Feedback activation of phospholipase C via intracellular mobilization and store-operated influx of Ca²⁺ in insulin-secreting β-cells

The ubiquitous enzyme phosphoinositide-specific phospholipase C (PLC) plays a key role in signal transduction by catalyzing the hydrolysis of the membrane phospholipid phosphatidylinositol-4,5-bisphosphate (PIP₂) to inositol-1,4,5-trisphosphate (IP₃) and diacylglycerol in response to various receptor stimuli. IP₃ mediates rapid mobilization of Ca²⁺ from the endoplasmic reticulum (ER), whereas diacylglycerol stimulates protein kinase C (Berridge et al., 2003). There are four families of PLC (PLC-β, -γ, -δ and -ϵ) with 11 different isoforms (Rhee, 2001). PLC-β is mainly activated by heterotrimeric G-proteins, PLC-γ by tyrosine kinases and PLC-ϵ by the small GTPase Ras. The activation mechanism for the δ isoforms is less clear, but it may be the elevation of the cytoplasmic Ca²⁺ concentration ([Ca²⁺]_i) alone, as this isoform is particularly sensitive to Ca²⁺ (Rhee, 2001).

In pancreatic β-cells, PLC mediates the potentiating action on insulin secretion of many hormones and neurotransmitters. For example, it is well established that cholinergic stimulation of insulin secretion is associated with accumulation of IP₃ and diacylglycerol (Gilon and Henquin, 2001). This effect is due to activation of muscarinic M3 receptors, which, as in other tissues, are believed to stimulate PLC-β via the Gq family of heterotrimeric G proteins. It was recognized early on that phospholipid hydrolysis and inositol phosphate production after cholinergic stimulation was larger in β-cells maintained in Ca²⁺-containing medium than in Ca²⁺-deficient medium (Best, 1986; Biden et al., 1987; Garcia et al., 1988). This phenomenon is poorly understood but may be explained by Ca²⁺-mediated activation of PLC (Biden et al., 1987). Analyses of [Ca²⁺]_i responses in β-cells have demonstrated that stimulation with the muscarinic-receptor agonist carbachol is associated with a biphasic increase of [Ca²⁺]_i with a rapid peak followed by a sustained plateau, sometimes with superimposed oscillations (Gylfe, 1991; Liu and Gylfe, 1997). Whereas the first phase reflects rapid IP₃-mediated mobilization of intracellular Ca²⁺, the second phase depends on Ca²⁺ influx through store-operated channels in the plasma membrane (Liu and Gylfe, 1997). As all PLC isoforms require Ca²⁺, it is possible that such receptor-induced Ca²⁺ signals result in feedback activation of the enzyme.

Owing to difficulties in measuring PLC activity in individual cells, little is known about how physiological changes of [Ca²⁺]_i influence the activity of the lipase. Most studies of PLC have employed radiotracer techniques in populations of cells, but with the advent of phosphoinositide-specific fluorescent biosensors, it has become possible to measure the enzyme activity in individual living cells (Stauffer et al., 1998; Varnai and Balla, 1998). The most commonly used single-cell biosensor for PLC activity is the pleckstrin homology (PH) domain from PLCδ1 fused to the green fluorescent protein (PH_PLCδ-GFP), which binds PIP₂ and IP₃ with high affinity and specificity (Stauffer et al., 1998; Varnai and Balla, 1998). In unstimulated cells, the construct is therefore located mainly to the plasma membrane. Upon PIP₂ hydrolysis and formation of IP₃, PH_PLCδ-GFP dissociates from the membrane and binds to IP₃ in the cytoplasm. This PH_PLCδ-GFP translocation can be used as an indicator of PLC activity. Using an evanescent wave microscopy approach for simultaneous measurements of PLC activity and [Ca²⁺]_i, we recently demonstrated that PLC activity in the electrically excitable insulin-secreting β-cell is tightly controlled by [Ca²⁺]_i elevations that result from voltage-dependent Ca²⁺ entry (Thore et al., 2004). In the present paper, we test the hypothesis that elevations of [Ca²⁺]_i following receptor stimulation result in positive feedback activation of PLC. After stimulation of endogenous muscarinic receptors in insulin-secreting cells, two distinct phases of PLC activation were resolved: an initial transient phase that is amplified by mobilization of intracellular Ca²⁺; and a second sustained phase, which is dependent on Ca²⁺ entry through store-operated channels in the plasma membrane. Moreover, activation of PLC by Ca²⁺ mobilized from intracellular stores was found to occur in primary mouse pancreatic β-cells after stimulation with the insulinotropic hormone glucagon.

Materials of analytical grade and deionized water were used. Thapsigargin and the acetoxymethyl esters of the Ca²⁺ indicator Fura Red and the Ca²⁺ chelator BAPTA were from Molecular Probes (Eugene, OR). HEPES was obtained from Roche Diagnostics (Bromma, Sweden), and Invitrogen (Carlsbad, CA) provided foetal calf serum, DMEM and RPMI 1640 culture media. Cyclopiazonic acid (CPA) and SKF96365 were from Calbiochem (San Diego, CA); 2-aminoethyl diphenylborate (2-APB) was from Aldrich (Gillingham, UK). Diazoxide was a kind gift from Schering-Plough (Kenilworth, NJ). All other chemicals were from Sigma (St Louis, MO). Plasmids encoding the fusion construct between the PH domain of PLCδ1 and GFP (Stauffer et al., 1998), and GFP targeted to the plasma membrane via covalent lipid modification (GFP-CAAX) were kindly provided by Professor Tobias Meyer, Stanford University.

Insulin-secreting MIN-6 β-cells (passage 16-30) (Miyazaki et al., 1990) were cultured at 37°C in a humidified atmosphere containing 5% CO₂ in DMEM containing 25 mM glucose and supplemented with 15% foetal calf serum, 2 mM glutamine, 70 μM 2-mercaptoethanol, 100 units/ml penicillin and 100 μg/ml streptomycin. After plating onto 25 mm coverslips at a density of 1.5×10⁵/ml, the cells were transiently transfected with 2 μg of plasmid DNA with Lipofectamine 2000 (Invitrogen) in a 1:2.5 DNA:lipid ratio (according to the manufacturer's protocol) and further cultured for 24-48 hours.

Mouse pancreatic β-cells were obtained from collagenase-isolated islets of Langerhans from ob/ob mice. Free cells were prepared by shaking the islets in a Ca²⁺-deficient medium (Lernmark, 1974). The cells were then suspended in RPMI 1640 medium supplemented with 10% foetal calf serum, 100 IU/ml penicillin, 100 μg/ml streptomycin and 30 μg/ml gentamycin, and allowed to attach to 25 mm coverslips for 1-3 days in culture at 37°C in a humidified atmosphere of 5% CO₂. The ob/ob mouse islets contain more than 90% β-cells (Hellman, 1965), which respond normally to glucose and to other regulators of insulin release (Hahn et al., 1974). Transfection of the primary mouse β-cells was performed by electroporation of in vitro transcribed mRNA, as outlined below.

As an alternative to conventional plasmid and viral methods commonly applied to transfect pancreatic β-cells, we used an mRNA transfection technique (Yokoe and Meyer, 1996) to express PH_PLCδ fused to yellow fluorescent protein (YFP) in the ob/ob mouse β-cells. First, we generated the vector pCS2-PH_PLCδ-YFP. The PH domain from PH_PLCδ-GFP was PCR amplified using primers 5′-TTTTGGATCCACCATGGGCCTACAGGATGATGAGGA-3′ (forward) and 5′-TTTTTCTAGAGCCTGGATGTTGAGCTCCTTCAG-3′ (reverse) containing BamHI and XbaI restriction sites, respectively (underlined). The product was subsequently ligated into the corresponding sites of the transcription vector pCS2-YFP (Tengholm and Meyer, 2002). The plasmid was linearized with NotI and subsequent in vitro transcription with SP6 RNA polymerase and poly-A tail addition were performed according to the manufacturers' protocol using commercial kits [mMESSAGE mMACHINE, and Poly(A) Tailing kit, respectively, Ambion Europe]. After purification of the mRNA by column chromatography (RNeasy, Qiagen), the eluent was dried and the mRNA dissolved at 2 μg/μl in phosphate-buffered saline (PBS; pH 7.00).

Electroporation of the adherent mouse pancreatic β-cells was performed using a custom-built small-volume electroporator (Teruel and Meyer, 1997). After replacement of the medium with electroporation buffer (PBS supplemented with 20 mM glucose at pH 7.00), 1-2 μl of the 2 μg/μl mRNA sample was applied to the ∼15 μl electroporation chamber. Electroporation was performed at 220 V/cm, using three voltage pulses, each 30 mseconds and 40 seconds apart. After transfection, the electroporation buffer was replaced with RPMI 1640 medium and the cells were kept in culture for 10-20 hours to allow for expression of PH_PLCδ-YFP.

Before experiments, the cells were transferred to a buffer containing 125 mM NaCl, 4.8 mM KCl, 1.28 mM CaCl₂, 1.2 mM MgCl₂, 3 mM glucose and 25 mM HEPES with pH adjusted to 7.40 with NaOH. Where indicated, transfected cells were loaded with the Ca²⁺ indicator Fura Red by a 40 minute incubation at 37°C with 10 μM of its acethoxymethyl ester. The coverslips were used as exchangeable bottoms of a 50 μl open chamber and superfused with buffer at a rate of 0.3 ml/minute. All experiments were performed at 37°C.

GFP, YFP and Fura Red fluorescence was measured using an evanescent wave microscopy setup built around an Eclipse TE2000 microscope (Nikon) as previously described (Thore et al., 2004). The 488 nm beam of an argon ion laser (Creative Laser Production, Munich, Germany) was homogenized, expanded and refocused onto the periphery of the back focal plane of a 60× 1.45-NA objective (Nikon) to achieve total internal reflection at the interface between the coverslip and the adherent cells. The fluorescence excited by the evanescent field was detected using an IEEE1394 Orca-ER camera (Hamamatsu) controlled by MetaMorph or MetaFluor software (Universal Imaging). Selection of emission wavelength was made with interference (525/25 nm for GFP; 550/30 for YFP) and long-pass (>630 nm for Fura Red) filters (Chroma Technology) mounted in a Lambda 10-2 filter wheel (Sutter Instruments) capable of changing positions within 60 mseconds. Images (or image pairs) were acquired every 5 seconds, except for in the experiments in Fig. 4, where image pairs were acquired at ∼1.5 Hz in the data-streaming mode of MetaFluor. To minimize exposure of the cells to the potentially harmful laser light, the beam was blocked by an electronic shutter (Sutter Instruments) between image captures.

Image analysis was made with MetaMorph, MetaFluor (Universal Imaging) or ImageJ (W. S. Rasband, National Institutes of Health, rsb.info.nih.gov/ij) softwares. GFP, YFP and Fura Red fluorescence intensities are expressed as changes relative to initial fluorescence (ΔF/F_o) after subtraction of background. All data are presented as mean values±s.e.m. Statistical significances were evaluated using Student's t-test.

When insulin-secreting MIN6 β-cells were stimulated with 100 μM of the muscarinic-receptor agonist carbachol, there was a rapid and pronounced activation of PLC, detected as a decrease in evanescent wave excited PH_PLCδ-GFP fluorescence (Fig. 1A). This activation consisted of two phases: an initial rapid peak (16±1% loss of fluorescence reached at 31±2 seconds; n=33), followed within 3 minutes by a sustained plateau (6.4±0.8% below baseline, n=33) that lasted throughout the stimulation period (more than 10 minutes; Fig. 1B,F). After washout of carbachol, the fluorescence returned to initial intensity, sometimes after temporarily overshooting the baseline. When the cells were stimulated in Ca²⁺-deficient medium containing 2 mM EGTA, the initial peak response to carbachol was indistinguishable from that in control cells (15.5±1.0% decrease, n=10), but there was no sustained plateau and fluorescence gradually returned to basal levels (Fig. 1C,F). Likewise, when extracellular Ca²⁺ was removed during the plateau phase of carbachol stimulation, fluorescence rapidly returned towards baseline. This effect was readily reversible with restoration of the plateau upon re-addition of Ca²⁺ (Fig. 1D). Irrespective of whether Ca²⁺ was present, carbachol produced no effect in control experiments with cells expressing GFP alone in the cytoplasm or targeted to the plasma membrane (data not shown). To verify that the Ca²⁺ requirement for sustained receptor-triggered PLC activity involved influx of the ion, we applied the non-selective Ca²⁺ channel inhibitor La³⁺. At 0.5 mM, this ion strongly enhanced evanescent wave excited PH_PLCδ-GFP fluorescence (45±4% increase of fluorescence, n=18; Fig. 1E, inset). A similar effect was observed with GFP when targeted to the plasma membrane (data not shown), indicating that La³⁺ might affect membrane properties and/or cell adhesion. Nevertheless, carbachol still triggered translocation. Although the initial peak translocation was unaffected by La³⁺, there was no sustained plateau (Fig. 1E,F). Thus, receptor-triggered PLC activity is enhanced by Ca²⁺ influx through the plasma membrane.

We investigated the nature of the Ca²⁺ influx pathway for sustained PLC activation in response to muscarinic-receptor stimulation. As carbachol-mediated stimulation of β-cells is associated with slight depolarization, which may activate voltage-dependent influx of Ca²⁺ through the plasma membrane (Gilon and Henquin, 2001), we tested the possible involvement of such a mechanism for the sustained PLC activation. The K_ATP channel opener diazoxide, which hyperpolarizes the β-cells, did not affect the plateau PH_PLCδ-GFP fluorescence (Fig. 2A,E). Likewise, 10 μM of the L-type Ca²⁺ channel inhibitor nifedipine was without effect on PH_PLCδ-GFP fluorescence (Fig. 2B,E), although the same concentration suppressed the elevation of [Ca²⁺]_i induced by depolarization with 90 mM KCl (data not shown), indicating that voltage-gated Ca²⁺ influx is not required for the sustained carbachol-induced activation of PLC.

Receptor-induced mobilization of intracellular Ca²⁺ in β-cells, as in most other cells, is associated with activation of Ca²⁺ entry through a store-operated pathway (Liu and Gylfe, 1997; Dyachok and Gylfe, 2001). To test the involvement of such a mechanism, we applied the relatively specific store-operated Ca²⁺ channel inhibitors 2-APB and SKF96365 at 100 and 50 μM, respectively. Both agents markedly reduced the sustained PH_PLCδ-GFP translocation in response to 100 μM carbachol (Fig. 2B,C,E,F), but they had no effect on the fluorescence of GFP in the cytoplasm (data not shown) or targeted to the plasma membrane (a representative experiment with 2-APB is shown in Fig. 2D). When carbachol was applied after 2-APB, there was transient initial PH_PLCδ-GFP translocation of the same magnitude as in control cells (18±4% fluorescence change, n=7; compare with 20±5% in control, n=7), but no sustained plateau (Fig. 2F).

To verify the role of store-operated Ca²⁺ entry for activation of PLC, we compared the influence of extracellular Ca²⁺ on membrane PH_PLCδ-GFP fluorescence under control conditions and after activating this Ca²⁺ influx pathway by depletion of the intracellular Ca²⁺ stores. Under control conditions, Ca²⁺ removal with addition of 2 mM EGTA and the following reintroduction of 1.28 and 2.56 mM of the ion resulted in less than 2% changes in PH_PLCδ-GFP fluorescence (n=9; Fig. 2G). Subsequent activation of the store-operated pathway by addition of 100 μM of the sarco(endo)plasmic reticulum Ca²⁺ ATPase (SERCA) inhibitor CPA was without significant effect on PH_PLCδ1-GFP fluorescence in Ca²⁺-deficient medium. By contrast, there was an immediate translocation of PH_PLCδ-GFP upon reintroduction of 1.28 mM Ca²⁺ to the medium (5.4±0.7% loss of fluorescence, n=9, P<0.01 for difference from Ca²⁺ addition before CPA; Fig. 2G). No further change of fluorescence was observed when raising Ca²⁺ to 2.56 mM. The Ca²⁺ effect was specific, as there was no change in the fluorescence in control cells expressing cytoplasmic or membrane-targeted GFP alone (data not shown). Taken together, the results indicate that sustained PLC activation depends on Ca²⁺ influx through store-operated channels.

In view of the potent effect of Ca²⁺ influx on PLC activity, we investigated the influence of Ca²⁺ mobilized from intracellular stores on receptor-triggered PLC activity. Simultaneous recording of PLC activity with PH_PLCδ-GFP and [Ca²⁺]_i with Fura Red demonstrated that the transient PLC activation induced by carbachol in Ca²⁺-deficient medium containing EGTA was associated with a rapid and pronounced spike of [Ca²⁺]_i (Fig. 3A). Increasing the Ca²⁺ buffering capacity of the cytoplasm by loading the cells with 1 mM of the acetoxymethyl ester of the Ca²⁺ chelator BAPTA resulted in altered response patterns for both [Ca²⁺]_i and PH_PLCδ-GFP translocation (Fig. 3B). The [Ca²⁺]_i response had a lower amplitude and longer duration in the BAPTA-loaded cells (Fig. 3B,E), and similar differences were observed for PH_PLCδ-GFP translocation (Fig. 3B,D), indicating that the [Ca²⁺]_i response is important for PLC activation kinetics. Further support for this idea was obtained from experiments in which intracellular Ca²⁺ stores had been depleted by SERCA inhibition with 100 μM CPA (Fig. 3C-E) or 1 μM thapsigargin (Fig. 3D,E). Accordingly, both agents not only abolished the [Ca²⁺]_i elevation induced by carbachol, but also strongly suppressed the PH_PLCδ-GFP translocation (Fig. 3C-E). These findings indicate that PLC activation in response to muscarinic-receptor stimulation is enhanced by positive feedback from Ca²⁺ that is mobilized from intracellular stores.

The requirement for intracellular Ca²⁺ mobilization for maximal PLC activation in response to carbachol raises the issue of whether intracellular Ca²⁺ release is sufficient for activation of PLC. This hypothesis was tested in primary mouse pancreatic β-cells, which show a Ca²⁺-induced Ca²⁺ release mechanism mediated by protein kinase A-dependent sensitization of IP₃ receptors (Liu et al., 1996; Dyachok and Gylfe, 2004). Mouse β-cells were transfected with PH_PLCδ-YFP and loaded with the Ca²⁺ indicator Fura Red. When the cells were exposed to medium containing 20 mM glucose and the voltage-gated Ca²⁺ influx was prevented by a combination of 250 μM diazoxide and 50 μM of the L-type Ca²⁺ channel blocker methoxyverapamil, [Ca²⁺]_i was low and stable, and evanescent wave-excited PH_PLCδ-YFP exhibited steady fluorescence. As previously described (Liu et al., 1996; Dyachok and Gylfe, 2004), addition of glucagon resulted in the appearance of pronounced spikes of [Ca²⁺]_i occurring from the baseline. Simultaneous recording of PLC activity demonstrated that the high [Ca²⁺]_i spikes were paralleled by transient drops of PH_PLCδ-YFP fluorescence (7.8±0.7 % loss of fluorescence, n=12; Fig. 4A). When these events were captured at a high time resolution (>1.5 Hz), it became evident that the [Ca²⁺]_i elevations preceded each increase of PLC activity by 0.85±0.08 seconds (difference in time to half-maximal change of fluorescence; n=12; Fig. 4B). These results indicate that intracellular Ca²⁺ mobilization can trigger PLC activation and that this mechanism operates in primary pancreatic β-cells.

In the present study, we used a novel evanescent wave microscopy approach to investigate the involvement of Ca²⁺ in regulating receptor-triggered PLC activation. Dissociation of PH_PLCδ-GFP from the plasma membrane has become a well-established single-cell assay for PIP₂ hydrolysis and IP₃ formation (Stauffer et al., 1998; Varnai and Balla, 1998; Hirose et al., 1999), now used as a measure of PLC activity in insulin-secreting β-cells. Detection of plasma membrane-associated fluorescence with evanescent wave microscopy, rather than with confocal or conventional epifluorescence microscopy, has several advantages, including lower background and less photobleaching and phototoxicity (Steyer and Almers, 2001). Using cells loaded with a fluorescent Ca²⁺ indicator allows simultaneous measurements of [Ca²⁺]_i and PLC activation dynamics. Using this approach, we recently showed that both membrane depolarization and the concomitant voltage-dependent influx of Ca²⁺ are sufficient to trigger PLC activation in insulin-secreting cells (Thore et al., 2004).

We now extend these findings by showing that also PLC activity triggered by a muscarinic-receptor agonist is tightly regulated by Ca²⁺. Our data indicate that Ca²⁺ exerts pronounced amplification of both the initial and sustained receptor-triggered PLC activity. Whereas the initial activation of the enzyme in response to carbachol was unaffected by omission of extracellular Ca²⁺, it was markedly suppressed after depletion of Ca²⁺ from the ER. The latter effect is probably explained by the failure of carbachol to mobilize Ca²⁺ from intracellular stores. A similar effect was thus observed when the [Ca²⁺]_i elevation in response to carbachol was prevented by increasing the cytoplasmic Ca²⁺ buffering capacity with BAPTA. Depletion of ER Ca²⁺ has previously been found to suppress α1B-adrenoceptor-mediated oscillations of IP₃ in CHO cells, indicating a role for Ca²⁺ feedback on PLC for periodic generation of IP₃ (Young et al., 2003). Direct support for the idea that Ca²⁺ mobilization from intracellular stores enhances PLC activity was now provided by the finding in primary mouse pancreatic β-cells that cAMP-sensitized intracellular Ca²⁺ mobilization is rapidly followed by activation of PLC. It is important to note that the glucagon receptor does not activate PLC in β-cells (S.T. and A.T., unpublished) and that the pronounced transients of [Ca²⁺]_i seen in this cell type after stimulation with cAMP-elevating agents is due to Ca²⁺-induced Ca²⁺ release via PKA-mediated sensitization of IP₃ receptors (Liu et al., 1996; Dyachok and Gylfe, 2004).

Whereas the initial PLC activation was amplified by intracellular Ca²⁺ mobilization, the second sustained phase of PLC activity after receptor stimulation crucially depended on Ca²⁺ influx from the extracellular medium. Although carbachol under some conditions may depolarize the β-cell sufficiently to reach the activation threshold for voltage-dependent Ca²⁺ entry (Gilon and Henquin, 2001), the involvement of such a mechanism is not required, as neither hyperpolarization with diazoxide nor direct inhibition of the Ca²⁺ channels with nifedipine had any effect on the sustained PLC activity. Instead, the PLC activity was suppressed by La³⁺ and commonly used inhibitors of store-operated Ca²⁺ channels.

None of the currently available inhibitors of the store-operated Ca²⁺ influx pathway is entirely specific. Although 2-APB was originally described as a membrane-permeable IP₃-receptor inhibitor (Maruyama et al., 1997), later studies have shown that this effect is weak and that, instead, 2-APB is a reliable inhibitor of store-operated Ca²⁺ entry in various types of cells, including pancreatic β-cells (Bootman et al., 2002; Dyachok and Gylfe, 2001). If the currently observed effect of 2-APB on PH_PLCδ-GFP fluorescence were due to inhibition of IP₃ receptors, the early and late phase of the carbachol response should have been affected equally. However, 2-APB only inhibited the late sustained PLC activation upon carbachol-mediated stimulation, consistent with a negligible effect on IP₃ receptors. The inhibition of PLC activity with unrelated blockers of store-operated channels and the stimulation obtained after activation of store-operated Ca²⁺ entry together support the conclusion that this pathway is involved in receptor-triggered PLC activation in insulin-secreting cells. In this context, it is worth noting that an inhibitory effect of 2-APB on acetylcholine-induced IP₃ production in pancreatic acinar cells was attributed to a novel unknown action of this drug (Wu et al., 2004). In view of the present data, this observation may represent suppression of IP₃ production after inhibition of store-operated Ca²⁺ entry.

The maintenance of PLC activity by store-operated Ca²⁺ entry can explain early observations that carbachol-stimulated inositol phosphate production is larger in β-cells maintained in Ca²⁺-containing than in Ca²⁺-deficient medium (Best, 1986; Biden et al., 1987; Garcia et al., 1988). Regulation of receptor-triggered PLC activity by Ca²⁺ entry has been described also in other types of cells. Using CHO cells expressing heterologous G-protein-coupled receptors, Nash et al. (Nash et al., 2001; Nash et al., 2002) have reported that Ca²⁺ influx stimulates PLC activity triggered by muscarinic M3 receptors, α1B adrenoceptors and mGlu1a metabotropic glutamate receptors, but inhibits that triggered by mGlu5 receptors. Receptor-activated Ca²⁺ entry has also been reported to promote PLC activity in bradykinin-stimulated PC12 cells (Kim et al., 1999), in B-cell receptor-ligated DT40 B lymphocytes (Nishida et al., 2003) and in rabbit gastric smooth muscle cells stimulated with some G_i/o-coupled receptor agonists (Murthy et al., 2004). These findings suggest that store-operated Ca²⁺ entry in various types of cells, including insulin-secreting β-cells, serves to amplify Ca²⁺ signalling not only by elevating [Ca²⁺]_i and replenishing intracellular stores, but also by directly stimulating PLC.

Although store-operated Ca²⁺ entry in pancreatic β-cells causes a rather modest elevation of [Ca²⁺]_i (Liu and Gylfe, 1997; Dyachok and Gylfe, 2001), it was associated with pronounced activation of PLC. Such high Ca²⁺ sensitivity may be explained by increased susceptibility of PLC to regulation by Ca²⁺ after activation by G-proteins (Rana and Hokin, 1990). It is also possible that PLC colocalizes with the sites for Ca²⁺ entry and senses the high Ca²⁺ concentration close to the store-operated channels. Indeed, via a PDZ domain-containing scaffold protein, some PLC isoforms have been found to associate with the mTRP4 channel, which forms store-operated channels in several types of cells (Tang et al., 2000).

It still remains to be clarified which PLC isoforms account for the different phases of receptor activation. All PLC isoforms depend on Ca²⁺ for activity (Rhee, 2001) and PLCδ has been suggested to be most sensitive (Allen et al., 1997). The store-operated Ca²⁺ entry following bradykinin-induced PLCβ activation in PC12 cells (Kim et al., 1999) and that induced by G_i/o-coupled receptor agonists in smooth muscle cells (Murthy et al., 2004) were thus found to stimulate PLCδ1, whereas B-cell receptor signalling in DT40 lymphocytes is amplified by Ca²⁺-mediated activation of PLCγ2 (Nishida et al., 2003). When overexpressed in MIN6 β-cells, the PLC-β1 and -δ1 isoforms were reported to be equally sensitive to elevation of [Ca²⁺]_i (Ishihara et al., 1999), while the overexpressed β1, but not the δ1, isoform was stimulated by [Ca²⁺]_i elevation in insulin-secreting RINm5F cells (Kelley et al., 2001).

In summary, we demonstrate that activation of PLC by endogenous muscarinic receptors in electrically excitable insulin-secreting β-cells is enhanced by positive feedback from Ca²⁺ entering the cytoplasm from intracellular Ca²⁺ stores and via store-operated channels in the plasma membrane (Fig. 5). Agonist binding to the G-protein-coupled receptor leads to partial activation of PLC and production of sufficient amounts of IP₃ to trigger Ca²⁺ release from the ER. The resulting elevation of [Ca²⁺]_i leads to marked amplification of PLC activity with further IP₃ production and elevation of [Ca²⁺]_i. The reduction of Ca²⁺ in the ER leads to opening of store-operated channels in the plasma membrane with entry of Ca²⁺, which also stimulates PLC and serves to maintain enzyme activity during sustained stimulation. Finally, we have shown that [Ca²⁺]_i spikes occurring as a result of PKA-mediated sensitization of IP₃ receptors, induce transient activation of PLC in primary mouse pancreatic β-cells. Thus, amplification of receptor signalling by feedback activation of PLC is involved in the physiological regulation of insulin secretion by hormones and neurotransmitters.

The authors are indebted to Professor Tobias Meyer (Stanford University) for generously making an electroporator available and for providing cDNA constructs. Professor Jun-Ichi Miyazaki (Osaka University) is acknowledged for providing MIN6 β-cells and Ms Elina Sandwall for skilful technical assistance. This work was supported by grants from Åke Wiberg's foundation, Carl Trygger's Foundation for Scientific Research, the Family Ernfors Fund, the European Foundation for the Study of Diabetes/Novo Nordisk, the Novo Nordisk Foundation, the Swedish Diabetes Association, the Swedish Research Council (32X-14643, 32BI-15333, 32P-15439 and 12X-6240) and the Wenner-Gren Foundations.