.Frelin for the gift of the A7r5 cell line. This work was supported by grants from ‘Telethon’, from the EU programs ‘Biomed2’, ‘Human Capital and Mobility’ and ‘Copernicus’, from the ‘Human Frontier Science Program’, from the Italian Research Council (CNR) Special Project ‘Oncology’, from the Italian University Ministry and from the British Research Council to T.P. and R.R. R.M. was supported by a EU fellowship ‘Human Capital and Mobility’.
- Copyright © 1997 European Molecular Biology Organization