Absinthin, an agonist of the bitter taste receptor hTAS2R46, uncovers an ER-to-mitochondria Ca2+–shuttling event

Type 2 taste receptors (TAS2R) are G protein–coupled receptors first described in the gustatory system, but have also been shown to have extraoral localizations, including airway smooth muscle (ASM) cells, in which TAS2R have been reported to induce relaxation. TAS2R46 is an unexplored subtype that responds to its highly specific agonist absinthin. Here, we first demonstrate that, unlike other bitter-taste receptor agonists, absinthin alone (1 μm) in ASM cells does not induce Ca2+ signals but reduces histamine-induced cytosolic Ca2+ increases. To investigate this mechanism, we introduced into ASM cells aequorin-based Ca2+ probes targeted to the cytosol, subplasma membrane domain, or the mitochondrial matrix. We show that absinthin reduces cytosolic histamine-induced Ca2+ rises and simultaneously increases Ca2+ influx into mitochondria. We found that this effect is inhibited by the potent human TAS2R46 (hTAS2R46) antagonist 3β-hydroxydihydrocostunolide and is no longer evident in hTAS2R46-silenced ASM cells, indicating that it is hTAS2R46-dependent. Furthermore, these changes were sensitive to the mitochondrial uncoupler carbonyl cyanide p-(trifluoromethoxy)phenyl-hydrazone (FCCP); the mitochondrial calcium uniporter inhibitor KB-R7943 (carbamimidothioic acid); the cytoskeletal disrupter latrunculin; and an inhibitor of the exchange protein directly activated by cAMP (EPAC), ESI-09. Similarly, the β2 agonist salbutamol also could induce Ca2+ shuttling from cytoplasm to mitochondria, suggesting that this new mechanism might be generalizable. Moreover, forskolin and an EPAC activator mimicked this effect in HeLa cells. Our findings support the hypothesis that plasma membrane receptors can positively regulate mitochondrial Ca2+ uptake, adding a further facet to the ability of cells to encode complex Ca2+ signals.

The existence of a family of bitter taste receptors was predicted over 20 years ago by Lush (1), and bitter taste receptor (TAS2R) genes were first described in 2000 (2). In humans, there are 25 subtypes of G protein-coupled receptors referred to as hTAS2Rs 4 (3)(4)(5) that vary in their selectivity toward bitter compounds: Some subtypes recognize few chemically restricted molecules, whereas some others respond to a wide range of diverse ligands (6). In the same way, some bitter compounds are known to be agonists for a single hTAS2R subtype, whereas others are not selective (7).
Recently, hTAS2Rs have been reported to be expressed, alongside tongue and palate epithelia, also on extraoral tissues such as gut, genitourinary system, brain, immune cells, and respiratory system (8,9). In airway smooth muscle (ASM), agonists of the most-represented bitter taste receptor subtypes (TAS2R10, 14, and 31) have been identified as novel pharmacological targets in obstructive pulmonary disease therapy (10 -14). The expression of TAS2R and the bronchodilatory effect of the agonists have been confirmed in mouse (15), guinea pig (16), and human (12,17,18). Paradoxically, the bronchodilation mechanism induced by activation of TAS2R10, 14, and 31 in response to a contractile stimulus is mediated by an increase in cytosolic calcium followed by membrane hyperpolarization through large-conductance potassium channels (BK Ca ) (11,16). Interestingly, it has been demonstrated that the level of calcium increase is correlated to the level of receptor expression (11). More recently, Tan and Sanderson (19) have found that agonists of TAS2R10, at concentrations required to dilate mouse constricted airways, do not increase cytosolic calcium in ASM, leading to the hypothesis that TAS2R agonists bronchodilate by inhibiting IP 3 receptors and reducing calcium sensitivity.
In the present manuscript, we focused on hTAS2R46, an unexplored hTAS2R subtype receptor in ASM (13). An extensive structure-function analysis of this subtype has been performed (20) and sesquiterpene lactones, such as absinthin, represent selective agonists (6,21).
cro ARTICLE reduced influx through the plasma membrane nor of a decrease in Ca 2ϩ efflux from the endoplasmic reticulum, but is a consequence of an increased Ca 2ϩ uptake by mitochondria. Indeed, cytosolic Ca 2ϩ decreases and simultaneous mitochondrial Ca 2ϩ increases were sensitive to the nonspecific mitochondrial calcium uptake inhibitor carbamimidothioic acid (KB-R7943), to the mitochondrial uncoupler carbonyl cyanide p-trifluoromethoxyphenyl-hydrazone (FCCP), to the selective hTAS2R46 inhibitor 3␤-hydroxydihydrocostunolide (3HDC) (7), to the cytoskeletal disrupter latrunculin, and to the EPAC antagonist ESI-09. Furthermore, the effect of absinthin was no longer evident in hTAS2R46silenced cells. Our observation that absinthin, activating hTAS2R46, modulates histamine-induced cytosol Ca 2ϩ rises by potentiating mitochondrial Ca 2ϩ uptake shows that regulated mitochondrial Ca 2ϩ uptake may participate in the encoding of Ca 2ϩ signals. We also provide data that this ER-to-mitochondria shuttling may be a general feature of Ca 2ϩ signaling as salbutamol, a ␤ 2 -agonist, also reduces histamine-induced Ca 2ϩ rises in the cytosol while simultaneously increasing mitochondrial calcium.

Absinthin reduces histamine-induced Ca 2؉ increases via hTAS2R46
First, we evaluated the effect of absinthin on Ca 2ϩ rises in ASM. Although other bitter taste receptor ligands have been shown to increase cytosolic Ca 2ϩ (11), absinthin (10 M) was unable to induce any Ca 2ϩ rise (n ϭ 7) (Fig. 1A).
We next evaluated the effect of absinthin on the Ca 2ϩ -rise induced by one of the most potent bronchoconstrictors, histamine (10 M). As shown in Fig. 1, A and B, histamine induced a rapid Ca 2ϩ rise, as determined by Fura-2. The simultaneous co-addition of absinthin, surprisingly, led to a significantly lower histamine-induced Ca 2ϩ peak. This effect was evident when pooling the data together but was just as evident at the single cell level in fluorescence microscopy (data not shown). Because histamine-induced Ca 2ϩ rises are the result of an initial Ca 2ϩ release from intracellular stores and a secondary Ca 2ϩ entry, we performed experiments in a Ca 2ϩ -free extracellular buffer to dissect the two components. As demonstrated in Fig.  1B, histamine induced a cytosolic Ca 2ϩ rise in this condition, which was of lower amplitude compared with the previous condition. Yet, absinthin was still able to significantly reduce the effect of histamine.
Absinthin is reported to bind different bitter taste receptor subtypes, although it is generally thought to act mainly on hTAS2R46 (6). We therefore evaluated whether hTAS2R46 was present in our cellular model. Indeed, RT-PCR experiments showed that mRNA coding for this subtype was present in ASM, although at lower levels compared with human bronchial muscle and human epithelial tongue (Fig. S1A). This was confirmed also by Western blotting, immunofluorescence, and immunohistochemistry of ASM with specific antibodies (Fig.  S1, B-F). We next evaluated whether 3HDC, a potent antagonist of hTAS2R46 (7), was able to revert the effect of absinthin on histamine-induced Ca 2ϩ rises. As shown in Fig. 1C, 3HDC dose-dependently reverted the effect of absinthin, with concentrations of 10 M abolishing completely the effect of the agonist. To further determine the specificity of this effect, we performed experiments in HeLa cells that are devoid of hTAS2R46 (Fig. S1B) but that have phospholipase C-coupled histamine receptors (22). As shown in Fig. 1D, absinthin was unable to affect histamine-induced Ca 2ϩ rises in HeLa cells, even at high concentrations (100 M). Last, we made use of shRNA to confirm that it was indeed hTAS2R46 mediating the absinthin response. The use of shRNA led to a decrease of the mRNA encoding for the receptor in ASM cells of 90 Ϯ 2.5% (as determined by RT-PCR and Western blotting) compared with control ASM cells (Fig. S1, A and B). In cells with a significantly reduced level of hTAS2R46, absinthin was unable to reduce histamine-induced cytosolic Ca 2ϩ rises (Fig. 1E).
These data demonstrated that absinthin reduced cytosolic Ca 2ϩ -rises induced by histamine by a receptor-specific mechanism mediated by hTAS2R46. This effect occurs rapidly and appears to affect the Ca 2ϩ -release component of the rise, given that it is apparent also in Ca 2ϩ -free conditions.

Absinthin does not affect subplasmallemal Ca 2؉
To evaluate whether the Ca 2ϩ -entry component of the histamine-induced Ca 2ϩ rise was also affected, we decided to use aequorin-based probes that can be targeted selectively to subcellular compartments. At first, we confirmed the effect of absinthin on histamine-induced Ca 2ϩ rises using a nontargeted cytosolic aequorin (cytAEQ) (23). In accordance with what was observed with Fura-2, absinthin dampened the peak of the Ca 2ϩ rise in a significant manner ( Fig. 2A). We next decided to evaluate whether Ca 2ϩ entry was affected by using an aequorin probe directed to the subplasma membrane compartment (pmAEQ) (24,25). In this instance, histamine-induced Ca 2ϩ entry was not affected by the co-addition of absinthin (Fig. 2B).
To further confirm that Ca 2ϩ entry was not involved, we performed a store depletion experiment investigating cytosolic calcium via Fura-2. Stores were depleted with tBHQ (50 M) in a calcium-free solution and calcium was re-added after 5 min in the presence or absence of absinthin. As shown in Fig. S2, absinthin was unable to affect the extent of store-operated Ca 2ϩ entry. These data therefore suggest that absinthin targets selectively the intracellular store component of histamine-induced Ca 2ϩ rises.

Absinthin controls mitochondrial calcium buffering
To reconcile the above, partly contradictory observations on absinthin (i.e. a decreased cytosolic calcium rise together with an unchanged Ca 2ϩ entry), we inquired whether mitochondrial Ca 2ϩ in response to histamine would change in accord to cytosolic Ca 2ϩ . If absinthin dampens Ca 2ϩ release via IP 3 R modulation or via IP 3 production, it would be expected that mitochondrial Ca 2ϩ uptake in the presence of absinthin after stimulation with histamine should be lowered as well. To test this hypothesis, we used an aequorin probe that targets selectively the mitochondrial matrix (mitAEQ) (24,26). As shown in Fig. 3A, stimulation of cells with histamine provoked an increase in mitochondrial Ca 2ϩ hTAS2R46 regulates mitochondrial calcium buffering (27) whereas the addition of absinthin alone had no effect. In striking contrast to the effect on cytosolic Ca 2ϩ , the histamine-induced mitochondrial Ca 2ϩ rise in the presence of absinthin was significantly greater compared with histamine alone (Fig. 3A).
The above data appeared to suggest that the reduced cytosolic Ca 2ϩ rise is paralleled by a paradoxical increase in mito-chondrial Ca 2ϩ uptake, i.e. it would suggest that histamineinduced Ca 2ϩ release is unaffected by absinthin, but the Ca 2ϩ is channeled to the mitochondria and not to the cytosol. If this were the case, it would be expected that disrupting the mitochondrial membrane potential, which drives Ca 2ϩ uptake by the mitochondrial calcium uniporter (MCU), with FCCP, and thereby not permitting mitochondrial Ca 2ϩ  (1,20) ϭ 16.45, F cr ϭ 4.35;°°°, p ϭ 6.17⅐10 Ϫ4 ), whereas no significant difference was found between Hist and 3HDC (10 M)ϩHistϩAbs (Student's t test, t (10) ϭ 0.13, t cr ϭ 2.23, p ϭ 0.90). D, cytosolic calcium release measurement in nonexpressing hTAS2R46 HeLa cells stimulated with histamine (Hist, 100 M) and absinthin at the indicated concentrations. No significant relation emerged by linearly regressing cytosolic free calcium against increasing doses of Abs (F (1,16) ϭ 0.009, F cr ϭ 4.49, p ϭ 0.93). E, ASM cells infected with LVs carrying shRNA-hTAS2R46 and treated as in A. Absinthin modulation of histamine-induced calcium release was analyzed in the two cellular models through a two-way ANOVA. The presence of a statistically significant interaction between the two factors (F (1,14) ϭ 11.40, F cr ϭ 4.6, p ϭ 0.0045) reflects the simple main effects for which the significant decrease in cytosolic calcium following absinthin addition observed in ASM (°°°, p ϭ 7.01⅐10 Ϫ4 ) completely vanished when moving to ASM shRNA-hTAS2R46 model (p ϭ 0.62).

hTAS2R46 regulates mitochondrial calcium buffering
uptake, should block the Ca 2ϩ increase in mitochondria induced by histamine in the presence of absinthin and simultaneously restore the histamine-Ca 2ϩ rise in the cytosol. Indeed, when FCCP was used, this resulted in a significantly reduced Ca 2ϩ rise in mitochondria ( Fig. 3A) and an increase in the cytosolic Ca 2ϩ rise that was equivalent, if not greater, to histamine alone (Fig. 3B). Subplasmalemmal Ca 2ϩ , on the contrary, was unchanged, suggesting that the extent of mitochondrial uptake does not influence Ca 2ϩ entry via the plasma membrane (Fig. 3C).
Ca 2ϩ -uptake in mitochondria is mediated by MCU, a protein complex whose main components have been recently identified (28,29). In the presence of KB-R7943, a nonspecific inhibitor of the uniporter (30), the effect of absinthin on histamine-induced mitochondrial Ca 2ϩ rises was abolished (Fig. 3A), as was the effect on cytosolic Ca 2ϩ (Fig. 3B), whereas again no difference was perceivable in subplasmalemmal Ca 2ϩ (Fig. 3C). FCCP or KB-R7943 alone did not affect basal Ca 2ϩ or Ca 2ϩ signals induced by histamine (Fig. S3, A-D).

An intact cytoskeletal structure is necessary for the action of absinthin
The above data would suggest that absinthin signals to the mitochondria to increase the uptake of the Ca 2ϩ released from the endoplasmic reticulum (ER). If this were the case, it could be expected that disrupting the ER-mitochondria juxtapositions would, at least in part, abolish the effect of absinthin. To this extent, we pre-incubated cells with latrunculin, a natural compound that disrupts the actin cytoskeleton thereby uncoupling organelle cross talk (31). As observed in Fig. 4A, in cells preincubated with latrunculin, the cytosolic Ca 2ϩ rises induced by histamine were similar in the presence or absence of absinthin in the cytosol ( Fig. 4A; see Fig. 1 for traces in the absence of absinthin) under the plasma membrane (Fig. 4B) and in mitochondria (Fig. 4C).

Neither a change in mitochondrial membrane potential nor a change in ATP/ADP ratio occurs upon absinthin addition
We reasoned that absinthin might modify the mitochondrial potential, and this might then be responsible for the increased mitochondrial accumulation. To test for this, we used the JC-1 probe, but found no difference of mitochondrial membrane potential in the presence or absence of absinthin (Fig. S4A).
Next, we investigated whether an increase in mitochondrial Ca 2ϩ during this time frame could significantly increase cytosolic ATP levels. If this were the case, it could be expected that extruding mechanisms could be more efficient and it could hTAS2R46 regulates mitochondrial calcium buffering account for part of the drop in cytosolic Ca 2ϩ . As shown in Fig.  S4B, absinthin, unlike the positive control FCCP, was unable to modify significantly ATP to ADP ratios.

␤ 2 -adrenoreceptor activation mimics the effects of hTAS2R46 activation
To evaluate whether other plasma membrane receptor agonists were able to trigger a similar mechanism, we investigated the Ca 2ϩ response of histamine in the presence of a ␤ 2 agonist, the most widely used bronchodilators in asthma and chronic obstructive pulmonary disease. As expected, salbutamol (10 M) lowered the Ca 2ϩ rise induced by histamine (Fig. 5A). This has been reported previously and has been attributed to a number of cAMP-and PKA-dependent effects, including IP 3 receptor modulation (32). Surprisingly, though, histamine-induced mitochondrial Ca 2ϩ increased significantly also in the presence of salbutamol (Fig. 5A).

cAMP or EPAC modulation modifies cytosolic and mitochondrial histamine-induced Ca 2؉ transients in HeLa and ASM cells
Given that ␤ 2 receptors are coupled to cAMP formation, we attempted to reproduce the effect with forskolin (10 M). In the presence of forskolin, cytosolic histamine-induced Ca 2ϩ rises were significantly blunted (Fig. 5A). When evaluating mitochondrial calcium, the concentration was also blunted, although not to the same extent (Fig. 5A). Yet, the effect of a strong cAMP production induced by forskolin prevented firm conclusions.
From the data on forskolin and, more importantly on salbutamol, it would appear that a cAMP-dependent pathway may hTAS2R46 regulates mitochondrial calcium buffering control ER to mitochondria Ca 2ϩ shuttling. It has been recently reported that Epac1 is localized on the mitochondrial inner membrane and matrix (and may control MCU activity) (33,34). To test whether the actions of absinthin were mediated by EPAC, we performed identical experiments to those described using the EPAC-specific inhibitor ESI-09. As shown in Fig. 5B, ESI-09 was able to counteract the effect of absinthin on both cytosolic and mitochondrial Ca 2ϩ .
We next proceeded to investigate cAMP rises upon histamine, histamine/absinthin, and forskolin in ASM cells using a cytosolic FRET probe. As expected, forskolin induced a rise in cytosolic cAMP in the same time frame as the regulation of histamine-induced Ca 2ϩ rises. On the contrary, however, absinthin did not appear to have any direct effect on cAMP levels in this time frame (Fig. S5).
Last, we investigated whether the same effect was observable in another cell type, HeLa. As shown in Fig. 5C both forskolin and the EPAC activator 8-pCPT-2Ј-O-Me-cAMP were able to reduce cytosolic Ca 2ϩ rises induced by histamine while increasing mitochondrial Ca 2ϩ rises, confirming that the phenomenon is likely to be present in other cell types.

Discussion
In the present report we show that absinthin, via hTAS2R46, inhibits histamine-induced cytosolic Ca 2ϩ rises in ASM cells via a positive modulation of mitochondrial Ca 2ϩ uptake possibly through the regulation of EPAC. The evidence that this occurs via a specific plasma membrane receptor is given by (i) the relative specificity of absinthin; (ii) the sensitivity of the effect to the specific receptor inhibitor 3HDC; and (iii) the inability of absinthin to elicit the same effect in HeLa cells, where hTAS2R46 is not expressed and in hTAS2R46-silenced ASM cells. The evidence that mitochondrial uptake is the mechanism by which this occurs is given by (i) the significant increase in mitochondrial calcium paralleled by the decrease in cytosolic calcium of the [Ca 2ϩ ]i-histamine induced in the presence of absinthin; (ii) the abolition of the effect in the presence of the mitochondrial uncoupler FCCP; and (iii) the abolition of the effect in the presence of the mitochondrial Ca 2ϩ uptake inhibitor KB-R7943, the similarity in subplasmalemmal calcium, and in tBHQ-triggered store-operated Ca 2ϩ entry, suggesting that store emptying is similar in the presence or absence of absinthin. The use of FCCP leads to a cell-wide decrease in ATP to ADP ratio and this is a limitation of our data as it might have confounded partly the results (given that Ca 2ϩ extrusion and reuptake should be hampered), but we believe that overall our data are consistent across experiments.
The evidence that EPAC is most likely the mediator of this effect is given by (i) the ability of salbutamol to elicit identical effects; (ii) the fact that forskolin is in part able to mimic the effect; more importantly (iii) the ability of ESI-09 to abolish the effects of absinthin; and (iv) the ability of the EPAC activator 8-pCPT-2Ј-O-Me-cAMP to increase mitochondrial Ca 2ϩ in HeLa cells.

hTAS2R46 regulates mitochondrial calcium buffering
We believe this is the first report of a positive modulation of mitochondrial Ca 2ϩ uptake by plasma membrane G proteincoupled receptors, although it had been long known that receptor agonists are able to increase mitochondrial Ca 2ϩ (35) and that postulations on the role of mitochondrial uptake in shaping cytosolic signals date back in time (36). Indeed, one of the most striking characteristics of this modulation is the speediness by which it occurs, as simultaneous activation of histamine receptors and TAS2R46/␤ 2 -adrenoreceptors is sufficient to elicit an effect. When superimposing traces, the mitochondrial Ca 2ϩ transient lags just a few seconds behind the cytosolic transient.
Although it is possible that TAS2R46 activation directly affects the signaling of histamine receptors (for example reducing the Gq coupling), this would explain the reduction in cytosolic Ca 2ϩ but would be unable to explain the increase in mitochondrial Ca 2ϩ and the absence of a difference in subplasmalemmal Ca 2ϩ .
It is thought that microdomains of high Ca 2ϩ at the mouth of IP 3 receptors are necessary in ER-mitochondria juxtapositions to drive Ca 2ϩ uptake into mitochondria. Ca 2ϩ flows through the voltage-dependent anion channel (VDAC), located on the mitochondrial outer membrane, and then through the mitochondrial Ca 2ϩ uniporter located on the inner membrane. Importantly, a Ca 2ϩ increase in mitochondria has physiological repercussions, including an increased activity of pivotal oxidative phosphorylation enzymes (37) and a protection from excessive Ca 2ϩ surges. It has been recently suggested that Epac1 favors Ca 2ϩ exchange between the ER and the mitochondrion in heart via the VDAC/GRP75/IP3R1 complex (34), and it is likely that a similar mechanism occurs in smooth muscle cells.
In this respect, it can be imagined that TAS2R46 activation not only controls histamine-induced Ca 2ϩ contractions but also provides an energetic boost to smooth muscle cells and protects them. Indeed, genetic ablation or inhibition of EPAC1 in cardiac myocytes has been shown as protective in myocardial ischemia/reperfusion injury (33). In this respect, it is interesting to note that a TAS2R ligand has been recently reported to decrease mitochondrial membrane potential and increase mitochondrial reactive oxygen species and mitochondrial fragmentation (38). Yet, when we tested membrane potential changes with absinthin, we were unable to detect any. Furthermore, an energetic boost stimulated by Ca 2ϩ (39) should be expected to increase ATP to ADP ratio in the cytosol (which in turn could also render Ca 2ϩ pumps more efficient and partly explain the cytosolic drop). Yet, by using a PercevalHR probe, which measures ATP/ADP ratios (see supporting Experimental procedures), we were unable to detect any significant changes in the time frame of treatment. The possibility that ATP/ADP ratios were below our threshold of detection is possible.
Although we have not investigated the molecular mechanism of action of the Ca 2ϩ shuttling, a few hypotheses can be made. First, TAS2R46 could be controlling ER-mitochondria juxtapositions, a mechanism that has been postulated by others to modify mitochondrial uptake (40 -42). In this respect, when we performed time-lapse microscopy using MitoTracker, we were unable to see any significant movements of mitochondria during the 100 s post treatment (data not shown). Alternatively, a posttranslational event could occur either on VDAC/MCU or one of its regulatory proteins (e.g. MICU, mitofusin). VDAC is known to increase its conductance for Ca 2ϩ in its closed conformation (43) and such an event could explain the increased amount of Ca 2ϩ entering the mitochondria. A regulation of MCU is also possible, although its localization on the inner membrane would create a topological problem, it should be noticed that the regulation is rapid, as simultaneous addition of histamine and absinthin/salbutamol is sufficient to unmask the effect.
Independently of the mechanism by which this occurs, we believe we have shown that activation of TAS2R46 leads to a redirection of IP 3 R Ca 2ϩ efflux from the cytosol to the mitochondria (i.e. from cytosolic modulation to mitochondrial regulation), and this is most likely mediated by EPAC, providing a further mechanism by which Ca 2ϩ signals can be encoded by cells.
A limitation of our study was our inability to detect cAMP rises upon absinthin addition. This may suggest that our FRET system does not have sufficient spatial resolution to detect discrete microdomains of cAMP (44), that the size of the cAMP increase is below the limit of detection, or that cAMP increases in mitochondria and not in the cytosol. This last hypothesis would be supported by a recent paper that showed that aldosterone increases selectively the activity of the soluble adenylyl cyclase in the mitochondria and that the inhibition of EPAC reduced mitochondrial Ca 2ϩ uptake, in a similar fashion to what was observed in the present contribution (45). It would be unlikely that EPAC, instead, is activated in a cAMP-independent fashion as this has never been reported previously. What links TAS2R46 to EPAC is therefore so far elusive.
It is generally accepted that high concentrations of cAMP, via PKA, are able to potentiate IP 3 -induced Ca 2ϩ release via phosphorylation of the IP 3 receptor (Ref. 46 and references therein). Our study does not contradict such notion as we have always used maximal concentrations of histamine whereas submaximal concentrations of phospholipase C-coupled receptor agonists are required to see this potentiation. Furthermore, we could not detect significant cAMP rises in the cytosol and it is therefore unlikely that a cytosolic PKA is activated in our conditions.
Recently, Dale et al. (47) have shown that in human bronchial airway smooth muscle cells, histamine-induced cytosolic Ca 2ϩ rises may be blunted by isoproterenol and proposed a model by which compartmentalized cAMP rises negatively feedback via PKA to the histamine receptor. In our study, we were unable to determine the absolute level of Ca 2ϩ released and therefore cannot rule out that IP 3 formation from histamine activation is reduced upon TAS2R46 activation. It is therefore possible that the two events occur simultaneously. Interestingly, Dale et al. (47) are able to observe an effect on histamine but not on other Ca 2ϩ -mobilizing agents, suggesting a certain degree of specificity. Preliminary data from our lab also shows that the mito-hTAS2R46 regulates mitochondrial calcium buffering chondrial uptake increase is not evident when carbachol or bradykinin are used to induce Ca 2ϩ mobilization. 5 It should be noted that we used a single Ca 2ϩ indicator for mitochondria (aequorin). Although it is unlikely that this may have resulted in artifacts, given also the consistency of results across protocols, this limitation should be acknowledged.
To describe this new Ca 2ϩ -signaling pathway, we made use of fluorescent imaging, a common technique in the field, coupled to aequorin-targeted probes and luminometry, which are less frequently employed. For these reasons, it might not be surprising that we have also unraveled a new mechanism that mediates the effect of ␤ 2 -agonists, among the most studied receptors in pharmacology.

Absinthin isolation
A voucher specimen of the Pancalieri chemotype of Artemisia absinthium is kept in Novara laboratories. 1300 g of leaves and flowers, powdered, were extracted with acetone (3 ϫ 7.5 liters) in a vertical percolator at room temperature, affording 97 g (7.5%) of a dark green syrup. The acetonic extract was dissolved into the minimal amount of acetone at 45°C and then 97 g silica gel was added (ratio extract/silica 1:1); finally the suspension was evaporated. The powder obtained in this way was stratified on a layer of 485 g of aluminum oxide (ratio extract/alumina 1:5) packed with petroleum ether (40:60) and protected on its surface by a filter paper in a sintered filtration funnel (9 ϫ 15 cm) with side arm for vacuum. The dissolved extract was purified with petroleum ether-ethyl acetate gradient from 90:10 to 60:40 (100 ml fraction) by vacuum chromatography. Fractions eluted with petroleum ether-ethyl acetate 40:60 afforded 8.7 g of pure absinthin (0.66%) whose structure elucidation and purity were confirmed from 1 H NMR (Fig. S6) according to Beauhaire et al. (48).

Lentiviral vectors (LVs) for hTAS2R46 shRNA
hTAS2R46 was silenced by lentiviral infection. Several lentiviral constructs targeting hTAS2R46 were obtained from TRC-Hs1.0 library (Dharmacon). Third-generation LVs were produced co-transfecting HEK293T packaging cells with plasmids pMDLg/pRRE, pMD2.VSVG, pRSV-Rev and transfer construct using the Lipofectamine (Life Technologies) transfection method and concentrated by PEG precipitation, as described previously (24). ASM cells were infected with the two LV-shRNAs (TCRN0000014110 and TCRN0000014112) together (referred as to shRNA-hTAS2R46) and the silencing was assessed by real time PCR and Western blotting as described in the supporting information.
Calcium imaging 5 ϫ 10 4 ASM or HeLa cells were plated on glass coverslips coated with poly-L-lysine (Sigma). The next day cells were loaded with Fura-2AM (5 M) and pluronic acid (0.005%) (all Thermo Fisher) in Vascular Cell Basal Medium for 30 min at room temperature in the dark. After washing and de-esterification (20 min) the coverslip was mounted in a chamber equipped with a thermostat and placed on the stage of a Leica epifluorescence microscope equipped with a S Fluor 40ϫ/1.3 objective. Cells were alternatively excited at 340/380 nm by the monochromator Polichrome V (Till Photonics, Munich, Germany) and the fluorescent signal was collected by a CCD camera (Hamamatsu, Japan) through band-pass 510 nm filter; the experiments were controlled and images analyzed with Meta-Fluor (Molecular Devices, Sunnyvale, CA) software. The cells were treated with the following stimuli alone or combined: histamine 10 M (Sigma-Aldrich); absinthin 10 M; KB-R7943 (Sigma-Aldrich) 10 M; FCCP (Sigma-Aldrich); 3HDC (7) 0.1, 1, and 10 M; salbutamol 10 M (Sigma-Aldrich); and forskolin 10 M (Sigma-Aldrich). To quantify the differences in the peaks of Ca 2ϩ transients the ratio values were normalized using the formula (F o Ϫ F c )/F o (referred to as normalized Fura-2 ratio; norm. ratio). To evaluate absinthin effect in calcium-free conditions, cytosolic Ca 2ϩ was monitored upon depletion of external Ca 2ϩ (by EGTA 100 M). Actin filament disrupter latrunculin A was used at 10 M (Sigma-Aldrich).

Aequorin-based Ca 2؉ measurements
We have monitored fluctuations of Ca 2ϩ concentrations in the cytosol, in the domains adjacent to the plasma membrane, and in the mitochondrial matrix. To this end, we have used third-generation LVs to infect cells with the constructs encoding the native aequorin (cytAEQ-LV) or aequorin N-terminally linked to the synaptic-associated protein 25 (SNAP-25) (pLV-pmAEQ-LV) or to the COXVIII cleavable leader sequence (mitAEQ-LV). Generation of the plasmids was reported elsewhere (24). Lentiviral particles were produced by Lipofectamine transfection of HEK293T cells and concentrated by PEG precipitation (24). To monitor Ca 2ϩ levels in different cellular compartments we used a custom built aequorinometer (CAIRN Research). ASM and HeLa cells were plated on 12 mm coverslips in a 24-well plate at a density 2 ϫ 10 4 cells/well. After 24 h, cells were infected with cytAEQ-, pmAEQ-, or mitAEQexpressing lentiviral particles. After 48 -72 h, cells were washed with Krebs-Ringer buffer (KRB; 135 mM NaCl, 5 mM KCl, 0.4 mM KH 2 PO 4 , 1 mM MgSO 4 , 5.5 mM glucose, 20 mM HEPES, pH 7.2) and reconstituted with native coelenterazine (catalog C2230, Sigma-Aldrich) at 37°C, in dark for 30 min. Then, coverslips were transferred into perfusion chamber of the aequorinometer. To assess subplasma membrane calcium, cells were firstly perfused with KRB containing 100 M EGTA for 100 s for 5 M. Talmon and L. Fresu, unpublished results. hTAS2R46 regulates mitochondrial calcium buffering baseline recording, and then the perfusion solution was switched to KRB-EGTA 100 M supplemented with 20 M tBHQ (Sigma-Aldrich). After 120 s cells were perfused with KRB-Ca 2ϩ 2 mM in presence of the stimuli of interest alone or combined: histamine 10 M, absinthin 10 M, salbutamol 10 M, KB-R7943 10 M, FCCP 10 M, forskolin 10 M. To assess calcium changes in the cytosol or mitochondria, cells were perfused with KRB containing Ca 2ϩ 2 mM for 100 s and then with KRB-Ca 2ϩ added by the already mentioned stimuli. HeLa cells were also challenged with histamine in presence of 8-pCPT-2Ј-O-Me-cAMP (10 M; Sigma-Aldrich) a cAMP analogue that specifically activates EPAC (49). For quantification of Ca 2ϩ concentration, at the end of each experiment cells were perfused with distilled water containing 0.1% Triton and 50 mM Ca 2ϩ to discharge the remaining aequorin pool. Emitted light was converted in Ca 2ϩ concentrations offline using a previously described algorithm (50). All measurements were carried out at 37°C.

Statistical analysis
Data are presented as mean Ϯ S.D. of n independent experiments. All samples were first tested for normality (Shapiro-Wilk test) and for homogeneity of variance (Levene's test). When possible, statistical significance was assessed by parametric tests, i.e. linear regression and t test or one-way analysis of variance (ANOVA) (followed by Tukey's or Dunnett's post hoc test) in the case of continuous and categorical independent variables, respectively. Otherwise, nonparametric alternatives were used, as detailed in the figure legends. Unless otherwise specified, all statistical tests were between unpaired samples, two-tailed, and a p value Ͻ 0.05 was considered statistically significant.