Phentolamine Inhibits Exocytosis of Glucagon by G i2 Protein-dependent Activation of Calcineurin in Rat Pancreatic a -Cells*

Capacitance measurements were used to investigate the molecular mechanisms by which imidazoline compounds inhibit glucagon release in rat pancreatic a -cells. The imidazoline compound phentolamine re-versibly decreased depolarization-evoked exocytosis > 80% without affecting the whole-cell Ca 2 1 current. During intracellular application through the recording pipette, phentolamine produced a concentration-dependent decrease in the rate of exocytosis (IC 50 5 9.7 m M ). Another imidazoline compound, RX871024, exhib-ited similar effects on exocytosis (IC 50 5 13 m M ). These actions were dependent on activation of pertussis toxin-sensitive G i2 proteins but were not associated with stim- ulation of ATP-sensitive K 1 channels or adenylate cy-clase activity. The inhibitory effect of phentolamine on exocytosis resulted from activation of the protein phosphatase calcineurin and was abolished by cyclosporin A and deltamethrin. Exocytosis was not affected by intracellular application of specific a 2 , I 1 , and I 2 ligands.

Thirty years has elapsed since the initial demonstration that the imidazoline compound phentolamine stimulated glucoseinduced insulin release in humans (1)(2)(3). Good evidence exists that the insulinotropic effects of imidazoline compounds do not result from antagonism of ␣ 2 -adrenergic receptors but rather from inhibition of ATP-sensitive K ϩ (K ATP ) 1 channels in the ␤-cell plasma membrane (6 -9), resulting in membrane depo-larization, stimulation of Ca 2ϩ influx, and exocytosis. In addition, imidazoline compounds also stimulate insulin release by a direct interaction with the exocytotic machinery (10).
Recent evidence suggests that imidazoline compounds stimulate not only insulin release but also somatostatin release while suppressing glucagon secretion (11). The mechanism underlying the inhibitory action of imidazoline compounds on glucagon release is not clear but may involve either a direct or a paracrine effect on the ␣-cells (11)(12)(13). Here we have combined the patch clamp technique with capacitance measurements of exocytosis to explore the effects of different imidazoline compounds on exocytosis in single rat pancreatic ␣-cells. We thereby provide the first direct evidence that imidazoline compounds inhibit Ca 2ϩ -dependent exocytosis of glucagon via G i2 -dependent activation of the serine/threonine protein phosphatase calcineurin.

EXPERIMENTAL PROCEDURES
Preparation of Islets and ␣-Cells-Male Lewis rats (250 -300 g; Møllegaard, Lille Skensved, Denmark) were anesthetized by pentobarbital (100 mg/kg intraperitoneally), and the pancreas was removed. The experimental procedures were approved by the local ethical committee. Islets were isolated by collagenase digestion and dispersed into single cells using dispase. Pancreatic ␣-cells were separated by fluorescenceactivated cell sorting as described elsewhere (14). Based on the hormone contents and their glucose sensitivity, we estimate that the preparations contain Ͼ80% ␣-cells and Ͻ3% ␤-cells (14,15). The cell suspension was plated on 35-mm diameter Petri dishes and incubated in a humidified atmosphere for up to 3 days in RPMI 1640 tissue culture medium (Life Technologies Ltd., Paisley, United Kingdom) supplemented with 10% (v/v) heat-inactivated fetal calf serum, 100 IU/ml penicillin, and 100 g/ml streptomycin.
Electrophysiology-Pipettes were pulled from borosilicate glass, coated with Sylgard near their tips, and fire-polished. When filled with pipette solutions, the electrodes had a resistance of 3-4 megaohms. The whole-cell K ATP conductance was estimated by applying 10-mV hyperand depolarizing voltage pulses (duration, 200 ms; pulse interval, 2 s) from a holding potential of Ϫ70 mV using either the perforated patch or standard whole-cell configuration. The currents were recorded using an Axopatch 200B patch clamp amplifier, digitized, and stored in a computer using the Digidata AD converter and the software pClamp (Axon Instruments, Foster City, CA). The ␣-cell membrane potential was recorded using the perforated patch whole-cell configuration. Exocytosis was measured as increases in cell capacitance using, except for Fig.  3, an EPC-9 patch clamp amplifier and the Pulse software (version 8.30; HEKA Elektronik, Lamprecht/Pfalz, Germany). The interval between two successive points was 0.2 s. All measurements of cell capacitance, except those in Fig. 3 in which the perforated patch whole-cell configuration was used, have been performed using the standard whole-cell recording mode. In Fig. 3, changes in cell capacitance were elicited by 500-ms voltage clamp depolarizations to 0 mV from a holding potential of Ϫ70 mV using an EPC-7 patch clamp amplifier (List Elektronik, Darmstadt, Germany) and in-house software written in AxoBasic (Axon Instruments, Foster City, CA) as detailed elsewhere (16). The volume of the recording chamber was 0.4 ml, and the solution entering the bath (1.5-2 ml/min) was maintained at 33°C.
Solutions-The extracellular medium consisted of 138 mM NaCl, 5.6 mM KCl, 2.6 mM CaCl 2 , 1.2 mM MgCl 2 , 5 mM HEPES (pH 7.4 with NaOH), and 0 or 5 mM D-glucose. The extracellular solution used for measurements of cell capacitance evoked by voltage clamp depolarizations contained 118 mM NaCl, 20 mM tetraethylammonium chloride, 5.6 mM KCl, 2.6 mM CaCl 2 , 1.2 mM MgCl 2 , 5 mM HEPES (pH 7.40 with NaOH), and 5 mM glucose. Tetraethylammonium chloride was included in the medium to block the outward delayed rectifying K ϩ current, which otherwise obscures the smaller Ca 2ϩ current (17). The pipette solution used for the infusion experiments consisted of 125 mM potassium glutamate, 10 mM KCl, 10 mM NaCl, 1 mM MgCl 2 , 8 mM CaCl 2 , 3 mM Mg-ATP, 10 mM EGTA, and 5 mM HEPES (pH 7.15 with KOH). The free Ca 2ϩ concentration of the resulting buffer was 0.87 M using the binding constants of Martell and Smith (18). The pipette solution used for measurements of membrane potential and K ATP channel activity, using the perforated patch configuration, was composed of 76 mM K 2 SO 4 , 10 mM KCl, 10 mM NaCl, 1 mM MgCl 2 , 5 mM HEPES (pH 7.35 with KOH). For measurements of exocytosis using voltage clamp depolarizations, K 2 SO 4 was replaced with Cs 2 SO 4 in the pipette solution. Electrical contact was established by adding 0.24 mg/ml amphotericin B to the pipette solution (16). Perforation required a few minutes, and the voltage clamp was considered satisfactory when the G series was stable and Ͼ35-40 nS. The pipette solution used for recording of K ATP channel activity using the standard whole-cell configuration contained 125 mM KCl, 30 mM KOH, 10 mM EGTA, 1 mM MgCl 2 , 5 mM HEPES, 0.3 mM Mg-ATP, and 0.3 mM K-ADP (pH 7.15). Pertussis toxin was obtained from RBI (Natick, MA). Deltamethrin and its inactive analog permethrin were from Alomone Labs (Jerusalem, Israel). All other chemicals were purchased from Sigma.
Glucagon Release-Glucagon release was measured at 37°C in static incubation. Groups of 10 size-matched rat islets were preincubated for 30 min in 200 l of extracellular solution consisting of 138 NaCl, 5.6 mM KCl, 2.6 mM CaCl 2 , 1.2 mM MgCl 2 , 5 mM HEPES (pH 7.4 with NaOH) and 0 -20 mM D-glucose in 96-well Durapore membrane plates (Millipore, Molsheim, France). The medium was aspirated using a vacuum control pump (Millipore) and discarded. The islets were resuspended in 200 l of extracellular solution in the absence and presence of test compounds and the indicated glucose concentration. At the end of the test incubation (1 h), the medium was aspirated and assayed immediately for glucagon using a glucagon radioimmune assay kit (GL-32K; Linco Research, St. Charles, MO).
Data Analysis-In the infusion experiments, the exocytotic rate is presented as the increase in cell capacitance occurring during the first 60 s following establishment of the whole-cell configuration, excluding any rapid changes occurring during the initial ϳ10 s required for equilibration of the pipette solution with cytosol. Results are presented as mean values Ϯ S.E. for the indicated number of experiments. Statistical significance was evaluated using Student's t test for paired or unpaired observations or Dunnett's test for multiple comparisons with a single control. Fig. 1 illustrates electrical activity recorded from a single rat ␣-cell using the perforated patch whole-cell configuration in the absence of glucose. Spontaneous electrical activity was observed in Ͼ80% of the tested cells (n Ͼ 80 cells), as expected for an ␣-cell-rich preparation. The application of phentolamine (0.1 mM) did not affect the ability of the ␣-cells to fire action potentials (Fig. 1A), whereas the subsequent addition of diazoxide (0.1 mM), which activates K ATP channels in rat ␣-cells, was associated with a reversible inhibition of electrical activity (Fig. 1B). Fig. 2A shows measurements of the whole-cell K ATP current from an intact ␣-cell using the perforated patch whole-cell configuration. In the absence of glucose, the 10-mV voltage steps applied from a holding potential of Ϫ70 mV elicited currents with amplitudes of 2 pA, corresponding to an input conductance of 0.2 nS. In a series of five experiments, the input conductance averaged 0.3 Ϯ 0.1 nS. The application of phentolamine or the sulfonylurea tolbutamide (both 0.1 mM) did not affect the current amplitude (phentolamine: 0.4 Ϯ 0.1 nS, n ϭ 5; tolbutamide: 0.4 Ϯ 0.2 nS, n ϭ 5), whereas the K ATP channel opener diazoxide (0.1 mM) produced a 500% increase in the membrane current, and a specific conductance of 1.8 Ϯ 1.1 nS (n ϭ 5) was observed. The lack of effect of tolbutamide suggests that the K ATP channels are already closed in the absence of glucose in the bathing solution. The imidazoline compound RX871024 (0.1 mM) likewise failed to affect electrical activity and K ATP channel activity in intact rat ␣-cells (data not shown).

Effects of Phentolamine on Electrical Activity and K ATP Channels in Rat ␣-Cells-
We used the standard whole-cell patch configuration and intracellular dialysis with 0.3 mM ATP and 0.3 mM ADP to activate the K ATP channels to evaluate whether the K ATP chan-  2B shows that phentolamine (0.1 mM) reduced the whole-cell K ATP current by 60%. The inhibitory effect of phentolamine on the whole-cell K ATP current was prompt and amounted on average to 55 Ϯ 5% (p Ͻ 0.001; n ϭ 5). The subsequent addition of 0.1 mM tolbutamide caused a complete but reversible block of the whole-cell K ϩ conductance (97 Ϯ 1% inhibition; p Ͻ 0.001; n ϭ 5).
Effects of Phentolamine on Depolarization-evoked Exocytosis- Fig. 3A illustrates whole-cell Ca 2ϩ currents and the associated changes in cell capacitance elicited by 500-ms depolarizations from Ϫ70 mV to 0 mV in an intact rat ␣-cell using the perforated patch configuration. In the presence of forskolin, which elevates cytoplasmic cAMP levels, the integrated Ca 2ϩ current amounted to 4.8 picocoulombs, and a capacitance increase of 87 fF was evoked. Two minutes after inclusion of 0.1 mM phentolamine in the bathing solution, the same membrane depolarization produced an integrated Ca 2ϩ current of 4.7 picocoulombs and a capacitance increase of 16 fF (82% inhibition). The depolarizations and increases in cell capacitance were not associated with any changes in cell conductance, and the capacitance measurements are accordingly likely to report exocytosis. On average (Fig. 3B), phentolamine produced 89 Ϯ 16% (p Ͻ 0.01; n ϭ 5) reversible inhibition of exocytosis, which was not associated with a change of the integrated Ca 2ϩ current (Fig. 3C).

Phentolamine Inhibits Exocytosis Evoked by Intracellular
Infusion of Ca 2ϩ -The effects of phentolamine on exocytosis were further investigated in standard whole-cell experiments in which secretion was evoked by intracellular dialysis with a Ca 2ϩ -EGTA buffer with a free Ca 2ϩ concentration of 0.87 M.
Following establishment of the whole-cell configuration, exocytosis was observed as a gradual capacitance increase (Fig. 4A, control). In general, cell capacitance reached a new steadystate level within 3-5 min. It is clear that inclusion of 0.1 mM phentolamine in the pipette solution exerted a strong inhibition of the increase in cell capacitance (Fig. 4A, phentolamine). On average, phentolamine evoked a 83% inhibition of the rate of capacitance increase measured over the first 60 s (excluding the first ϳ10 s) after the establishment of the whole-cell configuration (p Ͻ 0.01; n ϭ 10 (control) and n ϭ 4 (phentolamine)). The effect of phentolamine on exocytosis was dependent on dose (Fig. 4B). No inhibition of exocytosis was observed at Յ3 M. At higher concentrations, phentolamine decreased the rate of capacitance increase by 42-78%. Approximating the average data points of the inhibitory effect of phentolamine on exocytosis to the Hill equation yielded values of the half-maximal inhibitory concentration (IC 50 ) and cooperativity factor of 9 M and 3, respectively. The maximal effects were seen at phentolamine concentrations of Ն100 M (Fig. 4B). Table I shows that the inhibitory action of phentolamine on exocytosis was mimicked by RX871024 and efaroxan. When applied at a concentration of 0.1 mM, exocytosis was decreased by Ͼ70% for both compounds (p Ͻ 0.01; n ϭ 5). The inhibitory effect of imidazoline compounds on exocytosis does not result from ␣ 2 -adrenergic antagonistic activity or binding to I 1 or I 2 receptors, since clonidine (␣ 2 -adrenergic agonist), AGN 192403 (I 1 ligand), or BU-224 (I 2 ligand) failed to affect exocytosis (Table I). Furthermore, an irreversible blockade of either ␣ 2adrenergic receptors with benextramine or I 2 receptors with clorgyline did not affect the ability of phentolamine to inhibit exocytosis (Table I). These data suggest that the inhibitory action of phentolamine on exocytosis does not involve ␣ 2 -adrenergic, I 1 , or I 2 receptors. Table II shows that the inhibitory action of phentolamine on exocytosis was associated with suppression of glucagon release from batches of 10 size-matched rat islets. Phentolamine reduced glucagon release independently of the ambient glucose concentration, with the most pronounced effect at 2.5 mM glucose (54% inhibition). At this glucose concentration, phentolamine reduced glucagon release dose-dependently (Table III) with an IC 50 at 1.2 M, which is in fair agreement with that observed for inhibition of exocytosis.
Phentolamine Evokes G i2 Protein-dependent Inhibition of Exocytosis-We explored whether the ability of phentolamine to inhibit exocytosis involved activation of GTP binding proteins. Fig. 5A shows that inclusion of a 1 mM concentration of the stable GDP analog GDP␤S in the pipette solution abolished the inhibitory effect of phentolamine (0.1 mM) on exocytosis, and the exocytotic response amounted to 90% (n ϭ 5) of the control level. The effect of phentolamine on exocytosis was probably mediated by activation of inhibitory G proteins of the G i /G o type, since pretreatment of the ␣-cells with pertussis toxin (100 ng/ml for Ͼ20 h) abolished the inhibitory action of phentolamine (Fig. 5B).
To determine the type of pertussis toxin-sensitive G protein, we used antisense oligonucleotides against G␣ i1-3 and G␣ o . Fig.  5C shows that ␣-cells pretreated for 24 h with antisense oligonucleotides against G␣ i1 , G␣ i3 , or G␣ o did not affect the ability of phentolamine to inhibit exocytosis. In contrast, phentolamine did not suppress exocytosis in cells pretreated with antisense oligonucleotides against G␣ i2 . In cells treated with sense oligonucleotides against G␣ i2 , phentolamine decreased the rate of capacitance comparable with that observed in control cells (Fig. 5C). This suggests that G i2 proteins mediate the inhibitory action of phentolamine on ␣-cell exocytosis.
Inhibitory Effect of Phentolamine on Exocytosis Involves Activation of Calcineurin-Dephosphorylation catalyzed by the serine/threonine protein phosphatase calcineurin (PP2B) underlies inhibition of exocytosis produced by adrenaline, somatostatin, and ATP in pancreatic ␤-cells (22, 23). As illustrated   in Fig. 6A, this may also apply with regard to phentolamine, since the immunosuppressant cyclosporin A, an inhibitor of calcineurin, abolished the inhibitory action of this imidazoline compound on exocytosis. A similar abolition of phentolamineevoked inhibition of exocytosis was observed with the calcineurin inhibitor deltamethrin (Fig. 6B) but not in the presence of its inactive analogue permethrin (Fig. 6C). On the contrary, okadaic acid (an inhibitor of type 1, 2A, and 3 serine/ threonine protein phosphatases) failed to counteract the inhibitory action of phentolamine (Fig. 6D). On average, phentolamine reduced the exocytotic response by 81% (p Ͻ 0.05; n ϭ 5), similar to that observed in the absence of okadaic acid.
To ascertain that the decrease in exocytosis evoked by phentolamine infusion indeed reflects activation of calcineurin, we measured glucagon release from islets pretreated with inhibitors of this protein phosphatase. Table IV clearly demonstrates that deltamethrin and cyclosporin A prevented the inhibitory action of phentolamine on glucagon release. Under these conditions, glucagon release in the presence of phentolamine amounted to 95% (deltamethrin) and 93% (cyclosporin A) of the control level. On the contrary, phentolamine reduced glucagon release by Ͼ50% (p Ͻ 0.01; n ϭ 5) in islets pretreated with either permethrin or okadaic acid (Table IV). Finally, no inhibition of glucagon release was observed in islets pretreated overnight with pertussis toxin. DISCUSSION Imidazoline compounds have been shown not only to stimulate insulin release but also to improve insulin sensitivity (10,24), which constitutes two main defects underlying glucose intolerance in type 2 diabetic patients. Since patients with type 2 diabetes also exhibit exaggerated glucagon secretion, our present finding that phentolamine inhibits exocytosis of glucagon may constitute the basis for an additional target for the antidiabetogenic action of this class of compounds. The inhibitory action of phentolamine on exocytosis was not associated with a change in the activity of plasma membrane K ATP channels, the activity of voltage-gated Ca 2ϩ channels, or changes in cytoplasmic free Ca 2ϩ levels (data not shown) but results from a direct interference with the exocytotic machinery, an effect mediated by the protein phosphatase calcineurin.
In this study, we extend previous observations in ␤-cells (6,7,25) by showing that phentolamine blocked K ATP channel activity in standard whole-cell patch clamp experiments. This is consistent with the observation that rat ␣-cells are equipped with K ATP channels identical to those expressed in ␤-cells (26,27). The ␣and ␤-cell K ATP channel is a complex of two proteins: a pore-forming subunit, Kir6.2, and the sulfonylurea receptor, SUR1 (26,28,29). It has recently been demonstrated that phentolamine block of K ATP channels is mediated by Kir6.2 and results from a voltage-independent reduction in channel activity (9). Kir6.2 is also expressed in the heart, which may explain why native cardiac and ␤-cell K ATP channels share a similar sensitivity to phentolamine (6,30).
In keeping with previous observations (15,26), we found rat ␣-cells to be spontaneously active in the absence of glucose. Exposure of the ␣-cells to phentolamine in the absence of glucose was not associated with increased electrical activity. The failure of phentolamine to affect electrical activity is consistent with the inability of the imidazoline compound to reduce K ATP channel activity in metabolically intact cells. This suggests that the K ATP channels are already maximally inhibited in the absence of glucose and is consistent with the observation that tolbutamide failed to reduce channel activity under these conditions. Little information is available on how glucose inhibits glucagon secretion in rat ␣-cells, except that inhibition is mediated by glucose metabolism (31).
Our data suggest that dephosphorylation of components regulating exocytosis underlies the inhibitory action of phentolamine on glucagon secretion from intact islets and the perfused

TABLE III
Phentolamine produces dose-dependent inhibition of glucagon release from rat islets Glucagon release was measured from freshly isolated batches of 10 size-matched islets exposed to the indicated phentolamine concentration for 1 h in an extracellular medium with 2.5 mM glucose. rat pancreas (11). This is likely to be mediated by activation of the protein phosphatase calcineurin, since the action of phentolamine was abolished by maneuvers that suppressed the activity of the phosphatase. Calcineurin has been identified in rat pancreatic ␣-cells (32). Our data demonstrate that the ability of phentolamine to inhibit Ca 2ϩ -dependent exocytosis is rapid and readily reversible, suggesting that the magnitude of the secretory response depends principally upon phosphoryla-tion of as yet unidentified exocytotic proteins. Since activation of protein kinase A leads to enhancement of Ca 2ϩ -dependent exocytosis in rat ␣-cells (15), it could be argued that suppression of exocytosis by phentolamine is the result of reduced cAMP levels and inhibition of protein kinase A-mediated exocytosis. However, this possibility seems unlikely, since the ability of phentolamine to suppress exocytosis remained observable in experiments where the cytoplasmic cAMP concentration was elevated using forskolin or by inclusion of the cyclic nucleotide in the pipette solution dialyzing the cells (data not shown). Measurements of adenine nucleotide content in purified rat ␣-cells have revealed that they have a high ATP/ADP ratio already at 1 mM glucose and that it does not change significantly during glucose stimulation (33). This contrasts with the situation in the ␤-cells, where the ATP/ADP ratio increases severalfold following an elevation in the glucose concentration. The constant high ATP/ADP ratio in rat ␣-cells is likely to provide the energy to maintain the cells in a phosphorylated state and consequently to enable phentolamine to inhibit exocytosis in a glucose-independent manner.
Our results show that G i2 proteins mediate the inhibition of exocytosis by phentolamine in rat ␣-cells. This is indeed consistent with the observation that G i2 proteins have been identified in rat pancreatic islets (34). However, it remains to be established whether calcineurin activity is controlled by direct interaction of the G i2 protein or whether intermediate proteins are responsible for signal transduction. Recent studies have revealed that G i and G o proteins are involved in the regulation of intracellular transport processes, adding novel targets to the list of effectors for these versatile molecular switches. Heterotrimeric G i and G o proteins have been found on chromaffin granules and small vesicles from rodent and bovine brain (35,36). Interestingly, these heterotrimeric G-proteins differ in their composition of ␣-subunits. G␣ o1 , G␣ o2 , G␣ i1 , and G␣ i2

TABLE IV
Calcineurin mediates the inhibitory effect of phentolamine on glucagon release from rat islets Glucagon release was measured from batches of 10 size-matched islets in the absence or presence of 0.1 mM phentolamine for 1 h in an extracellular medium with 2.5 mM glucose. Islets were pretreated with deltamethrin and permethrin (20 nM for 1 h), cyclosporin A (1 M for 1 h), okadaic acid (100 nM for 30 min), and pertussis toxin (100 ng/ml for 20 h).  were detected on small synaptic vesicles, whereas chromaffin granules only contain G␣ o2 (36 -38). These G proteins are in an ideal position for controlling transport processes across the granular membrane and the priming and fusion steps regulating exocytosis. Indeed, G o proteins are involved in the exocytotic priming step in chromaffin cells (38), whereas G i3 proteins regulate swelling of zymogen granules, a potentially important prerequisite for granule fusion (39). These considerations raise the interesting possibility that phentolamine inhibits glucagon exocytosis by interfering with granular associated G i2 proteins.