Ca²⁺ dependency of ‘Ca²⁺-independent’ exocytosis in SPOC1 airway goblet cells

Abstract

SPOC1 airway goblet cells secrete mucin in response to P2Y₂ receptor agonists and to secretagogues, phorbol 12-myristate 13-acetate (PMA) and ionomycin, which mobilize elements of the phospholipase C pathway, PKC and Ca²⁺, respectively. Previous studies demonstrated that mucin secretion from SLO-permeabilized, EGTA-buffered SPOC1 cells was stimulated by PMA at low Ca²⁺ levels (< 0.1 μm), consistent with the notion that regulated exocytosis may occur by Ca²⁺-independent pathways. We tested the alternative hypothesis that PMA-induced mucin secretion is, in fact, a Ca²⁺-dependent process under the conditions of low bulk Ca²⁺, one that is permitted in the typical SLO-permeabilized cell model by the slow binding kinetics of EGTA. Both IP₃ and elevated bulk Ca²⁺ activated mucin secretion in SPOC1 cells buffered by EGTA, suggesting that IP₃ generates a local Ca²⁺ gradient in the vicinity of the secretory granules to the degree necessary to trigger exocytosis. BAPTA, which binds Ca²⁺ approximately 100-fold faster than EGTA, diminished IP₃-induced mucin release over a range of concentrations by ≥ 69%, yet maintained an essentially normal mucin secretory response to elevated bulk Ca²⁺ in permeabilized SPOC1 cells. BAPTA also diminished the mucin secretory response of permeabilized cells to PMA, relative to the EGTA-buffered control: at PMA below 30 nm, BAPTA abolished the secretory response, and at higher concentrations it was reduced significantly relative to the EGTA-buffered controls. PMA-induced secretion in EGTA was insensitive to heparin. These results suggest that Ca²⁺ is released locally during PMA-induced exocytosis, by an IP₃-independent mechanism.

Airway goblet cells and submucosal glands secrete the high molecular weight glycoconjugate mucin, the viscoelastic component of mucus. In healthy lungs, mucus plays a principal role in mucociliary clearance (Knowles & Boucher, 2002). However, in the airway obstructive diseases asthma, chronic bronchitis, bronchiectasis, and cystic fibrosis, mucus/mucin hypersecretion is a hallmark characteristic that arises from inflammation-induced goblet cell and submucosal gland hyper- and metaplasia (reviewed by Rogers, 2003). Despite the central role of mucin secretion in health and its importance in airway disease, the regulation of mucin granule exocytosis is poorly understood beyond the level of receptor activation and cellular messenger generation.

SPOC1 cells, derived from rat tracheal epithelium (Randell et al. 1996), secrete mucin in response to purinergic stimulation by ATP or UTP activation of P2Y₂ receptors (Abdullah et al. 1996), in agreement with studies on goblet cells in primary cultures from hamster trachea (Kim & Lee, 1991), canine tracheal epithelial explants (Davis et al. 1992), and human bronchial epithelial cells (Chen et al. 2001; Conway et al. 2003). P2Y₂ receptors typically couple to PLC (Harden et al. 1995), and mobilization of intracellular Ca²⁺ by ionomycin and activation of PKC by the DAG mimic, PMA, stimulate mucin secretion in SPOC1 cells (Abdullah et al. 1997; Abdullah et al. 2003). Much of the SPOC1 cell data on the regulation of mucin secretion by Ca²⁺ and PKC are consistent with the notion that these effectors function independently. Chief among these are the findings that ionomycin and PMA effects are fully additive at maximal concentrations (Abdullah et al. 1997), that the agonist-responsive isoform of PKC, nPKCδ, is a member of the novel, Ca²⁺-independent PKC subfamily (Abdullah et al. 2003), and that in permeabilized, EGTA-buffered cells PMA stimulates mucin secretion at 10 nm free Ca²⁺, an order of magnitude below generally accepted basal levels of intracellular Ca²⁺ (Scott et al. 1998). Ca²⁺-independent, regulated exocytosis has also been proposed for other secretory cells: (i) non-hydrolysable analogues of GTP stimulate degranulation of permeabilized, EGTA-buffered mast cells (Gomperts et al. 1986), and in mast cells dialysed with EGTA buffer by a whole-cell patch pipette, exocytotic events sensed by changes in membrane capacitance failed to correlate with Ca²⁺ transients observed by fura-2 fluorescence (Neher & Almers, 1986); (ii) in gonadotropes of the anterior pituitary, PMA stimulates luteinizing hormone (LH) secretion in the absence of observable changes in Ca²⁺ (Betz et al. 1998); and (iii) in pancreatic duct epithelial cells, exocytosis is stimulated by agents which elevate cAMP, without observable changes in Ca²⁺ (Koh et al. 2000). Hille et al. (1999) have postulated that such data indicate a Ca²⁺-independent regulation of exocytotic secretion by protein kinases such as PKC and PKA.

Contrary to the notion of a mechanism for regulated exocytosis that is independent of Ca²⁺, there is a strong argument that the process is strictly Ca²⁺ dependent. Ca²⁺ plays important roles in the regulation of exocytosis, beginning with cortical actin filament disassembly (Trifaro et al. 2000), granule docking to the plasma membrane (Martin, 2002), and fusion of the granule membrane to the plasma membrane during pore formation (Gerber & Sudhof, 2002). In fact, the postulated trigger for regulated exocytosis is synaptotagmin, an obligate accessory protein to the exocytotic SNARE complex. Though the number of studies is at present small, only the Ca²⁺-dependent synaptotagmin isoforms have been associated with regulated exocytosis (Chapman, 2002; Sudhof, 2002).

Interestingly, experiments yielding the most direct evidence favouring Ca²⁺-independent exocytosis are typically buffered by EGTA. Because this Ca²⁺ buffer has relatively slow binding kinetics (Tsien, 1980), one explanation of apparent Ca²⁺-independent exocytosis is that EGTA permits the generation of local Ca²⁺ gradients. Hence, we used IP₃ and other manoeuvres to test for the existence of Ca²⁺ gradients in permeabilized, EGTA-buffered cells, and using BAPTA as a probe we tested the hypothesis that regulated SPOC1 cell mucin secretion is, in fact, Ca²⁺ dependent.

Methods

Materials

Dulbecco's modified Eagle's medium–Ham's nutrient mixture F12 (DMEM–F12) was obtained from Gibco BRL (Gaithersburg, MD, USA) and the supplements from Collaborative Research (Bedford, MD, USA). Bisindolylmaleimide II (BIMII), calphostin C, d-myo-inositol 1,4,5-trisphosphate (IP₃; trilithium salt), heparin (sodium salt; low molecular weight), and phorbol-12-myristate-13-acetate (PMA) were purchased from Calbiochem (La Jolla, CA, USA). Ethylene gylcol bis(β-aminoethyl ether)-N,N,N′, N′-tetraacetic acid (EGTA) was purchased from Sigma (St Louis, MO, USA); 1,2-bis(o-aminophenoxy)ethane-N,N,N′ N′-tetraacetic acid (BAPTA; tetrapotassium salt) and TO-PRO were purchased from Molecular Probes (Eugene, OR, USA). Streptolysin-O (SLO) was purchased from Corgenix (Peterborough, UK).

SPOC1 cell culture

SPOC1 cells were seeded in 48-well cluster plates (Costar, Cambridge, MA, USA) at 25 000 cells per well (Scott et al. 1998) and maintained in rat tracheal epithelial cell medium (Randell et al. 1996), as previously described. The cells, passage 7–15, were used for experiments 18–22 days post-confluence.

SLO permeabilization and mucin release

SPOC1 cells were equilibrated before each experiment, as follows, using a Finnpipette multistepper, multichannel pipetter (Needham Heights, MA) and a custom fabricated, multichannel aspirator to quickly remove and replace all solutions from 48-well culture plates. Cells were placed in a 37°C water bath, washed twice with DMEM–F12 (200 μl well⁻¹), and incubated for 30 min; this procedure was repeated three times. Cells were washed twice with phosphate buffered saline (PBS) and then with intracellular buffer (Buf_i, composition given below). Immediately following the washes, SPOC1 cells were permeabilized with the bacterial toxin streptolysin-O (SLO) at a final concentration of 1 U ml⁻¹ in Buf_i for 30 s (150 μl well⁻¹), as previously described (Scott et al. 1998). Following the permeabilization, cells were washed once in Buf_i and incubated with the appropriate solution (200 μl well⁻¹; see Results) for 15 min, at which time the samples were collected for the determination of mucin content. TO-PRO, a membrane-impermeant, DNA-binding fluorescent dye, was used to confirm cell permeability, as previously described (Scott et al. 1998). Notably, TO-PRO and BAPTA are similar in size (relative molecular mass 645 versus 628.8) and EGTA is approximately 50% smaller (relative molecular mass 380.4), making the dye a relevant marker of small molecule permeation. In all of the studies reported below, TO-PRO fluorescence was observed to be uniformly distributed throughout SPOC1 cell cytoplasm indicating that the cells were well permeabilized.

Mucin enzyme-linked lectin assay

Samples collected from each well of the 48-well culture plate were assessed for mucin content using an enzyme-linked lectin assay (ELLA), previously described (Abdullah et al. 1996). Briefly, samples were diluted (100 μl final volume) and bound to 96-well high-binding microtitre plates (Costar), incubated overnight at 4°C or 37°C for 2 h, washed with PBST (0.05% Tween 20 in PBS), and incubated with 2.5 μg ml⁻¹ peroxidase-labelled soybean lectin for 1 h at 37°C. The plates were then washed and developed with O-phenylenediamine in a citrate phosphate buffer (0.0175 m, 0.01% hydrogen peroxide, pH 5.0). H₂SO₄ (4 m) was used to stop the reaction, and the plates were analysed using optical density at 490 nm (Dynatech microtitre plate reader, model MR5000; Chantilly, VA, USA). Known amounts of purified SPOC1 mucin were used to generate standard curves, on each plate, allowing the results to be expressed as nanograms mucin released per culture.

Intracellular Ca²⁺ buffer

EGTA stock solutions (50 mm) were prepared as previously described (Gomperts & Tatham, 1992; Scott et al. 1998) after first titrating the EGTA with CaCl₂.2H₂O to determine true buffer concentrations (Miller & Smith, 1984). BAPTA stock solutions (25 mm) were calculated directly, with allowance made for impurities as determined by HPLC (Molecular Probes, OR, USA). These stocks were stored at −20°C. Free Ca²⁺ activities were computed with the aid of the computer program Chelator (Schoenmakers et al. 1992) and CaEGTA or CaBAPTA were used to adjust the final free Ca²⁺ levels. Except where stated, both Ca²⁺ buffers were used at a final concentration of 3 mm. Intracellular buffer (Buf_i) contained 130 mm potassium glutamate, 20 mm Pipes, 1 mm MgATP and MgCl₂, and 3 mm EGTA or BAPTA (pH 6.8). Free Ca²⁺ in Buf_i was set to 0.1 μm, or, where stated in Results, was varied from 0.01 to 10 μm.

Ca²⁺ buffer simulations

Simulations were generated to compare the effectiveness of BAPTA and EGTA to buffer Ca²⁺ over time and space. For the temporal simulation (Fig. 2A), the free Ca²⁺ concentration was calculated from a consideration of the rate of change of bound Ca²⁺ following a step addition of EGTA or BAPTA to a solution containing 10 μm free Ca²⁺. The rate of change of bound Ca²⁺, D, is:

Figure 2. Binding kinetics of EGTA and BAPTA and their relative effectiveness in Ca²⁺-activated mucin secretion.

A, temporal simulation using eqn (1), illustrating the Ca²⁺ binding kinetics of EGTA and BAPTA and their effects on free Ca²⁺ over 500 μs following the step addition of 3 mm buffer to a solution of 10 μm free-Ca²⁺. B, spatial simulation of steady state free Ca²⁺ concentration plotted as a function of the distance along a radial extending from a channel opening in a planar membrane. C, comparison of mucin secretion activated in SPOC1 cells by bulk Ca²⁺ (0.01–10 μm) when buffered by EGTA or BAPTA (n = 6 SPOC1 passages).

with association rate constant k₊, dissociation rate constant k_–, buffer concentration B, Ca²⁺ concentration C, complex concentration BC, and time t. This equation can be put into the form:

which has a solution,

where, at D(0) = 0,

The spatial simulation (Fig. 2B) compares the steady state free Ca²⁺ concentration profiles in the presence of BAPTA, EGTA, or no buffer on one side of a planar boundary with a point Ca²⁺ source. The profile can be modelled using an approximation to the reaction–diffusion problem first described by Neher (1986) and now known as the excess buffer approximation (for review, see Stern, 1992; Bauer, 2001; Smith et al. 2001). The approximation is valid under conditions of high buffer concentration, low bound buffer to total buffer ratios, and rapid buffer diffusion. In this simulation, Bauer's excess buffer approximation for the case of one buffer was used, with the parameters listed in Table 1, to calculate steady state free Ca²⁺ concentration as a function of distance from a pore (eqn (9) in, Bauer, 2001). The calculated bound to total ratios for BAPTA and EGTA at the channel were 0.067 and 0.0029, respectively (eqn (12b) in, Bauer, 2001).

Table 1.

Parameters used in temporal and spatial Ca²⁺ gradient simulations

Ca2+ total influx (ions s−1)a	106
Ca2+ diffusion coefficient (μm2 s−1)a	220
Total buffer concentration (mm)c	3
Bulk Ca2+ concentration (μm)c	0.1
	EGTA	BAPTA
Buffer diffusion coefficient (μm2 s−1)b	113	95
Thermodynamic dissociation constant (μm)b	0.20	0.17
Association rate constant (μm−1 s−1)b	1.5	600

^aBauer (2001)

^bSmith et al. (2001).

^cParameters from permeabilized SPOC1 cell experiments.

Statistical analysis

Data collected from the ELLA were normalized to mucin secretion at 0.01 μm Ca²⁺ under control conditions. Data are presented as the mean ± s.e.m., for a specified number of SPOC1 cell cultures, each culture in a given experiment originating from a different passage. ANOVA and Student's t test were used to determine statistical significance between data sets. * indicates P-values < 0.05.

Results

IP₃-induced Ca²⁺ gradients

To test whether significant Ca²⁺ gradients can be generated in SLO-permeabilized, EGTA-buffered SPOC1 cells we used exogenous IP₃ to release Ca²⁺ from internal stores. Cells grown on 48-well culture plates were exposed to IP₃ (0.03–30 μm) in 0.1 μm Ca²⁺ EGTA buffer on one half of the plate. As controls, cells on the other half of the plate were exposed to a series of Ca²⁺ activities (0.01–10 μm) buffered by EGTA. Mucin secretion induced by increases in bulk Ca²⁺ was enhanced in a concentration-dependent manner at activities > 1 μm (Fig. 1A), consistent with previous studies in EGTA-buffered SPOC1 cells and many other secretory cells (Scott et al. 1998; and references therein). IP₃ also caused a concentration-dependent increase in mucin secretion with a maximum response at 0.3 μm, a value close to its reported EC₅₀ (Wojcikiewicz & Luo, 1998). For reasons that are not clear, IP₃ concentrations > 0.3 μm were less effective in stimulating mucin secretion in these cells. Notably, the maximal mucin secretory response to IP₃ (0.3 μm) was approximately equal to that induced by 10 μm bulk Ca²⁺, suggesting that IP₃ induces local Ca²⁺ release either in the vicinity of, or from, mucin secretory granules in EGTA-buffered cells generating a local gradient with a maximal effective concentration of ∼10 μm. Heparin, an IP₃ receptor inhibitor, blocked the mucin secretory response to IP₃ (Fig. 1B), suggesting that the IP₃ effect to release Ca²⁺ is specific.

Figure 1. Effects of IP₃ and bulk Ca²⁺ on SLO-permeabilized, EGTA-buffered SPOC1 cells.

A, cells grown on one half of a 48-well plate were incubated with IP₃ (0.03–30 μm; 0.1 μm Ca²⁺), whereas the other half received different concentrations of Ca²⁺ (0.01–10 μm; no IP₃). Following permeabilization in this and all other experiments, each treatment was applied to triplicate wells and the mucins released were collected and assessed after a 15 min incubation using a lectin-linked, microtitre plate binding assay. In each case, mucin release was normalized to the secretions at 0.1 μm Ca²⁺ and is represented as the mean ± s.e.m. (n = 6 SPOC1 passages). B, inhibition of IP₃-induced mucin release by heparin. Cells were exposed to 0.3 μm IP₃ ± 200 μg ml⁻¹ heparin (n = 6 SPOC1 passages).

Buffer kinetics: relative effects on free Ca²⁺ and mucin secretion

The two Ca²⁺ chelators commonly used in biomedical experiments, EGTA and BAPTA, have similar buffer affinities, in contrast to vastly different binding kinetics (see Table 1). To determine whether the faster binding properties of BAPTA might be used to our advantage, we first simulated the time course of Ca²⁺ binding by the two buffers using starting conditions that mimicked those likely to occur in permeabilized SPOC1 cells under our experimental conditions, i.e. 10 μm free Ca²⁺ and 3 mm buffer. The simulations indicate that addition of BAPTA, with instantaneous mixing, decreases free Ca²⁺ to vanishingly low activities in ∼2.5 μs, a time during which EGTA addition has no significant effect (Fig. 2A). As indicated in the figure, EGTA binds only ∼50% of the total free Ca²⁺ over an extended period of time (150 μs). Therefore, short-lived Ca²⁺ transients such as those that occur in nerve terminals will be effectively buffered by BAPTA, but not EGTA (e.g. see Adler et al. 1991).

Ca²⁺ release events associated with G protein-coupled receptor (GPCR) signalling, by comparison to those at synaptic terminals, generally have much longer durations (tens of seconds). In this case, steady-state Ca²⁺ gradients are likely to be established in the vicinity of Ca²⁺ release sites. Because small, mobile Ca²⁺ buffers will determine the Ca²⁺ concentration profile centred on such sites (see Bauer, 2001; Smith et al. 2001), we simulated Ca²⁺ gradients generated in the vicinity of a channel opening in the presence of EGTA or BAPTA. Under our experimental conditions, BAPTA is predicted to buffer Ca²⁺ much closer to the channel than would EGTA (Fig. 2B). At a distance < 50 nm from the mouth of an open channel, with a flux of 10⁶ ions s⁻¹, BAPTA will buffer Ca²⁺ to levels below 1 μm, whereas a gradient of similar magnitude in EGTA will extend > 350 nm. Hence, BAPTA is an excellent probe for revealing local Ca²⁺ gradients as well as Ca²⁺ transients.

As an initial control for potential buffer effects on mucin release, we tested the effects of elevated Ca²⁺ on mucin secretion from SPOC1 cells permeabilized into EGTA- or BAPTA-based buffers. As shown in Fig. 2C, mucin secretion activated by bulk Ca²⁺ was similar in both EGTA- and BAPTA-buffered conditions, with the exception of 3 μm Ca²⁺ where, though mucin secretion was still elevated in BAPTA-buffered cells, it was diminished relative to EGTA.

The findings in Fig. 1 indicate that IP₃ releases a significant amount of Ca²⁺ near sites of mucin granule exocytosis at the apical membrane in EGTA-buffered, permeabilized cells. In experiments conducted with cells of the same passage, we tested whether BAPTA, in the place of EGTA, would diminish mucin secretion elicited by IP₃. As shown in Fig. 3, mucin secretion in BAPTA-buffered cells was inhibited by 69–97% at all IP₃ concentrations, relative to EGTA.

Figure 3. Effects of rapid Ca⁺ buffering on IP₃-induced mucin secretion from SPOC1 cells.

SLO-permeabilized cells were exposed to IP₃ (0.03–30 μm), buffered by EGTA or BAPTA (0.1 μm Ca²⁺). The EGTA data duplicate those in Fig. 1; the experiments were conducted at the same time, on paired 48-well plates. Two-way ANOVA indicates a significant effect of BAPTA (n = 6 SPOC1 passages).

Lack of buffer effects

Previous studies suggested that BAPTA might have effects independent of Ca²⁺ buffering. For instance, in studies with purified IP₃ receptors from the cerebellum, BAPTA, EDTA and EGTA were shown to inhibit IP₃ binding. The inhibition was competitive between IP₃ and the free forms of the buffer (Richardson & Taylor, 1993). Similarly, free-BAPTA was implicated in an inactivation of PKC in liver macrophages (Dieter et al. 1993). Because BAPTA suppressed mucin secretion slightly, but significantly, at high Ca²⁺ in permeabilized SPOC1 cells (Fig. 2C), we tested for possible direct effects of EGTA and BAPTA on mucin secretion. First, to test whether PKC inactivation might explain the apparent inhibition by BAPTA (Fig. 2C) we used calphostin C and BIMII, in combination, to inhibit PKC (Fig. 4A) during stimulation by high Ca²⁺. The data indicate no effect of the inhibitors at 3 μm Ca²⁺ in EGTA- or BAPTA-buffered SPOC1 cells. Mucin secretion was stimulated, though diminished, with BATPA buffering, relative to EGTA, consistent with data of Fig. 2C. Second, as a more general test of Ca²⁺-independent effects of chelator we measured IP₃-induced mucin release at a constant Ca²⁺ activity of 0.1 μm, over a broad range of EGTA and BAPTA concentrations (0.03–10 mm; Fig. 4B). The results indicate that IP₃ elicited mucin secretion independent of EGTA concentration, and that BAPTA generally abolished the stimulatory effects of IP₃, again, independent of buffer concentration. Note that in this experiment (cf. Fig. 3), the only point where BAPTA failed to abolish IP₃-induced mucin secretion was at 10 μm, a concentration at which Ca²⁺ buffering by BAPTA is likely to be ineffective. Hence, from these studies we conclude that there were no apparent direct effects of either chelator on the secretory response. Importantly, the lack of buffer concentration effects on IP₃-stimulated mucin release in EGTA, or its inhibition by BAPTA, also makes it highly unlikely that the differences observed between the two buffers can be ascribed to differences in permeation of the buffer.

Figure 4. Lack of buffer effects on mucin secretion in SLO-permeabilized SPOC1 cells.

A, cells were exposed to high Ca²⁺ (3 μm), buffered by EGTA or BAPTA (0.1 μm Ca²⁺), ± the PKC inhibitors BIMII (1 μm) and calphostin C (1 μm; n = 4 SPOC1 passages). B, cells were exposed to IP₃ (0.3 μm) and buffered by either EGTA or BAPTA (0.03–10 mm) at a constant 0.1 μm Ca²⁺ (n = 4 SPOC1 passages).

Ca²⁺ dependency of PMA-induced mucin secretion

Because BAPTA suppressed IP₃-induced Ca²⁺ gradients, we used it to probe whether the stimulatory effects of PMA on mucin secretion are also Ca²⁺ dependent. Similar to our previous results (Scott et al. 1998), a concentration of PMA eliciting maximal effects (300 nm) stimulated mucin secretion from EGTA-buffered SPOC1 cells at all Ca²⁺ activities tested, including those clearly sub-basal to normal resting Ca²⁺ (Fig. 5). PMA increased mucin release by ∼100% at low Ca²⁺ activities and up to ∼350% at 3 μm Ca²⁺. Using BAPTA as the buffer, however, reduced PMA-stimulated mucin secretion by 32–58%, relative to EGTA-buffered cells. Experiments using 10 mm EGTA and BAPTA instead of 3 mm yielded similar results (data not shown).

Figure 5. Effects of rapid Ca²⁺ buffering on PMA-induced mucin secretion from SPOC1 cells.

SLO-permeabilized cells were exposed to Ca²⁺ (0.01–10 μm) buffered by EGTA or BAPTA, ± 300 nm PMA. Mucin secretion was normalized to that released by 0.1 μm Ca²⁺ in the absence of PMA (for clarity, controls not shown; compare with Fig. 2C). Two-way ANOVA indicates a significant reduction by BAPTA, as well as a significant effect of Ca²⁺ with both buffers (n = 6 SPOC1 passages).

Recent studies from our laboratory identifying the P2Y₂ agonist- and PMA-sensitive isoforms of PKC in SPOC1 cells suggested that the effects of PMA to stimulate mucin secretion at concentrations > 30 nm are independent of PKC (Abdullah et al. 2003). Hence, we compared the effects of EGTA and BAPTA buffering on concentration-dependent PMA stimulation of mucin secretion from SLO-permeabilized SPOC1 cells. Cells on 48-well cluster plates were buffered (0.1 μm free Ca²⁺) by EGTA on one half of a plate and by BAPTA on the other half, and incubated with 0.3–300 nm PMA. As shown in Fig. 6A, there was a biphasic response of EGTA-buffered SPOC1 cells to PMA. Mucin secretion was stimulated significantly over baseline between 1 and 30 nm PMA, the range within which the phorbol activates PKC (Abdullah et al. 2003). A much stronger degree of stimulation was observed for cells exposed to higher PMA concentrations, as previously shown (Abdullah et al. 1997). Mucin secretion elicited by PMA from BAPTA-buffered cells was inhibited significantly: at the lower PMA concentrations BAPTA abolished secretion and at higher concentrations it was diminished relative to EGTA-buffered cells.

Figure 6. Concentration effects of PMA in SLO-permeabilized SPOC1 cells.

A, cells were stimulated with 0.3–300 nm PMA at 0.1 μm Ca²⁺ in EGTA- or BAPTA-buffered solutions. Two-way ANOVA indicates a significant difference between EGTA and BAPTA (n = 6 SPOC1 passages). B, EGTA-buffered cells were stimulated with 0.3–300 nm PMA at 0.1 μm Ca²⁺ ± heparin (200 μg ml⁻¹). No significant differences were detected (n = 4 SPOC1 passages).

To test whether IP₃-mediated signalling might be responsible for the apparent Ca²⁺ dependency of PMA-stimulated secretion, we challenged SLO-permeabilized, EGTA-buffered SPOC1 cells with PMA in the presence of heparin. Figure 6B shows that heparin had no effect on the PMA response at any concentration (cf. Fig. 1), suggesting that the resulting secretion of mucin is stimulated by an IP₃-independent mechanism.

Discussion

EGTA has been used for nearly a quarter of a century to control Ca²⁺ in permeabilized secretory cells (Baker & Knight, 1981; Bennett et al. 1981). Remarkably, it continues to be the chelator of choice nearly 20 years after the synthesis of BAPTA (Tsien, 1980), a chelator that binds Ca²⁺ nearly 2 orders of magnitude faster despite having a similar affinity (Smith et al. 1984; Harrison & Bers, 1987). Theoretically, these kinetics allow BAPTA, but not EGTA, to buffer Ca²⁺ effectively at the mouths of open Ca²⁺ channels, up to limits imposed by the diffusivity of the chelator (Neher, 1986; Stern, 1992; Bauer, 2001; Smith et al. 2001). For example, our calculations show that under the conditions of our experiments BAPTA buffers Ca²⁺ ∼7-fold closer to the channel than does EGTA (Fig. 2B). The faster Ca²⁺ binding kinetics of BAPTA have been exploited successfully to probe the generation of Ca²⁺ transients and/or local gradients that occur in exocytosis, using permeable acetoxymethylester analogues to load, or a micropipette or patch pipette to inject/dialyse, the buffer into the target cells. For example, BAPTA markedly attenuated transmitter release at the squid giant synapse (Adler et al. 1991), diminished the amplitude of Ca²⁺ transients in mouse spinal cord neurones (Tymianski et al. 1994), and inhibited glucose-induced insulin secretion in pancreatic β-cells (Pertusa et al. 1999), while EGTA either had smaller or no significant effect. We took advantage of BAPTA's ability to buffer Ca²⁺ in near membrane environments to reveal Ca²⁺ dependency in pathways controlling exocytosis in SPOC1 cells previously thought to be independent of Ca²⁺ (Scott et al. 1998).

Ca²⁺ plays a central role in triggering exocytosis in numerous secretory cells; however, the process is particularly poorly defined in airway goblet cells. A useful model is the pancreatic acinar cell where presentation of agonist to the basolateral membrane induces an IP₃-dependent Ca²⁺ release selectively, within the secretory granule region in the apical pole to trigger zymogen granule exocytosis (Ito et al. 1997). Agonist-induced Ca²⁺ release in acinar cells is sensitive to heparin, indicating an involvement of IP₃ (Thorn et al. 1993b), and dialysis of the cell with IP₃ induces transient or oscillatory Ca²⁺ release from IP₃-sensitive intracellular Ca²⁺ stores (Thorn et al. 1993a; reviewed by Berridge, 1993; Petersen, 1992). In goblet cells, IP₃ releases Ca²⁺ from secretory granules (Nguyen et al. 1998).

IP₃-generated Ca²⁺ gradients in SPOC1 cell

IP₃ induced mucin secretion from permeabilized, EGTA-buffered SPOC1 cells in a heparin-sensitive manner (Fig. 1). The magnitude of IP₃-induced mucin release was similar to that induced by 10 μm bulk Ca²⁺ levels. Hence, our results suggest that EGTA allows a local elevation in Ca²⁺ to micromolar levels following its IP₃-mediated release from SPOC1 cell internal stores. This elevation is sufficiently robust and long-lived to initiate and trigger the exocytosis of mucin secretory granules. Mucin release elicited by IP₃ in the presence of BAPTA was inhibited by 69–100% (Figs 3 and 4), compared to EGTA, suggesting that IP₃ releases Ca²⁺ locally, generating Ca²⁺ gradients of a magnitude sufficient to trigger exocytosis. The differences in the mucin secretory response to IP₃ by cells buffered with EGTA or BAPTA appear to be specific to their binding kinetics, and not to be due to non-specific effects of the buffers (Fig. 4).

Ca²⁺ dependency of PMA-induced mucin secretion in SPOC1 cells

PMA and PDBu are classical phorbol ester secretagogues used to activate conventional and novel isoforms of PKC. Although they are often used at, or near, micromolar concentrations, maximal activation of PKC typically occurs at 10–30 nm (Liles et al. 1987; Kazanietz et al. 1993). Mucin secretion from SPOC1 cells, however, is stimulated maximally by PMA at 300 nm (Abdullah et al. 1997); hence, this was the concentration used originally to test the Ca²⁺ independence of PMA effects on mucin release from permeabilized, EGTA-buffered cells (Scott et al. 1998; see Introduction). In the study of Fig. 5, the original observation that 300 nm PMA stimulates mucin release from SPOC1 cells over a range of 0.01–30 μm Ca²⁺ was duplicated successfully. When BAPTA was used as the Ca²⁺ buffer in paired experiments, however, mucin secretion was inhibited substantially indicating that, in fact, PMA-induced mucin secretion has a strong Ca²⁺ dependency. Hence, this result supports the notion that regulated exocytosis from SPOC1 cells is Ca²⁺ dependent, as suggested by the Ca²⁺ dependency of regulated exocytotic pathways, in general (see Chapman, 2002; Sudhof, 2002).

In recent years, proteins other than PKC have been shown to possess the C1 domain necessary to be an effective phorbol ester receptor (e.g. see Kazanietz, 2002). Considering that PKC translocation to the membrane saturates between 10 and 30 nm PMA in SPOC1 cells (Abdullah et al. 2003), the effects of higher phorbol ester levels in SPOC1 cells are clearly PKC independent and may be due to the obligate, C1 domain, exocytotic accessory protein, ubMUNC13-2 (Koch et al. 2000; Rhee et al. 2002) that is expressed in these cells. Notably, the PMA concentration–effect studies conducted in BAPTA-based buffers were consistent with this scenario (Fig. 6A): at concentrations below 30 nm the effects of PMA in SPOC1 cells buffered by BAPTA were blocked completely, whereas EGTA proved permissive. PMA concentrations > 30 nm effectively stimulated mucin secretion from BAPTA-buffered cells to levels similar to those elicited in EGTA, except for the highest PMA concentration (300 nm) where mucin release in BAPTA was inhibited slightly. We speculate that the stimulatory effects of low PMA levels (≤ 30 nm) in EGTA-buffered cells correspond to the activation of PKC, which most likely phosphorylates MARCKS to initiate the disruption of actin cortical microfilaments (Trifaro et al. 2000). The stimulation that occurs in both EGTA- and BAPTA-buffered cells at higher PMA concentrations most likely reflects diffusion distances that are very short, a scenario consistent with the activation of ubMUNC13-2 to prime mucin secretory granules docked at the plasma membrane (Brose et al. 2000; Martin, 2002). BAPTA's superior ability to buffer Ca²⁺ near channel openings might explain the generally diminished secretion in BAPTA-buffered cells, relative to EGTA-buffered cells, at high PMA concentrations (Fig. 6A). As illustrated in Fig. 2B, under the conditions of our experiments BAPTA is expected to buffer Ca²⁺ to below 1 μm beyond a ∼50 nm radius of channel openings. Within this radius, Ca²⁺ gradients exceeding 10 μm are expected to develop and are likely to activate Ca²⁺-dependent proteins. Small distances of this order are appropriate to the plasma and secretory granule membranes as they approach one another during the exocytotic process. Interestingly, these molecular distances may explain why secretion occurred in the face of rapid Ca²⁺ buffering by BAPTA during stimulation by both IP₃ (Fig. 3) and PMA (Figs 5 and 6).

Chief among the unanswered questions in this scenario are the signal initiating Ca²⁺ release in PMA-treated SPOC1 cells and the source of the Ca²⁺. PMA activation of PKC in some cells results in a release of Ca²⁺ from internal stores through an undefined mechanism (e.g. see Xuan et al. 1994), and such could be the case for SPOC1 cells exposed to PMA concentrations ≤ 30 nm. The extra stimulatory effects of PMA at concentrations > 30 nm, however, indicates a PKC-independent mechanism (Abdullah et al. 2003), and the lack of inhibition of PMA-induced secretion by heparin (Fig. 6B) effectively excludes IP₃ as the potential signalling molecule. The Ca²⁺ store could be cisternae of the endoplasmic reticulum located near mucin granules, as it is in gonadotrophs (Tse et al. 1997), or potentially more interestingly, it could be the mucin secretory granule itself (Nguyen et al. 1998; and see Petersen, 1996). A granule-based store raises the intriguing possibility that Ca²⁺ is released from the granule as it interacts physically with the plasma membrane during the pre-exocytotic events involved in granule docking, priming and formation of the exocytotic pore complex. Precedence for such a mechanism exists in yeast where Ca²⁺ is released from the lumen of the vacuole as it interacts with its exocytotic docking site to stimulate the final steps of vacuolar secretion (Peters & Mayer, 1998).

In conclusion, by virtue of its slow binding kinetics EGTA is permissive with respect to the generation of local Ca²⁺ gradients in permeabilized SPOC1 cells, whereas BAPTA is a much more efficient Ca²⁺ buffer and inhibits mucin secretion. Hence, the SPOC1 cell mucin secretory response to PMA we observed previously in permeabilized, EGTA-buffered cells and proposed to be independent of Ca²⁺ does, in fact, possesses a definite Ca²⁺ dependency. What remains to be tested rigorously for Ca²⁺ dependency in other secretory cells are the cAMP-induced responses proposed to be Ca²⁺ independent (Koh et al. 2000), as well as other studies in which measurements of whole cell Ca²⁺ and exocytosis failed to exhibit the correspondence expected for Ca²⁺-dependent events (e.g. Neher & Almers, 1986). A major problem that needs to be resolved with such studies is the measurement of local Ca²⁺ – because the studies above were based on the use of fluorescent Ca²⁺ indicators using wide-field epifluorescence microscopy, the small signal from local Ca²⁺ gradients that might have occurred could have been buried in the noise of the larger signal being recorded from the whole cell. Ca²⁺ sparks, for instance, were not observed until the advent of confocal microscopy (Cannell et al. 1994). Clearly, before such data can be accepted as indicating genuine Ca²⁺-independent exocytosis, the measurements need to be repeated using confocal microscopy and the appropriate controls.