Muscarinic Control of MIN6 Pancreatic β Cells Is Enhanced by Impaired Amino Acid Signaling*

Background: Depletion of the GPCR T1R1/T1R3 increased calcium and ERK1/2 signaling by carbachol. Results: T1R3 depletion or reducing amino acids overnight increased M3 muscarinic receptor expression and altered calcium responses. Conclusion: M3 receptor expression in β cells is up-regulated by reduced amino acid availability. Significance: The M3 muscarinic receptor is a potential therapeutic target in β cells with impaired amino acid sensitivity. We have shown recently that the class C G protein-coupled receptor T1R1/T1R3 taste receptor complex is an early amino acid sensor in MIN6 pancreatic β cells. Amino acids are unable to activate ERK1/2 in β cells in which T1R3 has been depleted. The muscarinic receptor agonist carbachol activated ERK1/2 better in T1R3-depleted cells than in control cells. Ligands that activate certain G protein-coupled receptors in pancreatic β cells potentiate glucose-stimulated insulin secretion. Among these is the M3 muscarinic acetylcholine receptor, the major muscarinic receptor in β cells. We found that expression of M3 receptors increased in T1R3-depleted MIN6 cells and that calcium responses were altered. To determine whether these changes were related to impaired amino acid signaling, we compared responses in cells exposed to reduced amino acid concentrations. M3 receptor expression was increased, and some, but not all, changes in calcium signaling were mimicked. These findings suggest that M3 acetylcholine receptors are increased in β cells as a mechanism to compensate for amino acid deficiency.

The essential function of the pancreatic ␤ cell is to secrete insulin in response to increases in circulating glucose. Other nutrients, hormones, and paracrine agents influence pancreatic ␤ cell functions and insulin secretion to optimize glucose homeostasis. Ligands for several G protein-coupled receptors (GPCRs) 3 are among the most significant in tuning insulin secretion from ␤ cells, which express several different classes of GPCRs, including muscarinic acetylcholine receptors (mAChRs) (1).
Five mAChRs subtypes, M1-M5, have been identified (2). The M1, M3, and M5 subtypes are G q -coupled receptors, whereas M2 and M4 are G i -coupled receptors that are inhibited by pertussis toxin (3). ␤ cell muscarinic receptors are G q -coupled because binding of acetylcholine to these receptors results in the well characterized action of G q to activate phospholipase C ␤. The resulting hydrolysis of phosphatidylinositol 4,5-bisphosphate generates the second messenger inositol 1,4,5triphosphate, which binds to its receptor on the ER and induces calcium release from intracellular stores (4). Muscarinic agonist-induced mobilization of intracellular calcium (Ca 2ϩ i ) was absent in mice selectively lacking two members of the ␤ cell G q protein family, G␣ q and G␣ 11 (5).
The M3 mAChR (M3R) is the predominant receptor subtype expressed in ␤ cells and insulin-secreting cell lines (6,7). Parasympathetic nerve endings that innervate the pancreas release acetylcholine during the preabsorptive and absorptive phases of feeding (8) to activate this receptor. Studies utilizing M3R knockout mice have implicated M3R as the receptor subtype responsible for cholinergic potentiation of glucose-stimulated insulin secretion (9,10). Furthermore, mice selectively deficient in ␤ cell M3Rs demonstrated impaired glucose tolerance and reduced insulin release, whereas mice overexpressing M3Rs in ␤ cells exhibited a significant increase in glucose tolerance and insulin release (11). Similar observations were made in mice overexpressing constitutively active ␤ cell M3Rs (1).
Signaling by the M3R also activates ERK1/2 in ␤ cells, most likely downstream of elevated intracellular calcium (12)(13)(14). ERK1/2 activation enhances insulin gene transcription following nutrient-induced insulin secretion (15)(16)(17)(18). We have reported previously that the GPCR complex T1R1/T1R3 is an early amino acid sensor in the MIN6 pancreatic ␤ cell line and in other cell types (12). Similar to M3R, T1R1/T1R3 activation leads to a rise in Ca 2ϩ i and ERK1/2 phosphorylation that is partially dependent upon phospholipase C ␤ activation. Reduced expression of T1R3 in MIN6 cells resulted in a decrease of amino acid-induced ERK1/2 and mammalian target of rapamycin complex 1 activation. Signaling defects in cells in which the receptor had been depleted included a reduction in the ability of amino acids to induce changes in Ca 2ϩ i (12). Despite the impaired ability of amino acids to stimulate ERK1/2 in T1R3-depleted MIN6 cells, carbachol, a muscarinic receptor agonist, activated ERK1/2 better in T1R3-depleted cells than in control cells (12). We explored the underlying mechanisms for the enhanced carbachol response in MIN6 cells to determine whether similar mechanisms were enlisted to compensate for amino acid deficiency.

EXPERIMENTAL PROCEDURES
Materials-Fura-2/AM was purchased from Molecular Probes. Nifedipine was purchased from Calbiochem. 2-Aminoethoxydiphenyl borate (2-APB) was purchased from Sigma. Thapsigargin was purchased from Santa Cruz Biotechnology.
Cell Culture-MIN6 cells were cultured, and stable cell lines with T1R3 expression reduced following expression of a short hairpin were created and maintained as described previously (12).
Calcium Assays-Cells were plated at 80% confluency in white-walled, 96-well plates (Costar 3903). After 48 h, the cells were washed twice with PBS (0.2 ml/well) and incubated with 5 M Fura-2/AM diluted in Krebs-Ringer bicarbonate solution (KRBH) containing 115 mM NaCl, 5 mM KCl, 24 mM NaHCO 3 , 1 mM MgCl 2 , 2.5 mM CaCl 2 , 25 mM HEPES (pH 7.4), 0.1% BSA, and 4.5 mM glucose for 1 h (0.1 ml/well). Cells were then washed twice with KRBH (0.2 ml/well) and equilibrated in the same buffer for 30 min (0.1 ml/well). Agents were applied (0.1 ml/well) to triplicate wells at 2ϫ concentrations using injectors at a rate of 225 l/s. Changes in Ca 2ϩ i were assessed every 0.74 s by dual excitation of Fura-2 at 340/11 and 380/20 nm (center/ bandpass) and emission at 508/20 nm using the Synergy TM 2 multimode microplate reader (BioTek) with Gen5 TM software. Cells were pretreated with the indicated inhibitors for 30 min prior to stimulation. For experiments performed in the absence of calcium, cells were loaded, washed, and equilibrated with calcium-free KRBH in which MgCl 2 was substituted for 2.5 mM CaCl 2 . To assess store-operated calcium entry (SOCE), intracellular stores were depleted using 10 M thapsigargin. Calcium was then replenished with a second injection of KRBH containing 12.5 mM CaCl 2 (5ϫ concentration, 50 l/well). To assess receptor-operated calcium entry (ROCE), after calcium repletion, a third injection was required to apply 0.6 mM carbachol (6ϫ concentration, 50 l/well). Final concentrations of all agents were 1ϫ. For experiments involving nifedipine or 2-APB, cells were pretreated with inhibitors for 30 min prior to stimulation. All steps in each assay were performed at room temperature.
Nutrient Deprivation-MIN6 cells were plated as above for calcium assays or in 12-well plates for RNA or protein isolation. To examine the effects of reduced amino acids, cells nearing confluency were washed twice with PBS and incubated with KRBH supplemented with 10% dialyzed serum, 4.5 mM glucose, and either 1.0ϫ amino acids (12) or 0.1ϫ amino acids for 16 h at 37°C and 10% CO 2 prior to stimulation with carbachol or cell lysis. Calcium was measured as above with reduced amino acids throughout. To examine the effects of reduced glucose, cells were incubated as above in KRBH containing 10% dialyzed serum, 1ϫ amino acids, and either 25 or 2 mM glucose. Human islets were provided by the Integrated Islet Distribution Program. Islets were washed twice in KRBH and then once in KRBH containing 10% dialyzed serum, 4.5 mM glucose, and either 0.1ϫ or 1ϫ amino acids prior to treatment overnight.
Statistical Analysis-Results were expressed as means Ϯ S.E. determined from three independent experiments. Statistical significance was calculated using Student's t test.

Carbachol-induced Changes in Ca 2ϩ
i and M3R Expression Are Enhanced in MIN6 Cells after Depletion of T1R3-To determine the basis for the increased carbachol-induced ERK1/2 phosphorylation in MIN6 cells with reduced T1R3 expression (12), we first examined the effect of carbachol on Ca 2ϩ i after loading cells with the ratiometric calcium indicator Fura-2. Stimulation of ␤ cells with carbachol produced a biphasic rise in Ca 2ϩ i , composed of a rapid and transient peak followed by a sustained plateau phase (20,21). Compared with the control, we observed a larger rise in peak Ca 2ϩ i in MIN6 cells with depleted T1R3 as well as a faster decline during the second phase of the carbachol response ( Fig. 1, A and B).
One possible reason for enhanced carbachol-induced changes in Ca 2ϩ i in MIN6 T1R3 knockdown cells could be increased expression of mAChRs. Therefore, we performed real -time quantitative PCR to compare changes in expression of the G q -coupled mAChRs, M1, M3, and M5, in MIN6 control and T1R3 knockdown cells. There was 2.35 ϩ 0.06-fold increase in M3R mRNA in the T1R3knockdown cells compared with the control, whereas the expression of M1 and M5 receptors did not change significantly (Fig. 1C). The increase in M3R mRNA in T1R3 knockdown cells was mirrored by an increase in M3R protein expression that was found by immunoblotting lysates from control and T1R3 knockdown cells (Fig. 1D). Quantitation of the immunoblot analyses indicated a 3.8 ϩ 0.9fold increase in M3R protein in MIN6 cells with suppressed T1R3 expression (Fig. 1D).
In addition to mAChRs, we also examined the expression of the regulator of G protein signaling 4 (RGS4) and transient receptor potential channel 6 (TRPC6). RGS proteins are GTPase-activating proteins that enhance G␣-GTP hydrolysis, thereby decreasing the lifetime of active states of G protein subunits, and RGS4 is expressed in ␤ cells (22). TRPC6 has been reported to be a receptor-operated cation channel that is activated upon GPCR stimulation and subsequent phospholipase C ␤ activation (23,24). Calcium entry occurring through plasma membrane channels as a result of GPCR activation independent of the state of Ca 2ϩ i stores is referred to as ROCE (25). It has been demonstrated that carbachol is capable of inducing

T1R3 and Muscarinic Signaling
ROCE by activating and promoting cell surface expression of TRPC6 downstream of muscarinic receptor binding (26,27). Expression of neither RGS4 nor TRPC6 was significantly different in T1R3 knockdown cells compared with the control, suggesting that the enhanced carbachol response was not a result of decreased RGS4 or increased TRPC6 expression (Fig. 1C).
Carbachol-induced Changes in Ca 2ϩ i in MIN6 Cells with Suppressed T1R3 Expression Are Largely Dependent on Release of Ca 2ϩ from Intracellular Stores-The initial rapid rise in calcium observed upon carbachol stimulation in ␤ cells has been shown to be due to the inositol 1,4,5-triphosphate-mediated release of calcium from ER stores, whereas the second phase is maintained by SOCE (21). SOCE refers to calcium influx that occurs through store-operated calcium channels (SOCCs) as a result of Ca 2ϩ i store depletion (28). We investigated whether the contributions of Ca 2ϩ i stores or SOCE to carbachol-induced changes in Ca 2ϩ i differed in MIN6 cells with suppressed T1R3 expression compared with control cells. Cells were stimulated with carbachol in the absence of extracellular calcium to determine whether activation of M3R in the T1R3 knockdown cells induced a larger rise in Ca 2ϩ i as a result of the release of calcium from intracellular stores. We found that carbachol stimulated a larger rise in Ca 2ϩ i in the absence of extracellular calcium in the T1R3 knockdown cells compared with control cells (Fig. 2, A and B). This finding suggests that, in addition to elevated M3R, a larger release of calcium from intracellular stores may contribute to the enhanced first phase of the carbachol response in the MIN6 T1R3 knockdown cells.
We next determined whether the larger carbachol-induced release from intracellular stores was a consequence of more calcium stored in the ER of cells with reduced T1R3 by treating cells with thapsigargin, an inhibitor of the sarco/endoplasmic reticulum Ca 2ϩ ATPase, to deplete ER stores of calcium (29). In the absence of extracellular calcium, thapsigargin induced a similar rise in Ca 2ϩ i in T1R3 knockdown and control cells (Fig.  2, C and D). When extracellular calcium was replenished, allowing SOCE to occur, calcium entry through SOCCs was similar in both T1R3 knockdown and control cells. We also explored ROCE, after calcium restoration and SOCE had occurred, by stimulating cells with carbachol. Despite little or no change in TRPC6 expression in MIN6 cells depleted of T1R3, ROCE was absent (Fig. 2, C and D), suggesting that another, as yet unidentified channel may contribute to ROCE in MIN6 cells.
Consistent with the larger release of calcium from intracellular stores induced by carbachol, a greater portion of the peak carbachol response was sensitive to inhibition by thapsigargin in the T1R3 knockdown cells (Fig. 3A). Thapsigargin blocked a greater portion of carbachol-induced changes in Ca 2ϩ i at 2 min in the control cell line. It is possible that SOCE triggered by carbachol generates a smaller influx of Ca 2ϩ through SOCCs in the T1R3 knockdown cells compared with control cells. This  MAY 16, 2014 • VOLUME 289 • NUMBER 20 JOURNAL OF BIOLOGICAL CHEMISTRY 14373 might explain why the second phase of the calcium response declined at a faster rate and was less affected by thapsigargin in T1R3 knockdown cells. 2-APB, an inositol 1,4,5-triphosphate receptor antagonist, inhibited peak and 2 mincarbachol-induced rises in Ca 2ϩ i to a similar degree in both control and T1R3 knockdown cells (Fig. 3, A and B). It is conceivable that the inhibitory effect of 2-APB on carbachol-stimulated changes in Ca 2ϩ i was due to its actions on SOCCs rather than through inhibition of inositol 1,4,5-triphosphate receptors (30). Using an ER-localized FRET sensor, it has been reported that, in MIN6 cells, carbachol-induced reductions in ER calcium were not inhibited by pretreatment with 2-APB despite lower overall Ca 2ϩ i (14). To determine whether or not 2-APB had an effect on SOCE, calcium was depleted and then added back to MIN6 cells, as shown in Fig. 2C, and pretreated with 2-APB for 30 min. In 2-APB-treated cells, the ability of thapsigargin to deplete stores and the subsequent SOCE was reduced significantly, whereas ROCE was not altered (Fig. 3, C and D).

T1R3 and Muscarinic Signaling
It has been shown previously that nifedipine, an L-type voltage-gated calcium channel blocker, inhibits carbachol-induced changes in Ca 2ϩ i in MIN6 cells, suggesting that carbachol causes membrane depolarization (12). It has been proposed that phosphatidylinositol 4,5-bisphosphate hydrolysis, resulting from carbachol-induced phospholipase C activation, leads to decreased K ATP channel activity. Decreased K ATP channel activity results in membrane depolarization and activation of voltage-gated calcium channels. Nifedipine had a greater effect on the carbachol response in the control cells compared with those with reduced T1R3, consistent with the observation that calcium released from intracellular stores contributes a greater portion of the carbachol-induced rise in Ca 2ϩ i in the T1R3 knockdown cells than does extracellular calcium (Fig. 3, A and  B). To determine whether nifedipine was having an effect on SOCE or ROCE, MIN6 cells were pretreated with nifedipine for 30 min prior to stimulation. Unlike what was observed with 2-APB, nifedipine only had an effect on ROCE (Fig. 3, C and D). This is consistent with the data in Fig. 3, A and B, which demonstrated that nifedipine had a smaller effect on carbacholinduced changes in Ca 2ϩ i in the T1R3 knockdown cells in which ROCE was absent.

Carbachol-induced Changes in Ca 2ϩ i and ERK1/2 Phosphorylation Are Enhanced in MIN6 Cells Deprived of Amino
Acids-Because we observed these changes in carbachol signaling in MIN6 cells in which T1R3 expression was stably suppressed and because we have shown previously that the T1R1/T1R3 complex is an early sensor of amino acids, we hypothesized that depriving cells of amino acids may mimic some of the altered carbachol signaling observed in T1R3 knockdown cells. Therefore, we performed experiments in which MIN6 cells were incubated in KRBH with 10% dia- lyzed serum, 4.5 mM glucose, and either 1.0ϫ amino acids, representing the full complement of amino acids found in Dulbecco's modified Eagle's medium, or 0.1ϫ amino acids for 16 h to approximate amino acid starvation. Cells were then stimulated with carbachol for 1, 2, or 5 min, and pERK1/2 were assessed by immunoblotting. We observed lower basal pERK1/2 following amino acid deprivation compared with that in cells in complete medium (Fig. 4A). The pERK1/2 in carbachol-stimulated cells was approximately the same in both cell lines. However, the low basal activity resulted in a larger fold increase in pERK1/2 induced by carbachol in the amino acid-deprived cells compared with control cells (Fig. 4B).
We next determined whether the enhanced effect of carbachol on ERK1/2 phosphorylation was paralleled by a larger rise in Ca 2ϩ i in amino acid-deprived cells. Indeed, a larger rise in peak Ca 2ϩ i was observed with carbachol stimulation in MIN6 cells deprived of amino acids compared with non-deprived cells (Fig. 4, C and D). Although similar, the change was not as great as that observed in MIN6 T1R3 knockdown cells (Fig. 1, A and  B). We also noted significantly lower basal calcium in the deprived cells compared with non-deprived cells (average basal 340/380 value, 0.1ϫ amino acids ϭ 0.2200 Ϯ 1.191e-4; 1.0ϫ amino acids ϭ 0.2321 Ϯ 1.493e-4; p Ͻ 0.001), which may offer an explanation for the lower basal pERK1/2 observed under these same conditions (Fig. 4A). In addition, although peak Ca 2ϩ i induced by carbachol was significantly higher in deprived cells, the second phase of the response assessed at 2 min did not differ (Fig. 4, C and D).
Because M3 receptor expression increased in T1R3 knockdown cells, we wanted to determine whether amino acid deprivation also affected M3R expression. As shown in Fig. 5, A and B, M3R mRNA and protein were increased in MIN6 cells deprived of amino acids compared with non-deprived cells (1.89 ϩ 0.19-and 1.51 ϩ 0.13-fold, respectively). To determine whether this was a general response to nutrient deficiency, we examined effects of lowering the glucose concentration from 25 to 2 mM on M3R expression. In this case, we found that M3R expression was reduced, not increased (Fig. 5C), indicating a differential responsiveness of M3R expression to amino acids and glucose.
We verified, in human islets, that amino acid deprivation affected M3R expression. As shown in Fig. 5D, M3R expression was enhanced 1.5-fold in islets of higher purity (90 and 80%). In islets of lower purity (70%), no change was observed. The fold increase of M3R expression in deprived islets was less than that in deprived MIN6 cells (Fig. 5A, ϳ1.9-fold). This difference may be due to varied responses among other cell types in islets, in contrast to the relative homogeneity of the MIN6 cell line.

Carbachol-stimulated Mobilization of Calcium from Intracellular Stores Is Unaltered in MIN6 Cells Deprived of Amino
Acids-We further investigated whether reduced amino acids could mimic T1R3 knockdown in MIN6 cells by determining the ability of carbachol to induce the release of calcium from intracellular stores under these conditions. In contrast to the significantly enhanced ability of carbachol to stimulate the release of calcium from stores in T1R3 knockdown cells (Fig. 2,  A and B), there was no difference in the release from stores  MAY 16, 2014 • VOLUME 289 • NUMBER 20 comparing cells with normal and low amino acids (Fig. 6, A and  B).

T1R3 and Muscarinic Signaling
We also investigated whether depletion of intracellular calcium stores changed SOCE or ROCE in MIN6 cells deprived of amino acids to compare with changes in MIN6 T1R3 knockdown and control cells (Fig. 2, C and D). Contrary to what was observed in MIN6 cells with depleted T1R3, Ca 2ϩ i resulting from store depletion, SOCE, or ROCE was higher in amino acid-deprived cells compared with non-deprived cells (Fig. 6, C  and D). It is possible that, in amino acid-deprived cells, larger amounts of calcium are stored in the ER because of reduced activation of T1R3 in the absence of amino acids. The larger release of calcium from stores induced by thapsigargin could cause a greater influx of calcium through SOCCs, consistent with enhanced SOCE in deprived cells (Fig. 6, C and D). The altered ROCE observed in deprived cells could not be associated with an increase in TRPC6 mRNA, again suggesting the involvement of another channel in ROCE in MIN6cells (Figs. 5 and 6, C and D). Together, these findings indicate that, although M3R expression is up-regulated by lowering amino acids in the medium, leading to a larger rise in carbachol-induced changes in Ca 2ϩ i and ERK1/2 phosphorylation, the effects on intracellular stores, SOCE, and ROCE that were observed in MIN6 T1R3 knockdown cells were not mimicked. Finally, because we showed that fasting increased the expression of T1R3 in mouse tissues (12), we wondered whether amino acid deficiency also had an effect on T1R3 under these conditions. We found that expression of T1R3 mRNA was increased by incubation in 0.1ϫ amino acids (Fig. 6E).

DISCUSSION
We explored the underlying mechanisms for the enhanced carbachol response in MIN6 cells following depletion of T1R3 and found that there was a significant increase in M3R expression in T1R3-depleted cells. A similar increase in M3R was also observed in amino acid-deprived cells. Thus, the increase in M3R is, at least in part, a response to amino acid deficiency. This change in M3R expression is not observed by reducing glucose. Given the remarkable sensitivity of ␤ cells to changes in glucose concentration, perhaps it is not so surprising that lowering glucose from 25 to 2 mM, a concentration often used as the control condition in experiments with these cells, caused a decrease in M3R expression. This result further emphasizes that the details of how isolated ␤ cells or islets are handled ex vivo may have a large impact on signaling capability.
We evaluated the potential impact of other molecules known to impact ␤ cell function through connections to GPCRs. RGS4 terminates signaling from M3Rs, thereby inhibiting their function in MIN6 cells and primary mouse islets (7). It seemed possible that RGS4 might have been suppressed in T1R3 knockdown cells to increase M3R function, but experimental findings did not support this idea. We also examined TRPC6 because a microarray analysis suggested that it was up-regulated in T1R3 knockdown cells. Carbachol is capable of inducing ROCE by activating and promoting cell surface expression of TRPC6 downstream of muscarinic receptor binding (26,27). Although the suspected change in its mRNA could not be validated, TRPC6 seemed a logical candidate and may be involved in some manner we did not detect (23,24).
Differences in M3R signaling between shRNA depletion of T1R3 and amino acid limitation included effects on the kinetics of changes in intracellular free calcium. It is possible that differences in intracellular amino acid concentrations may have contributed to altered calcium signaling. Previously, we measured amino acid amounts in cells with T1R3 knocked down following amino acid withdrawal and subsequent amino acid repletion (12). Intracellular concentrations dropped rapidly when extracellular amino acids were removed. In contrast to amino acid withdrawal, T1R3 receptor knockdown did not reduce intracellular amino acid content. Branched side chain amino acids in particular were similar. Thus, intracellular amino acids may have led to some of the calcium responses that differed between receptor knockdown and amino aciddeprived cells. Finally, increased T1R3 expression in the amino acid-deprived cells may also have contributed to the differences noted.
Our findings suggest that multiple types of nutrient responses are linked with the M3R to support ␤ cells during nutrient stress. Beneficial effects of M3R activation in ␤ cells have been deduced from defects in M3R knockout mice and were also revealed in studies of a designer G q -coupled receptor engineered by incorporating mutations in M3R that rendered the receptor unresponsive to ACh but selectively sensitive to activation by the pharmacologically inert compound clozapine-N-oxide (13,(31)(32)(33). Chronic activation of the ␤ cell G q -coupled designer receptor in mice resulted in enhanced insulin release, decreased blood glucose concentrations, augmented ␤ cell mass because of stimulation of ␤ cell proliferation, increased insulin content, and amplified expression of several genes critical for ␤ cell function (13,33).
In addition to studies performed in mice, variations in the gene encoding M3R in humans are associated with a reduced acute insulin response and increased risk for early-onset type 2 diabetes (34). Type 2 diabetes is characterized by hyperglycemia resulting from the inability of ␤ cells to secrete sufficient insulin to overcome peripheral insulin resistance (35). The increased demand on ␤ cells to secrete insulin leads to ␤ cell exhaustion, reduced ␤ cell mass, and impaired insulin production (35,36). The increase in M3R expression in amino aciddeprived cells provides independent support for the idea that modulating the expression of and/or signaling through ␤ cell M3Rs enhances ␤ cell function and protects against some types of nutrient stress.