Agonists induce Ca2+ spikes, waves and oscillations initiating at a trigger zone in exocrine acinar cells via Ca2+ release from intracellular Ca2+ stores. Using a low affinity ratiometric Ca2+ indicator dye, benzothiazole coumarin (BTC), we found that high concentrations of agonists transiently increased Ca2+ concentrations to the micromolar range (>10 μM) in the trigger zone. Comparison with results obtained with a high affinity Ca2+ indicator dye, fura‐2, indicated that fura‐2 was in fact saturated with Ca2+ during the agonist‐induced Ca2+ spikes in the trigger zone. We further revealed that the micromolar Ca2+ spikes were necessary for inducing exocytosis of zymogen granules investigated using capacitance measurements. In contrast, submicromolar Ca2+ spikes selectively gave rise to sequential activation of luminal and basal ion channels. These results suggest new functional diversity in Ca2+ spikes and a critical role for the micromolar Ca2+ spikes in exocytotic secretion from exocrine acinar cells. Our data also emphasize the value of investigating the Ca2+ signalling using low affinity Ca2+ indicators.
The final Ca2+‐dependent steps of exocytosis in various types of cells recently have been shown to exhibit Ca2+ affinities in the micromolar range (Neher and Zucker, 1993; Thomas et al., 1993; Heidelberger et al., 1994; Muallem et al., 1995; Kasai et al., 1996). An extreme case is exocytosis involving synaptic vesicles in which the Kd for Ca2+ is expected to be 194 μM (Heidelberger et al., 1994). The synaptic vesicles are co‐localized with voltage‐gated Ca2+ channels in the plasma membrane, and the rate constants for Ca2+ binding to putative Ca2+ sensors are large. Therefore, large increases in cytosolic Ca2+ concentration ([Ca2+]i) at the inner mouth of a Ca2+ channel can be detected by the Ca2+ sensors before calcium ions bind to Ca2+ binding molecules in the cytosol. The high [Ca2+]i area formed by a Ca2+ channel is called a ‘Ca2+ domain’ (Chad and Eckert, 1984; Smith and Augustine, 1989), which is too small to be visualized by light microscopic Ca2+ imaging techniques.
Exocytosis is also triggered by Ca2+ release from intracellular Ca2+ stores which often induces a complex spatio‐temporal pattern of [Ca2+]i increases. The exocrine acinar cell is an excellent system for investigating the quantitative relationship between Ca2+ release and exocytosis, because most agonist‐induced [Ca2+]i increases are mediated by Ca2+ release from intracellular Ca2+ stores (Petersen et al., 1994), and they occurred in a fixed spatial pattern (Kasai et al., 1993). It has been reported that exocytotic secretion of enzyme from exocrine acinar cells requires micromolar [Ca2+]i (Knight and Koh, 1984; Cher et al., 1992; Stecher et al., 1992; Muallem et al., 1995; Padfield and Panesar, 1995). However, no Ca2+ imaging study has revealed large increases in [Ca2+]i that can account for exocytosis of zymogen granules in acinar cells (Kasai and Augustine, 1990; Tan et al., 1992; Toescu et al., 1992; Kasai et al., 1993; Maruyama et al., 1993; Thorn et al., 1993; Gerasimenko et al., 1996). There are two possible reasons for the discrepancy. First, the zymogen granules might utilize Ca2+ in the Ca2+ domains at the inner mouth of Ca2+ release channels as synaptic vesicles do for voltage‐gated Ca2+ channels. Second, there may be an area of a large [Ca2+]i increase in acinar cells that was not detected in the previous studies because high affinity Ca2+ indicators, such as fura‐2, that could be saturated with Ca2+ at a high [Ca2+]i, were used.
We therefore performed a Ca2+ imaging study using a low affinity Ca2+ indicator dye, benzothiazole coumarin (BTC) (Iatridou et al., 1994) and a sensitive cooled‐CCD camera that allowed us to capture images with a higher resolution (12 bits) (Messler et al., 1996) than in previous studies (8 bits) (Kasai and Augustine, 1990; Toescu et al., 1992). Ca2+ imaging with BTC revealed that agonists increased [Ca2+]i to levels >10 μM at the apical pole of the secretory granule area, called a ‘trigger zone’ (Kasai et al., 1993). The occurrence of these micromolar Ca2+ spikes appears to have a critical functional consequence: exocytosis in acinar cells requires [Ca2+]i increases to levels >5 μM during the agonist‐induced Ca2+ spikes as well as homogeneous [Ca2+]i increases that were produced by photolysis of a caged‐Ca2+ compound, nitrophenyl‐EGTA (Ellis‐Davies and Kaplan, 1994). Interestingly, lower concentrations of agonists induce submicromolar Ca2+ spikes which selectively activate Ca2+‐dependent ion channels, but do not induce the exocytosis. Thus, the distinct cellular functions are regulated by increases in [Ca2+]i that can be measured using a Ca2+ imaging technique and whose physiological dynamic range is larger (0.1–15 μM) than hitherto considered.
Micromolar Ca2+ spikes in pancreatic acinar cells
We examined the distribution of [Ca2+]i in mouse pancreatic acinar cells using the low affinity Ca2+ indicator dye BTC (KCa = 10 μM) (Iatridou et al., 1994), which was loaded into the acinar cells via a patch pipette. A highly sensitive cooled‐CCD camera with a 12 bit resolution (Messler et al., 1996) was utilized to detect changes in fluorescence of BTC in a range of [Ca2+]i between 0.3 and 100 μM (Materials and methods). Figure 1A shows the results of one such experiment. The dark areas in the fluorescence image correspond to secretory granule areas. Pseudocolour‐coded BTC ratio images indicate that acetylcholine (ACh; 1 μM) induced an increase in [Ca2+]i to 10 μM in the trigger zone and 5.5 μM in the basal area (Figure 1B). The peak [Ca2+]i averaged over 14 cells was 9.89 ± 1.96 μM (mean ± SD) in the trigger zone and 5.4 ± 2.05 μM in the basal area. A large increase in [Ca2+]i occurred in acinar cells upon their stimulation with cholecystokinin (CCK, 10 nM, n = 5) or with caged inositol trisphosphate (IP3; 100 μM, n = 5, data not shown). A large increase in [Ca2+]i also occurred in the cells in which the effects of whole‐cell perfusion were minimized by means of a short ‘whole‐cell’ episode (5 s) (Zhou and Neher, 1993) with a patch pipette containing 3 mM BTC, upon their stimulation with ACh (50 nM–1 μM, n = 9).
We noted two differences between the Ca2+ images obtained using BTC and those obtained using fura‐2 (Figure 1C). First, the Ca2+ gradient, with the [Ca2+]i peaking in the trigger zone, lasted >10 s in the BTC images, whereas the gradient was obscure or not present after the Ca2+ wave spread to the basal area in the fura‐2 images (Figure 1C) (Kasai and Augustine, 1990; Toescu et al., 1992). Second, the [Ca2+]i started to dissipate at the trigger zone within a few seconds after its peak in the BTC images (Figure 1B), whereas it remained high in this zone in the fura‐2 images (Figure 1C). These differences can be explained by the fact that the Kd of fura‐2 for Ca2+ (0.2 μM) is 50 times smaller than that of BTC for Ca2+ (10 μM) and thus fura‐2 can be saturated with Ca2+ at a high [Ca2+]i.
Two‐indicator four‐wavelength Ca2+ measurements
The differences in the spatio‐temporal patterns of [Ca2+]i measured with BTC and fura‐2 might be due to cell to cell variation or to different effects of the two Ca2+ indicators on cytosolic buffering in acinar cells. To exclude these possibilities, the acinar cells were loaded simultaneously with fura‐2 and BTC, and two ratios were acquired simultaneously for the same cell (see Materials and methods), taking advantage of the fact that the excitation wavelengths of fura‐2 and BTC differ greatly (Grynkiewicz et al., 1985; Iatridou et al., 1994). Consistent with the results obtained using a single indicator, the standing luminal–basal Ca2+ gradient (Figure 2A, 0.68–3.08 s) and its quick dissipation (Figure 2A, 3.08–16.04 s; Figure 2B) were detected only in the BTC images, and not in the fura‐2 images (Figure 2A and B). This indicates that fura‐2 was saturated during the Ca2+ spikes and confirms the occurrence of micromolar increases in [Ca2+]i.
Interestingly, the magnitude of the initial [Ca2+]i increase at the trigger zone estimated using BTC (Figure 2A, 0.68 s) tended to be larger than that estimated using fura‐2 [Figure 2A, 0.48 and 0.96 s, the period represented by the yellow bar (a) in Figure 2B] in all nine cells examined. This difference can be explained by saturation of fura‐2 within small areas of high [Ca2+]i which cannot be resolved using the present imaging methods. The small areas might reflect hot spots in the trigger zone (Thorn et al., 1996) and/or Ca2+ domains of Ca2+ release channels (Rizzuto et al., 1993). In contrast, the [Ca2+]i in the basal area estimated using fura‐2 was similar to that estimated using BTC unless [Ca2+]i increased to levels >2 μM which resulted in saturation of fura‐2 (Figure 2A and B), indicating that the two indicators were calibrated correctly. Details of the initial difference at the trigger zone will be reported in a separate paper. An apparent slow increase in [Ca2+]i at the trigger zone as detected in fura‐2 images [the period represented by the blue bar (b) in Figure 2B] occurred when the [Ca2+]i estimated using BTC decreased. This is considered to have been due to cross‐talk between fura‐2 and BTC signals, as predicted, when [Ca2+]i was higher than 2 μM (Materials and methods; Figure 7). A similar phenomenon was observed in the basal area (Figure 2B). In contrast, fura‐2 images did not exhibit such a slow phase in one‐indicator measurements (Figure 1C and D), supporting that the phenomenon was due to cross‐talk.
Low concentrations of agonist could induce Ca2+ spikes whose amplitudes were <2 μM. In the example shown in Figure 2C, fura‐2 images revealed submicromolar Ca2+ waves that originated in the trigger zone and spread throughout the basal area (Figure 2C). In contrast, BTC images revealed smaller [Ca2+]i increases or even a slight reduction in [Ca2+]i (Figure 2C and D). The slight reduction in [Ca2+]i estimated using BTC must have been due to cross‐talk between fura‐2 and BTC signals, as was predicted (Materials and methods; Figure 7). This confirmed that increases in [Ca2+]i never exceeded 2 μM. These patterns of Ca2+ spikes were detected in 7 out of 13 cells to which 0.1 μM ACh was applied. In addition, small Ca2+ spikes localized in the trigger zone were detected in 4 out of 11 cells to which 50 nM ACh was applied (data not shown) (Kasai et al., 1993; Thorn et al., 1993). Thus, we identified a total of three types of Ca2+ spikes.
Ca2+ dependence of exocytosis
We next examined Ca2+ and time dependences of exocytosis in the acinar cells by capacitance measurement and use of a method involving photolysis of the caged Ca2+ compound nitrophenyl EGTA (NP‐EGTA). We blocked most of the Ca2+‐activated currents in these experiments to reduce the conductance changes (see Materials and methods). The acinar cells exhibited a slow increase in membrane capacitance upon sudden increases in [Ca2+]i (Ca2+ jumps) to between 5 and 8 μM (Figure 3A). There was an initial lag in the sigmoidally increasing capacitance (Figure 3A). The mean value of the delay was 2.3 ± 1.0 s (mean ± SD, n = 23). Larger Ca2+ jumps induced a faster capacitance increase in addition to the slow one (Figure 3B and C). The slow increase lasted as long as the [Ca2+]i remained high (Figure 3B), while the capacitance sometimes decreased if the [Ca2+]i decreased (Figure 3C). In such cases, we often detected stepwise decreases in capacitance (Figure 3C, inset), which were also detected in endocrine cells (Rosenboom and Lindau, 1994; Thomas et al., 1994), and probably represented the formation of vacuoles (Baker and Knight, 1981; Knoll et al., 1992; Back et al., 1993). These features of exocytosis and endocytosis were qualitatively very similar to those reported in the case of endocrine cells (Thomas et al., 1994; Kasai et al., 1996), supporting the view that the changes in capacitance reflect exocytosis and endocytosis of secretory vesicles in the acinar cells.
The Ca2+ dependence of fast and slow increases in capacitance was determined by applying Ca2+ jumps of various magnitudes in acinar cells (Figure 4A). The slow component of exocytosis was detected at [Ca2+]i as low as 5 μM, while the fast component of exocytosis appeared at [Ca2+]i of 8 μM and became larger at higher [Ca2+]i. Rates of capacitance increases were quantified and are plotted in Figure 4B and C. The Ca2+ dependence of the slow exocytosis was fitted with a Hill equation with a dissociation constant of 8 μM and a Hill coefficient of 3 (solid curve). On the other hand, the Ca2+ dissociation constant and Hill coefficient of the fast exocytosis were 15 μM and 4, respectively. The Ca2+ dependence of the slow exocytosis is similar to results obtained by measuring amylase secretion from permeabilized acinar cells (Muallem et al., 1995), suggesting that the slow exocytosis represents secretion of zymogen granules.
Micromolar Ca2+ spikes and exocytosis
We then studied agonist‐induced exocytosis using capacitance measurement and Ca2+ imaging. Ion channels were not blocked in these experiments (Figure 5) in order to allow examination of the activation of Ca2+‐dependent ion channels. A high concentration (≥1 μM) of ACh increased the [Ca2+]i at the trigger zone to >5 μM, and induced increases in capacitance (Figure 5A). This conclusion was based on the results of 11 experiments in which Ca2+‐dependent currents and access resistance were <300 pA and 4 MΩ, respectively. Therefore, the capacitance changes should not be much affected by the conductance changes (Materials and methods). In fact, little change in capacitance was detected in acinar cells perfused with Ca2+ buffers (2 mM BAPTA and 0.8 mM CaCl2) to reduce the peak [Ca2+]i to <5 μM (Figure 5B), even though Ca2+‐dependent currents with larger amplitudes were recorded (Figure 5B). On the other hand, when the [Ca2+]i did not reach 5 μM, little exocytosis was detected (Figure 5C) (n = 7 for 0.1 μM, n = 6 for 50 nM). This indicates that the agonist‐induced exocytosis depends on [Ca2+]i, and that an increase in [Ca2+]i to at least 5 μM is required, as was the case with the slow component of exocytosis induced by photolysis of NP‐EGTA.
In contrast to the case for exocytosis, Ca2+‐dependent ionic currents were induced at [Ca2+]i <0.2 μM. Figure 5C shows the data obtained during the imaging experiments shown in Figure 2C. Membrane potentials were clamped at the reversal potential of non‐selective cation currents to allow selective recording of chloride currents (Kasai and Augustine, 1990). Chloride currents had two transient components as described previously (Kasai and Augustine, 1990). The first transient current appeared as soon as the initial increase in [Ca2+]i in the trigger zone was detected (Kasai and Augustine, 1990; Thorn et al., 1993), indicating that the chloride current is induced at the area of the plasma membrane adjacent to the trigger zone. The current decayed even though the [Ca2+]i at the trigger zone increased, suggesting that the decay was due to inactivation of chloride channels. When the Ca2+ waves reached the basal area, the second chloride current appeared, suggesting that the channels were distributed in the basolateral area (Kasai and Augustine, 1990). These results indicate that the ion channels and exocytosis were regulated differentially by Ca2+ spikes in a [Ca2+]i‐dependent manner (Figure 6).
Here, we have described the first investigation of Ca2+ signalling and exocytosis using a low affinity Ca2+ indicator, BTC, in pancreatic acinar cells. Our data demonstrate that agonists can evoke distinct Ca2+ spikes with widely differing amplitudes, i.e. micromolar and submicromolar ones. We will discuss below the characteristics and functional consequences of the multiple types of Ca2+ spikes in acinar cells.
Three types of Ca2+ spikes in acinar cells
Agonist‐induced increases in [Ca2+]i to levels >10 μM were detected using a low affinity Ca2+‐selective indicator, BTC, whose Kd for Ca2+ was estimated to be ∼10 μM in both in vitro (Iatridou et al., 1994) and in vivo calibration experiments (Materials and methods). Moreover, the two‐indicator four‐wavelength Ca2+ imaging results indicated that a high affinity Ca2+ indicator, fura‐2, was saturated during the spikes (Figure 2A). Thus, in our imaging study, we consistently demonstrated that IP3‐dependent mechanisms are involved in the increase of [Ca2+]i to levels as high as 10 μM. Results of an earlier study suggested the occurrence of micromolar [Ca2+]i increases in the Ca2+ domains of Ca2+ release channels (Rizzuto et al., 1993). Our findings further indicate that micromolar Ca2+ spikes that can be imaged directly using Ca2+ indicators occur in bulk cytosol.
The concentrations of Ca2+ measured using BTC routinely peaked at the trigger zone during the micromolar Ca2+ spikes (Figures 1 and 2). When we used a high affinity indicator, however, the [Ca2+]i peaked at the trigger zone only at the onset of the [Ca2+]i increases (Kasai et al., 1993) (Figures 1C and 2A). This difference is explained by the saturation of the indicator at high [Ca2+]i. Also, the highest estimates of [Ca2+]i in the basal area after the spread of Ca2+ waves to the basal area were obtained from fura‐2 images in some studies (Kasai and Augustine, 1990) but not in others (Kasai et al., 1993; Thorn et al., 1993). The results of imaging studies using fura‐2 should be interpreted with caution, because fluorescence ratios of fura‐2 are considerably affected by slight spatial heterogeneity in Rmax of fura‐2, because the estimated [Ca2+]i, KCaβΔR/(Rmax−R), critically depends on Rmax (Grynkiewicz et al., 1985) at [Ca2+]i higher than 2 μM, where R becomes close to Rmax. In contrast, the [Ca2+]i estimated using BTC is more reliable at high [Ca2+]i, because R is not close to the Rmax of the indicator if [Ca2+]i is lower than 20 μM.
We have thus demonstrated that the micromolar Ca2+ spikes are induced with physiological concentrations of ACh or CCK, and are always spatio‐temporally organized in such a manner that the Ca2+ spikes which initiate and peak at the trigger zone are transient and never exceed 15 μM (Figure 1A). Since IP3 receptors are ubiquitous in animal cells, similar micromolar Ca2+ gradients may occur in other cells. We therefore expect our findings to prompt investigation of Ca2+ homeostasis using low affinity Ca2+ indicators in efforts to understand the physiology and pathology of cells.
Importantly, low concentrations of agonists could trigger smaller Ca2+ spikes whose amplitudes do not greatly exceed 2 μM. This was confirmed by the results of the two‐indicator four‐wavelength imaging (Figure 2C). The submicromolar Ca2+ spikes originated at the trigger zone, and spread throughout the cell, as in the case of the micromolar Ca2+ spikes. Therefore, global Ca2+ spikes (Thorn et al., 1993) can be either submicromolar or micromolar, depending on the agonist concentration. This is, to our knowledge, the first report of two types of Ca2+ waves having widely different amplitudes in the same cell. Since we also detected local Ca2+ spikes, restricted to the trigger zone, at low agonist concentrations (Kasai et al., 1993; Thorn et al., 1993), altogether three types of Ca2+ spikes have been identified by Ca2+ imaging in the acinar cells. The existence of multiple types of Ca2+ spikes is consistent with the incremental detection mechanisms of IP3‐induced Ca2+ release reported for acinar cells (Muallem et al., 1989) and other preparations (Meyer and Stryer, 1990). The extent of spreading of Ca2+ waves and their amplitudes may be determined by a quantitative balance between the IP3‐dependent Ca2+ release and Ca2+ pumping (van de Put et al., 1994). A precise molecular and theoretical understanding of the diversity in Ca2+ spikes and waves remains to be gained.
Ca2+ dependence of exocytosis and electrolyte secretion
Experiments involving both capacitance measurements and photolysis of a caged‐Ca2+ compound revealed the occurrence of two components of exocytosis in exocrine acinar cells (Figures 3 and 4). We believe that slow exocytosis represents exocytosis of the zymogen granules, because its Ca2+ dependence correlates well with existing data on the secretion of amylase from permeabilized pancreatic acinar cells (Knight and Koh, 1984; Cher et al., 1992; Stecher et al., 1992; Muallem et al., 1995; Padfield and Panesar, 1995). This is also consistent with results of a study on PC12 cells, in which the exocytosis of large dense core vesicles was slower and showed a higher affinity for Ca2+ than that of small clear vesicles (Kasai et al., 1996; Ninomiya et al., 1997). The fast exocytosis might represent exocytosis of other types of secretory vesicles, for example those carrying newly synthesized proteins (Arvan and Castle, 1987) or small clear vesicles which have not been identified in acinar cells.
In order to compare the kinetic features of the Ca2+‐dependent exocytosis of the zymogen granules in acinar cells and the synaptic vesicles in neurones, the release rates and delays were fitted with a model used to explain exocytosis of the synaptic vesicles (Heidelberger et al., 1994):
where V0–V4 represent vesicles whose Ca2+ binding sites are occupied by 0–4 Ca ions, F is a fused vesicle, k is the rate constant of Ca2+ binding, C the concentration of Ca2+, l the rate constant of Ca2+ dissociation and a the rate constant of fusion. The parameter c accounts for the cooperativity of the Ca2+ binding (Heidelberger et al., 1994), and is given the value 0.2 to account for the large Hill coefficient of the Ca2+ dependence (Figure 4B). a was <0.1/s, because the slow exocytosis lasted >10 s (Figure 3B), if the [Ca2+]i increase was maintained. Therefore, the delay in exocytosis (2.3 s at [Ca2+]i <10 μM) is attributed chiefly to a small k, and the Ca2+ dependence of the release rates (Figure 4) depends both on k and l. A satisfactory fit of these data can be obtained by setting k equal to 1×105/s/M and l equal to 6/s (dashed curve in Figure 4B).
These results indicate that the slow exocytosis in the acinar cells can be accounted for by a and k being >100 times smaller than those for the nerve terminal (Heidelberger et al., 1994). The smaller a means that the [Ca2+]i increases must persist for longer than a few seconds in order to induce secretion. The smaller k and a suggest that the exocytosis of zymogen granules is little affected by Ca2+ in the Ca2+ domain, but is regulated by Ca2+ in bulk cytosol. Consistent with this hypothesis, ACh triggered exocytosis only when a Ca2+ indicator detected [Ca2+]i increases to levels >5 μM (Figure 5), and the agonist‐induced micromolar Ca2+ spikes lasted more than a few seconds. It is possible that exocytosis takes place mostly at the area of the plasma membrane adjacent to the trigger zone, because the largest increases in [Ca2+]i were detected there. It can then be speculated that disruption of the normal micromolar Ca2+ gradients triggers aberrant exocytosis, and results in acute pancreatitis (Ward et al., 1995).
In contrast, during the submicromolar Ca2+ spikes, Ca2+‐dependent ion channels were activated without exocytosis (Figure 5C). In such a case, sequential activation of the Cl− channels was detected, as reported previously, suggesting the existence of a push–pull mechanism of fluid secretion (Kasai and Augustine, 1990). Thus, the Ca2+ spikes can induce exocytosis and electrolyte secretion separately in a [Ca2+]i‐dependent manner, as shown in Figure 6. According to this model, the three types of Ca2+ spikes trigger distinct cellular activities: the local Ca2+ spikes at the trigger zone cause secretion of Cl− into the lumen (push phase), the global submicromolar Ca2+ spikes induce uptake of Cl− into the acinar cell (pull phase) and the micromolar Ca2+ spikes trigger exocytosis (exocytosis phase). During the push phase, Cl− currents decayed, although the [Ca2+]i was increasing at any part of the cell (Figure 2C). A similar decay in Ca2+‐dependent Cl− currents was reported to occur in Xenopus oocytes (Parker and Yao, 1994). On the other hand, the Cl− currents during the pull phase were sustained so long as the [Ca2+]i remained high (Figure 5A–C). Our data suggest, therefore, that two types of Cl− channels with distinctive distributions and regulation mechanisms are involved in electrolyte secretion from exocrine acinar cells.
In summary, our findings provide a new concrete example of Ca2+ release from intracellular stores regulating distinct cellular functions in an agonist concentration‐dependent manner. In particular, we have found that exocytosis and electrolyte secretion are regulated by Ca2+ spikes with a wider dynamic range in their amplitudes than hitherto considered, i.e. micromolar and submicromolar Ca2+ spikes, respectively. These findings emphasize the value of using low affinity Ca2+ indicators in elucidating the functional roles of Ca2+ signalling in living cells.
Materials and methods
Pancreatic acinar cells from 5‐ to 7‐week‐old mice were dissociated from pancreas by an enzymatic treatment as described (Maruyama, 1988). The dissociated cells were dispersed in a small chamber for electrophysiological recording in a recording solution (Sol A) containing 140 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM Na–HEPES (Dojin, Kumamoto) and 10 mM glucose at pH 7.4 and 310 mOsm. ACh (Wako, Osaka) and CCK (Peptide Institute, Osaka) were dissolved in Sol A and applied to acinar cells via a glass pipette. Ca2+ indicators were dissolved in a solution (basic internal solution) containing 120 mM Cs glutamate, 5 mM CsCl, 50 mM Cs–HEPES, 1 mM ATP, 0.2 mM GTP and 2 mM MgCl2 at pH 7.2, and indicator‐containing solution was loaded into cells using the patch–clamp method. We used BTC and/or fura‐2 (Molecular Probes, Eugene) as Ca2+ indicators. The concentrations of the indicators were 200 μM in single‐indicator measurements, and 200 and 350 μM for fura‐2 and BTC, respectively, in two‐indicator measurements. Caged‐IP3 [d‐myo‐inositol 1,4,5‐trisphosphate, P4(5)‐1‐(2‐nitrophenyl)‐ethyl ester, Calbiochem, San Diego] was added to the basic internal solution. As a caged‐Ca2+ compound, we used 10 mM NP‐EGTA (Molecular Probes) added with 5 mM CaCl2. In order to reduce Ca2+‐induced ion currents during photolysis of NP‐EGTA, we used an internal solution containing 105 mM N‐methyl‐d‐glucamine (NMDG, Wako) glutamate, 50 mM NMDG–HEPES, 1 mM ATP and 0.2 mM GTP at pH 7.2. In addition, the solution surrounding a cell was perfused locally using a glass pipette (φ = 10 μm) with an external solution containing 140 mM NMDG‐glutamate, 10 mM NMDG–HEPES and 10 mM glucose at pH 7.4. Photolysis routinely induced increases in [Ca2+]i to >10 μM. Smaller [Ca2+]i increases were produced by loading NP‐EGTA with lower concentrations of CaCl2, i.e. 2.5, 2, 1 or 0.5 mM. If necessary, the pH of the solutions was readjusted using HCl. Osmolarities of the external and internal solutions were estimated to be ∼310 mM after addition of all chemicals (Semi‐Micro Osmometer, Knauer, Berlin, Germany). We carried out all methods under yellow light illumination (FL40S‐Y‐F, National, Tokyo, Japan) at room temperature, 22–25°C.
Ca2+ imaging setup
A recording chamber was placed on an inverted microscope (IMT‐2, Olympus, Tokyo) and observed through an objective lens (DApo 40× UV/340 oil, Olympus). Measurement of [Ca2+]i was performed using the Ca2+ indicators BTC and/or fura‐2. Monochromatic beams of light with a wavelength of 340, 380, 430 or 480 nm were isolated from light emitted by a xenon lamp using a monochromator (T.I.L.L. Photonics, Munich), and fed into one port of a light guide (IMT‐2‐RFA caged, Olympus). The light was reflected by a dichroic mirror, DM500, placed beneath the objective lens, and fluorescent light emitted from the cells was captured using a cooled‐CCD camera system (T.I.L.L. Photonics) (Messler et al., 1996) fixed at the side port of the IMT‐2. During two‐wavelength imaging, two images were captured successively at 430 and 480 nm for BTC, and at 340 and 380 nm for fura‐2. The duration of image acquisition was 0.1 s, and the pairs of images were acquired every 0.24 s. During four‐wavelength imaging, images were successively acquired for 0.1 s at 340, 380, 430 and 480 nm, and the same acquisition sequence was repeated every 0.48 or 2 s. Ratios of fluorescence intensities in two images were calculated after subtracting the background fluorescence. [Ca2+]i was estimated from these ratios using the method described below, and represented by pseudocolour coding where 0.1, 0.3, 1, 3 and 10 μM were expressed as blue, sky blue, green, yellow and red, respectively (Figures 1 and 2).
One‐indicator two‐wavelength ratiometric Ca2+ imaging
Ratiometric estimation of [Ca2+]i using either fura‐2 or BTC was performed using the method developed by Grynkiewicz et al. (1985). Using their notation, fluorescence of an indicator (D), FD, at four excitation wavelengths, λj (340, 380, 430 and 480 nm for j 0–3, respectively), is given by
where SfDj and SbDj represent the fluorescence of free and Ca2+‐bound indicators, respectively, and CfD and CbD represent the concentrations of free and Ca2+‐bound indicators, respectively. The fluorescence ratio RDj = FD(j−2)/FD(i−1) for j = 2 or 4 is expressed as a monotonic function of Ca2+ concentration, [Ca], as,
where RminDj = SfD(j−2)/SfD(j−1), RmaxDj = SbD(j−2)/SbD(j−1), βDj = SfD(j−1)/SbD(j−1) and KD represents the dissociation constant of the indicator for Ca2+. Calibration experiments for fura‐2 and BTC were carried out in acinar cells in vivo (Almers and Neher, 1985), and six constants were obtained: RminA2 = 0.17, RmaxA2 = 2.5 and KAβA2 = 0.87 for fura‐2, and RminB4 = 0.514, RmaxB4 = 2.0 and KBβB4 = 112 for BTC (n = 5–12). Four [Ca] calibration solutions were prepared from the basic internal solution by adjusting its [Ca] to either 0 μM, 0.15 μM, 10 μM or 10 mM using Ca2+ buffers and CaCl2. For the 0 μM and 10 mM [Ca] calibration solutions, we simply added 10 mM EGTA and 10 mM CaCl2, respectively, to the internal solution. To prepare the 0.15 and 10 μM [Ca] calibration solutions, we used 1,2‐bis(2‐aminophenoxy)ethane‐N,N,N′,N′‐tetracetic acid (BAPTA, Molecular Probes) and 2‐aminophenol‐N,N,O‐triacetic acid (APTRA, Molecular Probes), respectively, as Ca2+ buffers. We first added 20 mM BAPTA or 20 mM APTRA plus 8 mM CaCl2 to the internal solution, and titrated [Ca] of the solution to 0.15 and 10 μM, respectively, with CaCl2 using a mini Ca2+ electrode, which was made of polyethylene tubing (φ = 1.5 mm), and a Ca2+ ionophore, ETH 129 (Fluka, Switzerland) (Baudet et al., 1994). The [Ca] calibration solutions were whole‐cell dialysed into the cells in the presence of a Ca2+ indicator, and fluorescence ratios were obtained after equilibration to give RminDj, RmaxDj and KDβj. Our estimations were based on measurements in >5 cells.
Using Ca2+ images obtained using fura‐2, distributions of RA2 were first calculated, and [Ca] was estimated as KAβA2(RA2−RminA2)/(RmaxA2−RA2) (Grynkiewicz et al., 1985). Using Ca2+ images obtained using BTC, the distribution of RminB4 in individual cells was first estimated by averaging several frames of resting distribution of RB4. This procedure was used to compensate for small heterogeneity in RminB4 (SD <0.01) within a cell, and to reduce noise levels particularly at [Ca2+]i lower than 1 μM. Distributions of ΔRB4 were then calculated by subtracting the distribution of RminB4 from that of RB4. Using ΔRB4, [Ca] could be obtained as KB4βB4ΔRB4/(RmaxB4−m[RminB4]−ΔRB4), where m[RminB4] represents the spatial average of RminB4.
Two‐indicator four‐wavelength ratiometric Ca2+ imaging
assuming a linear summation of the fluorescence of the two indicators. The ratio Rj = Fj−2/Fj−1 for j = 2 and 4 is expressed as
where γj = SfB(j−1)/SfA(j−1), and [A] and [B] are concentrations of fura–2 and BTC, respectively. KA and KB were estimated as 0.2 and 10 μM, respectively, which are close to the values reported elsewhere (Grynkiewicz et al., 1985; Iatridou et al., 1994). Using whole‐cell dialysis of [Ca] calibration solutions into the acinar cells and by calculating the ratios, we estimated RminA4 = 15.3, RmaxA4 = 2.7, KAβA4 = 0.16, RminB2 = 0.18, RmaxB2 = 0.124 and KBβB2 = 3 (n = 5–9). The in vivo values of γj were estimated as 0.263 and 131.6 for γ2 and γ4, respectively, in the presence of 200 μM fura‐2, 350 μM BTC and 10 mM EGTA, according to the relationship γj = [A](Rj−RminAj)/[B](RminBj−Rj), which is derived from Equation 5. The small γ2 and large γ4 indicate that the amount of cross‐talk between the two indicators is relatively small. On the basis of these values and Equation 5, the [Ca2+]i dependences of R2 and R4 were predicted as smooth curves, shown in Figure 7. The predicted curves were confirmed by direct measurement of R2 and R4 in the presence of two indicators loaded into acinar cells at five different [Ca2+]i, 0, 0.15, 3, 10, 185 μM and 10 mM, as plotted in Figure 7 (circles). The 3 μM Ca2+‐buffered solution was prepared using 20 mM BAPTA, and its Ca2+ concentration was verified using the mini Ca2+ electrode.
The calibration curves can be approximated by the equation, Rj = (Rminj+Rmaxj[Ca]/Kj)/(1+[Ca]/Kj) for j = 2 and 4, where Rmin2 = 0.185, Rmax2 = 0.85, K2 = 0.3, Rmin4 = 0.545, Rmax4 = 2.35 and K4 = 150 μM (dashed curves in Figure 7). However, R2 and R4 are not monotonically increasing functions of [Ca] between 2 μM and 10 mM for R2 and between 0 and 2 μM for R4 where a decrease in R2 and R4 is detected (Figure 7, solid curves). The decreases are due to the small amount of cross‐talk between the two indicators, and were also seen in actual experiments. R4 decreased even when [Ca2+]i increased, as shown in Figure 2D, and R2 increased when [Ca2+]i decreased (Figure 2A). Therefore, [Ca] could not be determined uniquely from R2 and R4 in the respective ranges of [Ca]. Outside the ranges mentioned above, however, R2 and R4 predicted [Ca] with <20% error (Figure 7, dashed and solid curves). Therefore, Ca2+ images were obtained from distributions of R2 and ΔR4 as was the case in the two‐wavelength Ca2+ measurement, and [Ca] was estimated from K2(R2−Rmin2)/(Rmax2−R2) and K4ΔR4/(Rmax4− m[Rmin4]−ΔR4), respectively, in these ranges. In order to facilitate comparison between R2 and ΔR4 images (Figure 2A), Rmax2 was adjusted slightly in each cell when necessary so that [Ca] during the recovery phase of agonist‐induced Ca2+ transients was estimated similarly using both R2 and ΔR4. This correction only slightly affected (<10%) the estimate of [Ca] lower than 1 μM.
Photolysis of a caged‐Ca2+ compound
Photolysis of a caged compound, NP‐EGTA, was achieved as described (Kasai et al., 1996; Ninomiya et al., 1996). In brief, we used a mercury lamp (IMT‐2‐RFC, Olympus) as an actinic light source. The light from the mercury lamp was filtered through a 360 nm band‐pass filter, and fed into one port of the light guide (IMT‐2‐RFA caged, Olympus). Incorporating a dichroic mirror, DM400, the light guide can accommodate two light sources, one for the actinic light and the other for excitation of a Ca2+ indicator. Irradiation with the actinic light was gated through an electric shutter (Copal, Tokyo). The duration of the opening of the shutter was set at 33 ms, which was sufficient for full activation of the caged compounds within the cells. The speed of the photolysis was not expected to affect the time course of exocytosis, which was slow in the case of the acinar cells.
In most studies in which a caged‐Ca2+ compound was used, [Ca2+]i was estimated by microfluorimetric measurement using a photomultiplier (NT5783, Hamamatsu Photonics, Hamamatsu, Japan) attached to the side port of an IMT‐2 instead of the cooled‐CCD camera. One ratiometric Ca2+ measurement and capacitance measurement were carried out simultaneously at 44 Hz. In vivo calibration experiments of BTC were repeated in the presence of NP‐EGTA (Zucker, 1992). We prepared a total of six calibration solutions that contained the basic internal solution plus 10 mM of either the caged or photolysed form of NP‐EGTA. [Ca] of the calibration solutions was adjusted to either 0 μM, 20 μM or 10 mM with Ca2+ buffers and CaCl2, and confirmed using the mini Ca2+ electrode as described above. RmaxB4 and RminB4 were 2.45 and 0.55, respectively for both caged and uncaged NP‐EGTA. Values of KBβB4 were 100 and 120 μM for caged and uncaged NP‐EGTA, respectively. The values for photolysed caged‐Ca2+ compounds were used to estimate the peak values of [Ca2+]i, since most of the caged‐Ca2+ compounds were photolysed in our experiments (see below). Thus, [Ca2+]i was estimated as KBβB4ΔRB4/(RmaxB4−RB4). The error in the estimates of [Ca2+]i during re‐equilibration of caged Ca2+ compounds after photolysis was considered to be <16%.
Capacitance measurements were carried out using a patch–clamp amplifier, AxoPatch 1D (Axon Instrument, Foster City) (Kasai et al., 1996). When Ca2+ images were acquired, the time of image acquisition was monitored by recording the readout signal of the CCD camera during the capacitance measurement. The membrane potential of the cells was maintained at either 0 or −27 mV (these values were corrected for liquid junction potential), at which current–voltage relationships were almost linear and to which 1 kHz (ω = 6283 radian/s) sine waves with a peak–peak amplitude of 50 mV were applied. Our experiments were performed using a total of 84 cells, where the initial membrane capacitance ranged between 8 and 14 pF (mean = 11.5, n = 84). Changes in capacitance were calculated from 10 cycles of sine waves and stored at 44 Hz. The phase‐offset was corrected several seconds before application of Ca2+ jumps or agonists using the phase‐tracking method (Fidler and Fernandez, 1989). For those experiments where ion currents were not blocked, ΔG of 10 nS (200 pA current at −20 mV) resulted in error of <23 fF [C(ΔG/Ga)2] (Maruyama et al., 1993) when C was 10 pF, access resistance (Ga) was <5 MΩ and the phase offset was assumed to be corrected accurately. The inaccuracy in the correction of phase offset (θ) was estimated to be <0.05 radian using a capacitance calibration command generated by DC‐1 option (Axon Instrument) [θ ≑ EGerror(ΔC)/ ωECerror(ΔC)] (Maruyama et al., 1993). The inaccuracy in the correction of phase offset resulted in <83 fF error (θΔG/ω) at ΔG of 10 nS (Maruyama et al., 1993). Thus, overall error in the estimate of ΔC due to ΔG is <106 fF, which cannot account for the larger changes in ΔC actually detected (>500 fF, Figure 5A).
This work was supported by Grants‐in‐Aid for Scientific Research from the Japanese Ministry of Education, Science and Culture, CREST (Core Research for Evolutional Science and Technology) of the Japan Science and Technology Corporation (JST), research grants from the Human Frontier Science Program and the Takeda Foundation. K.I. is a research fellow of the Japan Society for the Promotion of Science.
- Copyright © 1997 European Molecular Biology Organization