Glucose Evokes Rapid Ca2+ and Cyclic AMP Signals by Activating the Cell-Surface Glucose-Sensing Receptor in Pancreatic β-Cells

Glucose is a primary stimulator of insulin secretion in pancreatic β-cells. High concentration of glucose has been thought to exert its action solely through its metabolism. In this regard, we have recently reported that glucose also activates a cell-surface glucose-sensing receptor and facilitates its own metabolism. In the present study, we investigated whether glucose activates the glucose-sensing receptor and elicits receptor-mediated rapid actions. In MIN6 cells and isolated mouse β-cells, glucose induced triphasic changes in cytoplasmic Ca2+ concentration ([Ca2+]c); glucose evoked an immediate elevation of [Ca2+]c, which was followed by a decrease in [Ca2+]c, and after a certain lag period it induced large oscillatory elevations of [Ca2+]c. Initial rapid peak and subsequent reduction of [Ca2+]c were independent of glucose metabolism and reproduced by a nonmetabolizable glucose analogue. These signals were also blocked by an inhibitor of T1R3, a subunit of the glucose-sensing receptor, and by deletion of the T1R3 gene. Besides Ca2+, glucose also induced an immediate and sustained elevation of intracellular cAMP ([cAMP]c). The elevation of [cAMP]c was blocked by transduction of the dominant-negative Gs, and deletion of the T1R3 gene. These results indicate that glucose induces rapid changes in [Ca2+]c and [cAMP]c by activating the cell-surface glucose-sensing receptor. Hence, glucose generates rapid intracellular signals by activating the cell-surface receptor.


Introduction
Secretion of insulin is regulated by nutrients, neurotransmitters and hormones in pancreatic βcells [1]. Among them, glucose is a primary stimulator of insulin secretion and is able to induce secretion by itself. Thus, when ambient glucose concentration rises, insulin secretion is initiated after a certain lag period [1]. The mechanism by which glucose stimulates insulin secretion has been investigated extensively for several decades [1,2]. It was shown some decades ago that glucose induces complex changes in ion fluxes and membrane potential [3][4][5][6]. The resting membrane potential of mouse β-cells is between -60 and -70 mM [3][4][5], which is determined mainly by high permeability of K + . Elevation of ambient glucose leads to a gradual depolarization of 10 to 15 mV, which is followed by an initiation of action potentials. Initial depolarization induced by glucose is brought about by a decrease in K + permeability of the plasma membrane. It is now known that glucose enters the cells, is metabolized through the glycolytic pathway and in mitochondria, and the resultant increase in ATP/ADP ratio causes closure of the ATP-sensitive K + channel (K ATP channel) [2,[5][6][7]. Closure of the K ATP channel leads to gradual depolarization to a threshold, at which action potential driven by Ca 2+ is initiated [4,5,7,8]. Since it takes a minute or more for glucose to be metabolized, action potential starts after one to several minutes of lag time [7][8]. After the initial burst of action potential, the membrane potential returns to the level slightly below the resting potential, which is followed by cyclic changes in the membrane potential [4][5][6].
When changes in cytoplasmic Ca 2+ concentration ([Ca 2+ ] c ) are monitored in pancreatic βcells, the addition of a high concentration of glucose reduces [Ca 2+ ] c rather rapidly [9][10][11]. This initial decrease in [Ca 2+ ] c lasts for a few minutes and is followed by an oscillatory elevation of [Ca 2+ ] c [9][10][11]. The initial decrease in [Ca 2+ ] c is thought to be due to sequestration of Ca 2+ mainly to endoplasmic reticulum (ER) via the ER Ca 2+ pump (SERCA) [12,13]. In fact, initial decrease in [Ca 2+ ] c is accompanied by an increase in Ca 2+ concentration in ER [14,15]. The role of this sequestration of Ca 2+ to ER is not totally certain but it may be important for subsequent loading of Ca 2+ into mitochondria. More importantly, the exact mechanism by which glucose stimulates sequestration of calcium into ER is not certain at present.
Besides changes in Ca 2+ , glucose also increases cyclic 3', 5' AMP (cAMP) in pancreatic βcells [16][17][18]. Elevation of cytoplasmic cAMP concentration ([cAMP] c ) induced by a high concentration of glucose has been thought to be secondary to elevation of [Ca 2+ ] c [18,19]. In fact, pancreatic β-cells express adenylate cyclase (AC) isoforms, ACIII and ACVIII [20,21]. ACVIII is a Ca 2+ -calmodulin-activated AC and is also regulated by G s . Presumably, elevation of [Ca 2+ ] c activates calcium-dependent AC such as ACVIII, and increases production of cyclic AMP [19]. However, in a study using islets obtained from transgenic mice expressing a cAMP sensor Epac1-camps, Kim et al. [22] showed that glucose evoked a rapid elevation of [cAMP] c , which preceded elevation of [Ca 2+ ] c . This observation raises a possibility that increase in [cAMP] c is rapid and at least partly independent of elevation of [Ca 2+ ] c .
We have shown recently that subunits of the sweet taste receptor [23] are expressed in pancreatic β-cells [24]. Specifically, T1R3 subunit is abundantly expressed in β-cells while the protein expression of T1R2 is negligible [25]. Furthermore, the actions of sweet molecules are blocked by knockdown of T1R3 whereas knockdown of T1R2 was without effect. Based on these observations, we have speculated that a homodimer of the T1R3 functions as a cell-surface glucose-sensing receptor. Alternately, a heterodimer of T1R3 and another class C G protein-coupled receptor (GPCR) may function as a glucose-sensing receptor [26]. This receptor is activated by glucose and a nonmetabolizable analogue 3-O-methylglucose [26,27]. Interestingly, activation of the glucose-sensing receptor by glucose facilitates its own metabolism in pancreatic β-cells [27].

Animals
The animal experiment was approved by the Animal Experiment and Ethics Committee, Gunma University School of Medicine (#25-0112), and was conducted according to the guidelines for animal care issued by the Committee. B6; 129-Tas1r3 <tm1Csz>/J mice were purchased from the Jackson Laboratory (Bar Harbor, ME). They were kept in an experimental animal facility controlled at 23°C room temperature with a 12-hr light and dark cycle, and with free access to standard chow and water. The animal experiment was conducted according to the guidelines for animal care issued by Animal Experiment and Ethics Committee, Gunma University School of Medicine.

Measurement of cAMP and Activation of PKC
MIN6 cells were transiently transfected with plasmid encoding the cAMP indicator Epac1camps [30] kindly provided by Dr. Lohse of the University of Würzburg (Germany) or the myristoylated alanine-rich C kinase substrate (MARCKS) fused with a green fluorescent protein (GFP), MARCKS-GFP [31]. The transfected cells were used to measured cAMP levels and PKC activation as previously described [24]. Knockdown of T1R3 was performed as described previously [29].

Imaging of Cytoplasmic Ca 2+
Measurement of [Ca 2+ ] c was performed by using a fluorescent Ca 2+ indicator Fluo-8 (AAT Bioquest, Sunnyvale, CA). MIN6 cells and isolated β-cells were loaded with 2 μM Fluo-8/ AM dissolved in HBSS for 20 min at room temperature. To monitor [Ca 2+ ] c using Cameleon-nano15, MIN6 cells were transfected with PM-Cameleon-nano15 [32] by electroporation as described previously [27]. The cells were placed on a 35 mm glass bottom culture dish. The cells were visualized with an Olympus UPlanAPO 10x Water Objective lens (Olympus, Tokyo, Japan). To detect fluorescence images, we used AQUACOSMOS/ASHURA, 3CCD based fluorescence energy transfer imaging system (Hamamatsu Photonics, Hamamatsu, Japan). Fluo-8 fluorescence was obtained by a U-MGFPHQ cube (Olympus), and expressed as the ratio of cytosolic fluorescence and initial intensity (F/F 0 ). PM-Cameleon-nano15 fluorescence was obtained by a 440AF21 excitation filter (Omega Optical, Brattleboro, VT) and DM455DRLP dichroic mirror (Omega optical) and expressed as the ratio of CFP/YFP. These Images were captured with at a C7780-22 ORCA3CCD camera (Hamamatsu Photonics) at 10-second intervals.

Statistical Analysis
Values are expressed as mean ± SE. Statistical analysis was done by using Mann-Whitney's Utest. A p value of less than 0.05 was considered statistically significant.

Effect of Glucose on Subplasmalemmal Free Calcium Concentration
We first examined whether glucose evoked a rapid effect on cytoplasmic Ca 2+ . To this end, we monitored changes in [Ca 2+ ] c in MIN6 cells using an ultrasensitive Ca 2+ indicator, yellow Cameleon-nano15 [34] targeted to the plasma membrane (PM-Cameleon-nano15) [32]. PM-Cameleon-nano15 enabled us to monitor subtle changes in [Ca 2+ ] c in these cells [32,34]. When ambient glucose concentration was raised from 2.8 mM to 25 mM, an immediate transient elevation of [Ca 2+ ] c was observed. A typical response of [Ca 2+ ] c is shown in Fig 1A. This immediate elevation of [Ca 2+ ] c was observed within seconds and [Ca 2+ ] c peaked within 10 sec.
[Ca 2+ ] c then rapidly decreased to a plateau level, which was lower than the basal level. [Ca 2+ ] c remained decreased for 1 to 5 min and then [Ca 2+ ] c rose abruptly. The initial rapid peak of [Ca 2+ ] c was observed in approximately 60% of the cells examined while the following plateau phase and the second large elevation of [Ca 2+ ] c were observed in almost all of the cells tested. In cells without the rapid peak of [Ca 2+ ] c , a small hump of [Ca 2+ ] c was instead observed immediately after the addition of glucose ( Fig 1B). [Ca 2+ ] c was then reduced to a plateau level, which was lower than the basal level, and remained reduced for some minutes. Then [Ca 2+ ] c rose abruptly and oscillatory elevation of [Ca 2+ ] c was observed. Note that in cells without a rapid peak of [Ca 2+ ] c , the area under the curve (AUC) of the subsequent oscillatory elevation of [Ca 2+ ] c was smaller than that in cells with an initial rapid peak of [Ca 2+ ] c . Thus AUC was significantly smaller than that in cells with a rapid peak of [Ca 2+ ] c ( Fig 1C). The initial rapid peak of Ca 2+ or small hump of [Ca 2+ ] c detected using PM-Cameleon-nano15 was never observed by monitoring [Ca 2+ ] c using a conventional Ca 2+ indicator fura-2. Thus, when [Ca 2+ ] c was monitored by fura-2, [Ca 2+ ] c was decreased by the addition of 25 mM glucose, and after some minutes of interval, elevation of [Ca 2+ ] c was observed (data not shown). More importantly, when cells expressing PM-Cameleon-nano15 were simultaneously loaded with fura-2, the initial rapid elevation of [Ca 2+ ] c monitored by PM-Cameleon-nano15 was abolished ( Fig 1D). In other words, loading of fura-2 abolished the initial rapid peak of [Ca 2+ ] c . Triphasic changes in [Ca 2+ ] c were observed when cells were stimulated with 8.3 mM and 16.7 mM glucose ( Fig 1E  and 1F) and approximately 50% of the cells presented a rapid peak of [Ca 2+ ] c . Both the first rapid peak and AUC of the second oscillatory elevation of [Ca 2+ ] c were increased in dosedependent manners (Fig 1G and 1H). The first rapid peak of [Ca 2+ ] c was markedly inhibited by the addition of a non-specific inhibitor of phospholipase C (PLC) U73122 [35] (Fig 2A and  2B). Subsequent reduction of [Ca 2+ ] c was also abolished by U73122 ( Fig 2B). Note that an inactive analogue U73343 was without effect (data not shown). Likewise, the first rapid peak of [Ca 2+ ] c and subsequent reduction of [Ca 2+ ] c were markedly inhibited by the addition of a G q inhibitor YM254890 [36] (Fig 2C). Quantitatively, the first peak of [Ca 2+ ] c was significantly reduced by U73122 and YM254890 (Fig 2D). Similarly, the reduction of [Ca 2+ ] c was abolished by both U73122 and YM254890 ( Fig 2E). It should be noted that YM254890 also significantly inhibited large oscillatory elevation of [Ca 2+ ] c induced by glucose ( Fig 2F).
We tried to detect the rapid peak of [Ca 2+ ] c by using conventional Ca 2+ indicators and found that the initial rapid elevation of Ca 2+ was detected by using fluo-8 in a fraction of the cells. As shown in Fig 3A, a rapid elevation of [Ca 2+ ] c induced by 25 mM glucose was observed in approximately 40% of the fluo-8-loaded cells. In these cells, triphasic changes in [Ca 2+ ] c were qualitatively similar to those observed by using PM-Cameleon-nano15. When extracellular Ca 2+ was removed, glucose-mediated initial peak of [Ca 2+ ] c was observed but the large elevation of [Ca 2+ ] c observed several minutes later was markedly inhibited (Fig 3B). Similar results were obtained when cells were stimulated by 25 mM glucose in the presence of 1 μM nifedipine, an inhibitor of L-type voltage-gated calcium channel ( Fig 3C). Note that the rapid peak of [Ca 2+ ] c was smaller compared to Fig 3A. We then assessed the involvement of the glucose-sensing receptor in glucose-induced changes in [Ca 2+ ] c . We first blocked glucose metabolism by adding mannoheptulose, an inhibitor of glucokinase. In the presence of 10 μM mannoheptulose, glucose induced a rapid elevation of [Ca 2+ ] c and subsequent sustained reduction of [Ca 2+ ] c . The magnitude of the first peak of [Ca 2+ ] c was smaller compared to that in the absence of mannoheptulose. A large elevation of [Ca 2+ ] c observed after a lag period of some minutes was blocked (Fig 3D). Similar results were obtained by adding a nonmetabolizable glucose analogue, 3-O-methylglucose, instead of glucose. Thus, 3-O-methylglucose induced a rapid elevation of [Ca 2+ ] c and subsequent reduction of [Ca 2+ ] c . However, a large elevation of [Ca 2+ ] c after a lag period of some minutes was not observed (Fig 3E). When the glucose-sensing receptor was blocked by adding lactisole, an inhibitor of T1R3 [37], glucose did not induce a rapid elevation of [Ca 2+ ] c (Fig 3F). Subsequent reduction of [Ca 2+ ] c was not observed. In contrast, elevation of [Ca 2+ ] c was observed several minutes later. Quantitatively, the decrease in [Ca 2+ ] c was abolished by lactisole (Fig 3G).

Effect of Glucose on cAMP
We next monitored changes in [cAMP] c induced by glucose using a cAMP indicator Epac1camps [30]. When Epac1-camps-expressing MIN6 cells were stimulated by 25 mM glucose, a rapid and sustained elevation of [cAMP] c was observed (Fig 4A). Elevation of [cAMP] c was observed within 10 sec of the addition of glucose, and [cAMP] c reached a plateau level approximately 1 min after the addition of glucose. When changes in [cAMP] c and [Ca 2+ ] c were monitored simultaneously in Epac1-camps-expressing cells loaded with fura-2, glucose induced reduction of [Ca 2+ ] c and the elevation of [cAMP] c were observed even in the same period (Fig 4B), indicating that [cAMP] c was increased when [Ca 2+ ] c was reduced. Glucose-induced elevation of [cAMP] c was observed at a concentration of 8.3 mM and was increased in a dosedependent manner (Fig 4C). Elevation of [cAMP] c was observed even in the absence of extracellular Ca 2+ (Fig 4D). AUC of the [cAMP] c response in the absence of extracellular calcium was not changed significantly compared to that in the presence of extracellular calcium ( Fig  4E). On the other hand, the elevation of [cAMP] c was abolished by introducing dominant-negative G s (Fig 4F and 4G). Note that dominant-negative G s did not affect the elevation of [cAMP] c induced by depolarizing concentration of KCl (data not shown). This rapid elevation of [cAMP] c was observed in the presence of mannoheptulose (Fig 4H). Furthermore, 3-Omethylglucose, a nonmetabolizable analogue of glucose, essentially reproduced the effect of glucose (Fig 4I). When T1R3 was knocked down, glucose-induced elevation of [cAMP] c was markedly inhibited (Fig 4J).

Effect of Glucose on Activation of Protein Kinase C
We then measured changes in DAG. As shown in Fig 5A, the addition of 25 mM glucose induced a rapid and monophasic increase in DAG. Elevation of DAG was detected within 10 sec and the peak was observed around 3 min of stimulation by glucose. We then monitored activation of protein kinase C by monitoring phosphorylation state of MARCKS in cytosol, a substrate for protein kinase C that translocates from the plasma membrane to cytosol upon phosphorylation [31]. As shown in Fig 5B, the addition of 25 mM glucose induced biphasic elevation of phosphorylated MARCKS in cytosol. An initial increase was observed within 20 sec and peaked at 2 to 3 min after the addition of glucose. The second larger peak of phosphorylated MARCKS in cytosol was observed 5 to 7 min after the addition of glucose. When mannoheptulose was included to block metabolism of glucose, the first peak was observed whereas the second peak was abolished (Fig 5C). Addition of 3-O-methylglucose, a nonmetabolizable glucose analogue, evoked the first peak whereas the second peak was not observed and the level of phosphorylated MARCKS in cytosol was decreased gradually (Fig 5D). Results of the quantitative analyses of the effects of mannoheptulose and 3-O-methylglucose are shown in Fig 5E  and 5F.  (Fig 6A). In cells without the rapid peak of [Ca 2+ ] c , glucose induced a gradual decrease in [Ca 2+ ] c , which was followed by an abrupt elevation of [Ca 2+ ] c (Fig 6B).
[Ca 2+ ] c remained elevated for more than 10 min. In T1R3-null β-cells, the glucose-induced rapid peak of [Ca 2+ ] c was never observed. Moreover, a high concentration of glucose did not cause a decrease in [Ca 2+ ] c . After a lag period of approximately 7 min, an increase in [Ca 2+ ] c was observed (Fig 6C). Compared to that in normal β-cells, elevation of [Ca 2+ ] c was blunted. Quantitatively, initial decrease in [Ca 2+ ] c induced by glucose was abolished in T1R3-null βcells (Fig 6D). The onset of the large elevation of [Ca 2+ ] c was significantly delayed in T1R3-null β-cells (Fig 6E). In addition, AUC of the elevation of [Ca 2+ ] c induced by high concentration of glucose was markedly reduced in T1R3-null β-cells (Fig 6F). We then examined the effect of 3-O-methylglucose on [Ca 2+ ] c in normal β-cells. As shown in Fig 7A, addition of 3-O-methylglucose induced a rapid elevation of [Ca 2+ ] c , which was followed by a sustained decrease in [Ca 2+ ] c in normal β-cells. The initial rapid peak of [Ca 2+ ] c was observed in approximately 40% of the cells. In the rest of cells, the initial rapid peak was not observed and [Ca 2+ ] c was decreased in response to 3-O-methylglucose ( Fig 7B). In contrast, 3-O-methylglucose did not induce a rapid peak of [Ca 2+ ] c nor a sustained decrease in [Ca 2+ ] c in T1R3-null β-cells ( Fig  7C). Quantitatively, sustained decrease in [Ca 2+ ] c induced by 3-O-methylglucose was abolished in T1R3-null β-cells (Fig 7D).
We then measured changes in [cAMP] c induced by glucose in normal β-cells. As shown in Fig 7E, glucose evoked a rapid and sustained elevation of [cAMP] c . Elevation of [cAMP] c was observed within 10 sec. In contrast, glucose did not increase [cAMP] c in T1R3-null β-cells (Fig 7F). Quantitatively, elevation of [cAMP] c induced by glucose was markedly reduced in T1R3-null β-cells (Fig 7G).

Discussion
The present study was conducted to detect receptor-mediated rapid signals evoked by glucose in pancreatic β-cells. Using an ultrasensitive Ca 2+ indicator Cameleon-nano15 [34] targeted to the plasma membrane [32], we could detect subtle changes in [Ca 2+ ] c in β-cells. As shown in ] c has never been reported to date. This is because the elevation of [Ca 2+ ] c was rapid and subtle, and was only detected by using a very sensitive indicator PM-Cameleon-nano15 but not by conventional Ca 2+ indicators such as fura-2. As demonstrated in Fig 1D, loading of fura-2 abolished this rapid Ca 2+ signal detected by PM-Cameleon-nano15. The chelating action of fura-2 may have buffered the tiny changes in [Ca 2+ ] c . In any case, a sensitive method enabled us to detect the immediate Ca 2+ signal induced by high concentration of glucose. It should be noted that, when we monitored [Ca 2+ ] c by using a Ca 2+ indicator with stronger fluorescence fluo-8, we could detect the rapid peak of [Ca 2+ ] c in a fraction of the cells (Fig 1E). This elevation is due at least partly to release of Ca 2+ from an intracellular pool(s) since the rapid elevation of [Ca 2+ ] c was observed in the absence of extracellular Ca 2+ . Since the rapid elevation of [Ca 2+ ] c was inhibited by an inhibitor of PLC and an inhibitor of G q , it may be due to a release of calcium from ER caused by inositol trisphosphate. This rapid elevation of [Ca 2+ ] c was observed even if glucose metabolism was blocked by mannoheptulose, indicating that it is not dependent on glucose metabolism. Instead, the rapid elevation was blocked by lactisole, an inhibitor of T1R3 [37]. Furthermore, the rapid peak of [Ca 2+ ] c was never observed in T1R3-null β-cells. The rapid peak of [Ca 2+ ] c evoked by glucose is therefore caused by activation of the glucose-sensing receptor. In agreement with this notion, activation of the receptor by an administration of nonmetabolizable glucose analogue 3-O-methylglucose reproduced the rapid elevation of [Ca 2+ ] c . Collectively, the rapid elevation of [Ca 2+ ] c is a signal due to activation of the glucose-sensing receptor. It should be mentioned that the rapid elevation of [Ca 2+ ] c induced by glucose is different from the elevation of [Ca 2+ ] c induced by typical calcium-mobilizing agonists, for example, carbachol [24] in many respect. Although both glucose and carbachol may activate PLC and elevate [Ca 2+ ] c , [Ca 2+ ] c response induced by glucose is subtle and is not detected by fura-2 whereas that induced by carbachol is easily detectable by fura-2 [24]. Also, glucose-induced rapid elevation of [Ca 2+ ] c was only transient and [Ca 2+ ] c decayes quickly. In addition, glucose-induced rapid elevation of [Ca 2+ ] c is followed by a sustained decrease in [Ca 2+ ] c while that induced by carbachol is not. Presumably, glucose and carbachol activate PLC differently and elicit slightly different [Ca 2+ ] c signals.
The present results indicate that there are two populations of β-cells with regard to the [Ca 2+ ] c response to glucose: roughly half of the cells presented a rapid peak of [Ca 2+ ] c induced by glucose whereas rest of them did not. Difference in the expression levels of T1R3 would explain the different patterns of [Ca 2+ ] c response in two populations. However, when we measured the expression of T1R3 in pancreatic islets [25] and in MIN6 cells [24], the expression of T1R3 was not so different among β-cells. Presumably, difference in the expression of other signaling molecules would be responsible for different properties found in two populations.
Previous studies have shown that glucose induced reduction of [Ca 2+ ] c , which is due largely to sequestration of Ca 2+ into ER through activation of SERCA [9][10][11][12][13]. Nonetheless, the mechanism by which glucose promotes sequestration of Ca 2+ into ER is not certain [38]. Indeed, inhibition of the glucose-sensing receptor by lactisole [37] abolished the reduction of [Ca 2+ ] c induced by glucose (Fig 2C).  [39]. The reduction of [Ca 2+ ] c is due to activation of the plasma membrane Ca 2+ pump (PMCA) [39]. Although the precise mechanism by which the receptor activates PMCA is still unclear, this is another example showing that binding of a ligand to the glucose-sensing receptor causes activation of the Ca 2+ pump leading to sustained reduction of [Ca 2+ ] c . In β-cells, activation of the receptor by glucose may cause activation of SERCA and reduces [Ca 2+ ] c . Further study is obviously needed to identify the mechanism of activation of SERCA by the glucose-sensing receptor. In this regard, it is possible that increase in subplasma membrain Ca 2+ activates SERCA and promotes uptake of Ca 2+ to ER. Alternately, since elevation of cAMP stimulates sequestration of Ca 2+ to ER [40], it is possible that cAMP produced by the receptor activation of G s would cause sequestration of Ca 2+ to ER. Further studies are needed to identify the mechanism. It should be mentioned that the reduction of [Ca 2+ ] c induced by glucose is also dependent on glucose metabolism to some extent. This is because SERCA is an ATP-requiring enzyme and glucose elevates intracellular ATP within a minute [27]. Elevated ATP would facilitate uptake of Ca 2+ to ER.
In the present study, we monitored the glucose effect on [cAMP] c by paying particular attention to a rapid action of glucose. The results obtained in MIN6 cells and mouse β-cells show that glucose induces an immediate elevation of [cAMP] c . Glucose-induced elevation of [cAMP] c preceded a large oscillatory elevation of [Ca 2+ ] c and was observed even in the absence of extracellular calcium. These results suggest that glucose-mediated elevation of [cAMP] c is at least partly independent of elevation of [Ca 2+ ] c . We have previously shown that the glucosesensing receptor causes elevation of [cAMP] c by a G s -dependent mechanism [29]. In addition,  (Fig 4I). Our results are in agreement with the report by Kim et al. [22] and Fridlyand et al. [41]. However, our results are not consistent with previous reports showing that elevation of [cAMP] c is secondary to the elevation of [Ca 2+ ] c [18]. The reason for the discrepancy is not totally certain. However, our measurement was focused on the early response of [cAMP] c in β-cells. It is likely that at least rapid elevation of [cAMP] c is caused by receptor-mediated activation of G s .
In the present study, we also showed that glucose induced rapid and long-lasting activation of protein kinase C by showing that glucose induces biphasic increase in phosphorylated MARCKS in cytosol. It has been generally thought that glucose activates protein kinase by elevating [Ca 2+ ] c [42]. In β-cells, elevation of [Ca 2+ ] c would activate calcium-dependent PLC leading to generation of DAG [43]. Alternately, high concentration of glucose stimulates de novo synthesis of DAG by supplying a substrate [44]. In any event, it has been thought that activation of protein kinase C is dependent on glucose metabolism and is rather slow in onset. As shown in Fig 3B, however, phosphorylation of MARCKS was detected within 10 seconds and increase in DAG was also detected within 10 sec. Furthermore, at least the first phase of elevation of phosphorylated MARCKS was independent of glucose metabolism and was reproduced by a nonmetabolizable glucose analogue. Our previous results show that the glucose-sensing receptor is coupled to the PLC-calcium messenger system [24]. Collectively, a rapid activation of protein kinase C may be independent of glucose metabolism and is caused by receptormediated activation of PLC.
The present study demonstrates for the first time that glucose evokes rapid intracellular signals, Ca 2+ and cAMP, in pancreatic β-cells. These actions are independent of the glucose metabolism and are mediated by the glucose-sensing receptor. Importantly, inhibition of the receptor or deletion of the T1R3 gene attenuates glucose-induced insulin secretion [37,45]. Specifically, insulin secretion is markedly delayed and blunted in β-cells obtained from T1R3 knockout mice [45]. Hence, these receptor-mediated rapid signals are critical for the glucose action in β-cells. Collectively, the action of glucose is not solely dependent on its metabolism [46]. The glucose-sensing receptor generates rapid signals and, by priming the metabolic pathway, enhances the pathway dependent on the glucose metabolism [26]. Our results are different from the postulate by Kyriazis et al. [47] that the sweet taste receptor inhibits basal secretion of insulin. As we mentioned previously [37], their conclusion is solely dependent on the results obtained by using a T1R3 inhibitor lactisole. Moreover, their data may result from an inadequate use of lactisole [37]. In fact, their results obtained by lactisole are even contradictory to their own results [47,48]. More importantly, glucose-induced insulin secretion is obviously attenuated in T1R3-null β-cells both in vitro [37,45] and in vivo [49], it is certain that T1R3 is involved in the action of glucose in pancreatic β-cells.
In summary, we identified the intracellular signals evoked by glucose via the activation of the glucose-sensing receptor. These rapid signals may be important for the priming of the metabolic pathway [26,27] and for rapid secretion of insulin [37,45].