The Cellular and Molecular Basis of Bitter Tastant-Induced Bronchodilation

Bitter tastants can activate bitter taste receptors on constricted smooth muscle cells to inhibit L-type calcium channels and induce bronchodilation.


Introduction
Airway obstructive diseases (asthma and chronic obstructive pulmonary disease [COPD]) have become increasingly prevalent, currently affecting more than 300 million people worldwide. Dysfunction of airway smooth muscle (ASM) cells, a major cell type in the respiratory tree, plays a pivotal role in promoting progression of these diseases and in contributing to their symptoms of these diseases [1][2][3]. With their ability to contract and relax, these cells regulate the diameter and length of conducting airways, controlling dead space and resistance to airflow to and from gasexchanging areas. Their excessive contraction, as seen in patients with asthma and COPD, can fully close the airways, thereby preventing gas exchange and threatening life. Not surprisingly, bronchodilators have been used as the medication of choice for asthmatic attacks and as a standard medicine for managing COPD [4,5]. However, available bronchodilators have adverse side effects, and are not sufficiently effective for severe asthmatics and many other COPD patients. A better understanding of the mechanisms regulating ASM thus holds the promise of developing more effective and safe bronchodilators, which in turn would have a significant impact in reducing mortality and morbidity caused by asthma and COPD.
Bitter tastants represent a new class of compounds with potential as potent bronchodilators. Deshpande et al. recently found that cultured ASM cells express G-protein-coupled bitter taste receptors (TAS2Rs) [6], a class of proteins long thought to be expressed only in the specialized epithelial cells in the taste buds of the tongue that allow organisms to avoid harmful toxins and noxious substances characterized by bitterness [7][8][9][10]. Importantly, bitter tastants with diverse chemical structures cause greater ASM relaxation in vitro than b2 adrenergic agonists, the most commonly used bronchodilators to treat asthma and COPD [6,11]. Moreover, these compounds can effectively relieve in vivo asthmatic airway obstruction than b2 adrenergic agonists in a mouse model of asthma [6], making them highly attractive bronchodilators for asthma and COPD.
Bitter tastant-induced bronchodilation was unexpected, because these agents appeared to increase intracellular Ca 2+ concentration ([Ca 2+ ] i ) to a level comparable to that produced by potent bronchoconstrictors [6], which should have led to smooth muscle contraction [12]. To reconcile this apparent paradox, it was proposed that bitter tastants activate the canonical bitter taste signaling pathway (i.e., TAS2R-gustducin-phospholipase Cb [PLCb]-inositol 1,4,5-triphosphate receptor [IP3R]) to increase focal Ca 2+ release from endoplasmic reticulum, which then activate large-conductance Ca 2+ -activated K + channels thereby hyperpolarizing the membrane [6]. However, we subsequently demonstrated through patch-clamp recordings that bitter tastants do not activate large-conductance Ca 2+ -activated K + channels but rather inhibit them [11]. Moreover, three different large-conductance Ca 2+ -activated K + channel blockers did not affect the bronchodilation induced by bitter tastants [11]. Therefore, a different mechanism must be responsible for the bitter tastantinduced bronchodilation.
The apparent conundrum of putative [Ca 2+ ] i elevation leading to relaxation may be attributed to the fact that Ca 2+ responses to bitter tastants were assessed in cultured human ASM cells, while the contractile responses to them were investigated in freshly dissected ASM tissues [6]. It is well known that cultured smooth muscle cell lines alter their phenotype (i.e., losing their ability to contract and relax [13,14]) and it is likely their Ca 2+ response is also modified. Therefore, to understand bitter tastant-induced bronchodilation, it is necessary to study the contraction and the underlying signaling in freshly isolated ASM tissues and cells. Using this approach in the present study, we found that bitter tastants activate the canonical bitter taste signaling cascade, slightly increasing global [Ca 2+ ] i in resting cells, but not to a level sufficient to cause contraction. However, bitter tastants reverse the increase in [Ca 2+ ] i evoked by bronchoconstrictors, leading to bronchodilation. This reversal is mediated by the suppression of Ltype voltage-dependent Ca 2+ channels (VDCCs) in a gustducin bc subunit-dependent, yet PLCb-and IP3R-independent manner. Hence, we propose that TAS2R activation in ASM stimulates two opposing Ca 2+ signaling pathways, both mediated by Gbc subunits, which increases [Ca 2+ ] i at rest but blocks activated Ltype VDCCs reversing the contraction they cause. These results provide the cellular and molecular basis of bitter tastant-induced bronchodilation that occurs in vitro and in vivo. They further reveal a Ca 2+ signal that is well suited for screening and identifying potent bronchodilators from among the many thousands of available bitter tastants.

Results
To uncover the mechanism underlying bitter tastant-induced bronchodilation as demonstrated in both in vitro and in vivo normal and asthmatic models of mice, and in vitro human airways ( [6,11,15,16], we examined how bitter tastants affected both [Ca 2+ ] i and ASM contraction in freshly isolated airway cells and tissues from mouse and human. Fluo-3 was used to assess the effect of bitter tastants on [Ca 2+ ] i ; chloroquine and denatonium, two substances commonly used to study bitter taste signaling, were used as bitter tastants.

Bitter Tastants Modestly Raise Global [Ca 2+ ] i with No Change in Force Generation in Native ASM at Rest
We started our analysis by examining the Ca 2+ response to bitter tastants in resting cells. In contrast to the marked increase in global [Ca 2+ ] i reported in resting cultured human ASM cells [6], we observed, in resting native ASM cells from mouse, that chloroquine (0.1 mM-1 mM) only modestly raised global [Ca 2+ ] i (and to a level much lower than when cells contracted after application of Mch at 0.1 mM-100 mM) ( Figure 1A and Figure  S1A). Chloroquine (330 mM) increased fluo-3 fluorescence (DF/ F 0 ) (i.e., [Ca 2+ ] i ) both in the presence of extracellular Ca 2+ (37.7%68%, n = 19) and in its absence (29.3%66%, n = 15; p.0.05), indicating that the source for this chloroquine response is from internal Ca 2+ stores.
To examine whether this modest increase in [Ca 2+ ] i is sufficient to trigger contraction, we measured smooth muscle force formation in mouse airways. As shown in ( Figure 1B and S1B), chloroquine (10 mM-1 mM) did not cause contraction of mouse airways, although there was a tendency to decrease the basal tone of airways. As a comparison, Mch at concentrations between 0.3 mM and 10 mM induced contraction markedly and in a dosedependent manner ( Figure 1B and S1B).

Bitter Tastants Do Not Generate Localized Ca 2+ Events
Mouse ASM cells exhibit spontaneous Ca 2+ sparks resulting from the opening of ryanodine receptors in the sarcoplasmic reticulum [17]. To test whether bitter tastants generate local Ca 2+ events as proposed by Deshpande et al. [6], we stimulated ASM cells with chloroquine (10 mM, a concentration around EC50) for 2 min and measured Ca 2+ sparks. Off

Author Summary
Bitter taste receptors (TAS2Rs), a G-protein-coupled receptor family long thought to be solely expressed in taste buds on the tongue, have recently been detected in airways. Bitter substances can activate TAS2Rs in airway smooth muscle to cause greater bronchodilation than b2 adrenergic receptor agonists, the most commonly used bronchodilators. However, the mechanisms underlying this bronchodilation remain elusive. Here we show that, in resting primary airway smooth muscle cells, bitter tastants activate a TAS2R-dependent signaling pathway that results in an increase in intracellular calcium levels, albeit to a level much lower than that produced by bronchoconstrictors. In bronchoconstricted cells, however, bitter tastants reverse the bronchoconstrictor-induced increase in calcium levels, which leads to the relaxation of smooth muscle cells. We find that this reversal is due to inhibition of L-type calcium channels. Our results suggest that under normal conditions, bitter tastants can activate TAS2Rs to modestly increase calcium levels, but that when smooth muscle cells are constricted, they can block L-type calcium channels to induce bronchodilation. We postulate that this novel mechanism could operate in other extraoral cells expressing TAS2Rs.  [19,20], we studied whether bitter tastants activate this TAS2R signaling pathway. In native ASM cells, PTX (1 mg/ml, and 6-8 h pretreatment), reduced the chloroquine-induced increase in global [Ca 2+ ] i was measured with fluo-3 in the form of acetoxymethyl ester, loaded into isolated mouse ASM cells, and expressed as DF/F 0 (%). (B) 1 mM chloro did not contract airways (using tension as its proxy) while 100 mM Mch caused a robust contraction. Data are mean 6 SEM (n = 6 for chloro, and n = 5 for Mch). (C) PTX, gallein, anti-bc (MPS-phosducin-like protein C terminus, a Gbc blocking peptide), U73122, and 2-APB inhibited chloro-induced increase in [Ca 2+ ] i (n = 19-24 cells). Isolated mouse ASM cells were either pretreated with 1 mg/ml PTX for 6-8 h or with 1 mM anti-bc for 1-2 h or with each of the other compounds listed for 5-10 min. The effects of PTX and anti-bc were calculated by normalizing the response of chloro to that from the time matched cells without the pretreatments, and the effects of other three compounds were analyzed by normalizing the response of chloro to its own control without the compound. (D) RT-PCR transcripts after amplification with primers to TAS2R107, a-gustducin, Gb3, Gc13, PLCb2, and b-actin. Note that no transcript was detected with TAS2R108 primers. RNAs were isolated from mouse tracheas and mainstem bronchi, and reactions without complementary DNA were used as a negative control. (E) Cellular distribution of TAS2R107 in three focus planes (bottom, middle, and top) of an isolated mouse ASM cell. The TAS2R107 immunostaining intensity after 3D deconvolution (see Methods) was pseudocolored with the color map on the right. This makes positive (but dim) pixels more easily distinguished from background. Eight cells showed a similar subcellular distribution pattern. doi:10.1371/journal.pbio.1001501.g001 bc-Gustducin Inhibits L-Type Ca 2+ Channel PLOS Biology | www.plosbiology.org [Ca 2+ ] i to 21.1%68.6% of the control cells (n = 20; Figure 1C). Also both gallein (20 mM and 30 min pretreatment), a blocker of the Gbc dimer of PTX sensitive G proteins, and MPS-phosducinlike protein C terminus, a Gbc blocking peptide (anti-bc; 1 mM, and 1 h pretreatment) [21,22] reduced the bitter tastant-mediated increase in [Ca 2+ ] i to 19.9%68.5% (n = 19; Figure 1C) and 18.4%64.8% of the controls, respectively. Finally, U73122 (3 mM), a blocker of PLCb, and 2-aminoethoxydiphenyl borate (2-APB) (50 mM), an IP3R antagonist, suppressed the bitter tastant-induced increases in [Ca 2+ ] i to 18.0%65.5% (n = 24) and 210.5%67.3% of controls, respectively ( Figure 1C). These results indicate that bitter tastants do activate the TAS2R signaling transduction pathway (i.e., TAS2R-PTX-sensitive G protein-PLCb-IP3R) to release Ca 2+ from internal stores. This conclusion is further supported by the finding that mouse ASM cells express transcripts for TAS2R107, a-gustducin, Gb3, Gc13, and PLCb2 ( Figure 1D), and display peripheral localization of TAS2R107 ( Figure 1E).

Bitter Tastant-Induced Bronchodilation Is Due to Reversal of the Rise in Global [Ca 2+ ] i Caused by Bronchoconstrictors
Bitter tastants at mM levels can modestly increase [Ca 2+ ] i in resting cells, but this raises a conundrum as they also can fully relax airways precontracted by bronchoconstrictors [6,11]. In light of the fact that an increase in [Ca 2+ ] i is the primary signal for contraction in all smooth muscle, we explored how bitter tastants affect [Ca 2+ ] i evoked by bronchoconstrictors. To better quantify these effects, we measured ASM Ca 2+ response and cell shortening at the same time. The cells were stimulated with methacholine (Mch), a stable analogue of acetylcholine, which is the major neurotransmitter in parasympathetic nerves. As expected, Mch (100 mM) rapidly increased [Ca 2+ ] i as fluo-3 fluorescence increased by 162%626% (DF/F 0 ), and concurrently caused cell shortening by 49%68% (n = 21; Figure 2A and 2B). Strikingly, chloroquine (1 mM) almost completely reversed this [Ca 2+ ] i increase (i.e., bringing [Ca 2+ ] i down to a level only 15%62% higher than pre-stimulation levels, n = 12, p,0.01 Mch versus Mch+chloroquine). The reversal of the increase in [Ca 2+ ] i was closely associated with relaxation in ASM cells from both mouse (back to 89%67% of the pre-stimulation length; Figure 2B and Video S1) and human (back to 94%65% of the control length; Figure S2A). Denatonium (1 mM) generated similar effects on [Ca 2+ ] i and cell shortening in response to Mch in mouse ASM cells (n = 9).
The inverse relationship between changes in [Ca 2+ ] I and the resulting cell length (i.e., lowering [Ca 2+ ] i results in cell lengthening) in response to bitter tastants suggests that bitter tastants reduce [Ca 2+ ] i , leading to bronchodilation. If this is the case, one would expect that bitter tastant-induced bronchodilation could be prevented if [Ca 2+ ] i was clamped to a physiologically high level. To test this possibility, we used staphylococcal a-toxin (16,000 units/ml) to make the ASM membrane permeable to ions such that the intracellular [Ca 2+ ] i could be controlled at will. A major advantage of using this toxin is that it does not damage the cells; thus signaling processes such as the G-protein-coupled receptor mediated signaling remain intact [23]. As shown in Figure 3A, raising [Ca 2+ ] i to 3 mM caused a robust increase in tension in mouse airway. More importantly, at this fixed [Ca 2+ ] i level, denatonium, chloroquine and quinine (all at 1 mM) failed to relax ASM in the time frame they would have in Mch contracted airways without a-toxin treatment. Therefore, clamping [Ca 2+ ] i at mM levels can prevent bitter tastant-induced bronchodilation, strongly arguing that reduction of [Ca 2+ ] i by bitter tastants is necessary for their relaxation action. These results further imply that a decrease in Ca 2+ sensitivity is probably not a major mechanism underlying bitter tastant-induced bronchodilation.

The Prominent Role of L-type Ca 2+ Channels in Mediating Mch-Induced Contraction
Having established that the suppression of [Ca 2+ ] i is necessary for bitter tastant-induced bronchodilation, we next addressed the molecule(s) that bitter tastants act on to reduce [Ca 2+ ] i . Before addressing this critical question, it is appropriate to determine the Ca 2+ pathways underlying Mch-induced contraction because a controversy persists [24]. Mch activates both the M3 muscarinic acetylcholine receptor (M3R)-Gq-PLCb-IP3 pathway and the M2 muscarinic acetylcholine receptor (M2R)-Gi/o pathway to raise [Ca 2+ ] i by releasing Ca 2+ from internal stores and inducing Ca 2+ influx from the extracellular space respectively [25,26]. It has also been suggested that Ca 2+ release from the internal stores contributes to the early phase of Mch-induced contraction, and Ca 2+ influx is required to sustain elevated [Ca 2+ ] i and contraction. Indeed, we found that the sustained contraction by Mch in mouse ASM is largely dependent on Ca 2+ influx ( Figure S3). However, the route of Ca 2+ influx upon muscarinic receptor activation among different species is highly debatable [24]. Many studies suggest that L-type Ca 2+ channels are the major path of the Ca 2+ influx for contraction [27][28][29][30][31][32]. To determine the role of this channel in mediating Mch-induced contraction in mouse airways, we examined whether L-type channel specific blockers inhibit Mch-induced contraction and the rise in [Ca 2+ ] i . Previous studies showed that diltiazem, a well-known L-type Ca 2+ channel blocker that belongs to the benzothiazepine class, dose-dependently reverses Mch-induced contraction when administrated after Mch's response reaches a plateau [31][32][33]. We therefore examined whether this blocker produces similar inhibition of Mch-induced airway force generation in mouse and human airways. Figures  S2B and S4A show that diltiazem dose-dependently inhibited Mch-induced contraction of airways from both species, and, 6100 (i.e., the decrease due to chloro divided by the increase due to Mch). **p,0.01, control versus +FPL. (C) FPL dose-dependently reversed chloro-induced bronchodilation (using tension as a proxy measure) in Mch precontracted airways (n = 5-7 independent experiments). Data on the right panels are mean 6 SEM. % relaxation = tension decrease due to chloro divided by tension increase due to Mch, 6100. The tension decrease at each concentration of chloro is measured once the tension stabilizes. The tension decrease at each increased concentration is always measured relative to the peak tension (i.e., it is total decrease, not the incremental decrease due to the additional chloro which was added). doi:10.1371/journal.pbio.1001501.g003 moreover, it reversed the Mch-induced increase in [Ca 2+ ] i by 90.2%62.9% in single isolated mouse ASM cells (n = 12 cells). To further examine whether diltiazem inhibits muscarinic receptormediated contraction, we treated the airways with diltiazem at different concentrations before Mch administration. Since a unique feature of diltiazem in inhibiting L-type Ca 2+ channels is its use-dependence of action (i.e., it more likely binds to, and therefore blocks, channels as they open), in this series of experiments we challenged the airways twice with KCl to activate the Ca 2+ channels. As shown in Figure S4B, under this condition, diltiazem dose-dependently suppressed Mch-induced contraction and at 100 mM it inhibited the force by 85%66% (n = 5). To directly demonstrate that diltiazem inhibits L-type Ca 2+ channels, we studied the effect of diltiazem on L-type Ca 2+ currents with patch-clamp recording. We found that diltiazem dose-dependently reduced depolarization-induced L-type Ca 2+ currents at concentrations over the same range as that which blocked contraction ( Figure S4C).
Dihydropyridines (e.g., nisoldipine, nifedipine, and isradipine) are another well-known class of L-type Ca 2+ channel blockers [34][35][36]. Previously we found that dihydropyridines can inhibit L-type Ca 2+ current in mouse ASM [17]. To further establish the role of this class of Ca 2+ channels in mediating Mch-induced contraction, we assessed the effect of nisoldipine on the contraction evoked by Mch. It is known that dihydropyridines bind stronger to inactivated Ca 2+ channels, thus displaying a so-called voltage dependent inhibition [37]. Therefore, in this series of experiments, to facilitate the interaction between nisoldipine and L-type Ca 2+ channels, we examined the effect of nisoldipline on Mch-induced contraction in the presence of 20 mM KCl to modestly depolarize the membrane. Under this condition, 1 mM nisoldipine inhibited Mch-induced contraction by 74%68% (n = 5; Figure S4D). This result is similar to the inhibition of this blocker on carbacholinduced contraction in rat ASM, when it is similarly depolarized [29]. Hence, with specific L-type Ca 2+ channel blockers of distinct structures we have established that L-type VDCCs are the major contributor to Ca 2+ influx and sustained contraction in response to Mch in mouse airways.

Bitter Tastants Inhibit L-Type VDCCs to Decrease [Ca 2+ ] i Evoked by Bronchoconstrictors
Given the prominent role of L-type VDCCs in Mch-induced sustained contraction in mouse airways, and our and others' findings that bitter tastants reverse Mch-induced sustained contractile [6,11], we hypothesized that bitter tastants inhibit Ltype VDCCs, leading to relaxation of airways precontracted by Mch. To test this possibility, we investigated whether the L-type VDCC agonist FPL 64176 [38,39] Figure 3B). At the tissue level, FPL prevented chloroquine from relaxing Mch precontracted mouse ASM in a dose-dependent manner ( Figure 3C). These results suggest that bitter tastants inhibit L-type VDCCs, which in turn leads to a decrease in [Ca 2+ ] i and resulting bronchodilation.

Bitter Tastants Reverse the [Ca 2+ ] i Rise and Contraction Evoked by Depolarization-Induced Activation of L-Type VDCCs
To directly examine the inhibitory role of bitter tastants on Ltype VDCCs, we studied the effect of bitter tastants on L-type VDCC currents using patch clamp recording. Figure 4A shows two representative traces of L-type Ca 2+ currents evoked by a depolarizing pulse to 0 mV from a holding potential of 270 mV (left panel) and mean peak currents at different depolarizing voltages (right panel) before and after 1 mM chloroquine. As is evident, chloroquine inhibited the L-type Ca 2+ current when depolarizing voltages are between 230 mV and +40 mV.
To study whether this inhibition by chloroquine of L-type VDCCs could produce relaxation, we evaluated the effect of this bitter tastant on the depolarization-induced increase in [Ca 2+ ] i and contraction. KCl is a standard and common reagent used to study cellular processes mediated by depolarization. In airways, depolarization is expected to not only activate VDCCs in ASM but also those in cholinergic nerves as well (leading to release of Ach). Indeed, in ferret and pig, KCl activates both mechanisms to cause airway contraction [40,41]. However, the Ach release mechanism does not operate in dog and rabbit airway, as demonstrated by several reports that showed that atropine, a muscarinic receptor antagonist, had no measureable effect on the magnitude of tension generated by high K + in these species [42][43][44]. Therefore to determine which mechanisms are activated by KCl in mouse airways, we examined the influence of atropine on KCl-induced contraction. Atropine dose-dependently inhibited Mch-induced contraction of mouse airways, and at concentrations greater than 100 nM it fully blocked the contraction ( Figure S5A). These results confirm the efficacy of atropine in inhibiting muscarinic receptors in mouse airways. We further found that atropine (100 nM) reduced KCl (60 mM)-induced contraction to 58%62.5% of the time matched control ( Figure S5), implying that KCl does activate VDCCs in both ASM and cholinergic neurons to cause airway contractions via a combined effect.
Although L-type VDCCs are the major Ca 2+ channel for Ca 2+ influx upon depolarization in ASM, and mouse ASM cells exhibit only L-type Ca 2+ currents [17], it is possible that KCl-induced contraction might involve Rho and Rho kinase via a Ca 2+independent mechanism [45]. To examine this possibility, we studied the effect of extracellular Ca 2+ on the KCl-induced increase in [Ca 2+ ] i and contraction. In Ca 2+ containing medium, KCl (60 mM) induced a marked increase in [Ca 2+ ] i ( Figure S6A) in isolated ASM cells and airway contraction ( Figure S6B). Yet in Ca 2+ free medium, the same KCl failed to cause any increase in [Ca 2+ ] i or a significant contraction ( Figure S6B and S6C), consistent with published results in mouse and rat ASM [46,47]. These results indicate that KCl depolarizes the membrane, leading to a rise in [Ca 2+ ] i and a resultant contraction in mouse airway. They also demonstrate that Ca 2+ influx is necessary to produce the KCl-induced contraction, and that a Ca 2+ -independent mechanism (such as the suggested Rho and Rho kinase pathway) [45,48,49] is not sufficient (if needed at all) to produce contraction.
To further establish the role of L-type Ca 2+ channels in the KCl-induced rise in [Ca 2+ ] i and contraction, we investigated the influence of diltiazem on these two effects of KCl. We found that 100 mM diltiazem pretreatment reduced the KCl-induced increase in DF/F 0 from 122%619% to 16.8%610% in isolated ASM cells (n = 9; Figure S6C); it also reversed the KCl-induced contraction by 93.1%64.8% in airway tissue (n = 6; Figure S6D). Therefore, in mouse ASM, high KCl seems to increase [Ca 2+ ] i and cause contraction by depolarizing the membrane and activating L-type VDCCs.
Considering the action of KCl as just described ( Figures S5 and  S6), we reasoned that bitter tastants would be able to relax airways precontracted by KCl if bitter tastant's inhibition of L-type Ca 2+ channels underlies its relaxation of airways pre-contracted by Mch ( Figure 3). Indeed, we found that a 60 mM KCl-induced contraction in mouse and human airways was fully reversed by either chloroquine (1 mM) or denatonium (1 mM) (Figures 4B and  S2C). This reversal is due, at least in part, to a direct inhibition of VDCCs in ASM because (1) in the presence of 100 nM atropine, chloroquine can fully block atropine-resistant contraction ( Figure 4B), and (2) when nerve action potentials were blocked by 1 mM tetrodotoxin, a voltage-dependent Na + channel blocker, and arachidonic acid metabolism was inhibited by 1 mM indomethacin, a nonselective inhibitor of cyclooxygenase, the bitter tastants still fully relaxed airways precontracted by KCl ( Figure 4B). Similar to their effects on Mch-induced responses ( Figure 2B), chloroquine reversed the KCl-induced increase in [Ca 2+ ] i and shortening of isolated single ASM cells ( Figure 4C; n = 7). Moreover, FPL dose-dependently reversed chloroquineinduced relaxation in ASM pre-contracted by KCl (60 mM; Figure 4D), and prevented the reduction of [Ca 2+ ] i by chloroquine in cells stimulated by KCl ( Figure 4E).

Gbc Activation Mediates Bitter Tastant Suppression of the Rise in [Ca 2+ ] i Evoked by Activation of L-Type VDCCs
To address the signaling basis underlying bitter tastant inhibition of L-type VDCCs, we studied the impact of perturbing TAS2R signaling on bitter tastant-induced reversal of the [Ca 2+ ] i increase in response to KCl in isolated single ASM cells. Pretreatment with PTX at 1 mg/ml for 6-8 h prevented chloroquine-induced reversal of the KCl-induced increase in [Ca 2+ ] i , as did gallein (20 mM) and anti-bc, a Gbc blocking peptide (1 mM) ( Figure 5). However, U73122 and 2-ABP, at the concentrations that block the bitter tastant-induced increase in [Ca 2+ ] i in resting cells (Figure 1), failed to alter chloroquine's ability to reverse a KCl-induced increase in [Ca 2+ ] i ( Figure 5). These results indicate that activation of Gbc but not PLCb and IP3R is required for bitter tastant-induced inhibition of L-type VDCCs.

Discussion
Our results demonstrate that bitter tastant's reversal of the rise in [Ca 2+ ] i evoked by bronchoconstrictors is required for its bronchodilation effect. They also reveal that bitter tastants can generate different and opposing Ca 2+ signals depending upon the cellular environment. When administered alone to ASM cells at rest, bitter tastants activate the canonical TAS2R signaling pathway to modestly raise [Ca 2+ ] i ( Figure 5C) without affecting the contraction. Yet when applied in the presence of the bronchoconstrictors Mch and KCl, they inhibit L-type VDCCs, leading to a reversal of both the evoked [Ca 2+ ] i rise and the contraction ( Figure 5C). Remarkably, both types of Ca 2+ signals require Gbc, while only the increase in resting [Ca 2+ ] i depends on PLCb2 activation and IP3 generation.
Bitter taste receptors (35 in mouse and 25 in human) belong to seven transmembrane domain G-protein-coupled receptors. Long thought to only be expressed in the epithelium cells of the taste buds of the tongue, recent studies have revealed that these receptors also express in several extraoral tissues including brain, testis, immune cells, gastrointestinal tract, and respiratory system [50][51][52][53][54][55][56][57]. In airways, these receptors are found to be expressed in ciliated epithelial cells and nasal solitary chemosensory cells [51,52]. Deshpande et al. [6] reported that multiple TAS2Rs can be detected in cultured human ASM cell lines. In this study, for the first time, to our knowledge, we found that this class of Gprotein-coupled receptors is expressed in native mouse ASM cells. Specifically, we determined that TAS2R107, to which both chloroquine and denatonium are ligands, localizes in the cell periphery, a location well suited for mediating bronchodilation in response to these two bitter tastants. We cannot rule out that other types of TAS2Rs also contribute to the bronchodilation induced by chloroquine and denatonium, since both of them can activate multiple mouse and human TAS2Rs [58]. Nevertheless, these ligands do activate TAS2R signaling transduction, resulting in a bronchodilation effect, because pharmacological blocking of multiple downstream components of bitter taste receptors can prevent chloroquine and denatonium-induced cellular responses ( Figures 1C, 5A, and 5B). It is also worth noting that chloroquine, denatonium, and all the bitter tastants examined so far are not endogenous ligands for bitter taste receptors. Hence a major question remains as to whether bitter taste receptors in ASM cells have physiological ligands. Interestingly, a recent study revealed that acyl-homoserine lactones, quorum-sensing molecules for Gram-negative pathogenic bacteria, can activate bitter tastant receptors in nasal solitary chemosensory cells to evoke trigeminally mediated reflex reactions, which may trigger an epithelial inflammatory response before the bacteria reach population densities capable of forming destructive biofilms [52,53]. It would be of great interest and significance to investigate whether these quorum-sensing molecules can activate bitter taste receptors in ASM to induce bronchodilation.
This study revealed two major differences in Ca 2+ signaling compared to the study by Deshpande et al. [6]. First, these authors reported that bitter tastant increased [Ca 2+ ] i to a level comparable to bronchoconstrictors. In freshly isolated ASM, we found that bitter tastants only modestly increase [Ca 2+ ] i to a level much lower than that produced by bronchoconstrictors. Second, Deshpande et al. [6] reported that bitter tastants generate local Ca 2+ events. However, in freshly isolated ASM, we found that bitter tastants do not increase local Ca 2+ releases such as Ca 2+ puffs and Ca 2+ sparks. A reason for these two discrepancies may be that Deshpande et al.'s studies were conducted in cultured ASM cell lines; compared to freshly isolated ASM, these cells display a different phenotype by altering the expression of receptors, ion channels, and contractile proteins [13,14]. The aforementioned two differences and another difference in which we found that bitter tastants do not activate large-conductance Ca 2+ -activated K + channels strongly argue that bitter tastant-induced bronchodilation is highly unlikely to result from the generation of local Ca 2+ events, which in turn activate large-conductance Ca 2+activated K + channel and hyperpolarize the membrane as proposed [6].
Since bitter tastants relax precontracted airways [6,11,15,16], it is imperative to use a similar stimulating paradigm in order to understand the underlying mechanism of this relaxation. By simultaneously measuring [Ca 2+ ] i and cell shortening, we found that bitter tastant's ability to reverse the increase in [Ca 2+ ] i caused by bronchoconstrictors is the underlying signal producing the bronchodilation. Three lines of evidence support this conclusion. First, in the presence of bronchoconstrictors, bitter tastants lowered [Ca 2+ ] i while at the same time relaxing the precontracted cells, and this response was reversible. Second, clamping intracellular [Ca 2+ ] i to levels produced by the bronchoconstrictors (low mM) prevented bitter tastants from relaxing airways. Third, enhancing and blocking Ca 2+ influx via L-type Ca 2+ channels can oppositely regulate the relaxation mediated by bitter tastants. These results reinforce the idea that [Ca 2+ ] i is the critical signal governing ASM contractility.
The opposing Ca 2+ signals mediated by Gbc upon activation of TAS2Rs revealed in this study are unique. It is expected that gustducin Gbc activates PLCb to generate IP3 and release Ca 2+ from endo/sarcoplasmic reticulum to raise [Ca 2+ ] i in ASM cells. But, unexpectedly, gustducin Gbc also suppresses Ca 2+ signaling mediated by Mch, which largely activates M3R, a Gq family receptor. In general, Gbc from the G i /G o family (to which TAS2Rs belong) tends to potentiate, rather than, inhibit the Ca 2+ responses caused by the Gq family [59,60]. It remains to be determined whether the inhibition of Ca 2+ signaling by TAS2R activation is Gbc isoform specific. Since Gbc also mediates the ASM contractions induced by activation of M2R and caminobutyric acid-B receptors [61,62], our present findings suggested that Gbc reversal of the rise in [Ca 2+ ] i caused by bronchoconstrictors is isoform specific, and is likely via Gb3c13 dimers, which are released upon activation of TAS2Rs [63]. Further studies using ASM cells with genetic deletions of these isoforms should facilitate studying this possibility. It is worthy of mention that virtually all of the studies of bitter taste signaling in taste buds [7][8][9][10] and extraoral tissues [51][52][53] have focused on the responses mediated by bitter tastants alone; the opposing Ca 2+ signaling mediated by Gbc as revealed in the present study likely operates in these systems when they are stimulated by a combination of bitter tastants and other activators.
L-type VDCCs in smooth muscle can be modulated by a variety of means including phosphorylation and Ca 2+ [64][65][66][67][68]. Yet for the first time, to the best of our knowledge, we show that bc subunits of G-protein gustducin can inhibit these channels in smooth muscle, extending the similar findings for cloned Cav1.2 in heterologous expression cells and Cav1.1 in skeletal muscle fiber [69,70]. This interpretation is buttressed by the experiments showing that the contraction mediated by KCl-induced activation of presynaptic Ca 2+ channels can also be fully blocked by bitter tastants ( Figure 4B). What remains unknown is whether Gbc directly or indirectly inhibits these channels, and the structural basis for this inhibition. Given that Gbc can directly inhibit K + channels and N-type Ca 2+ channels in several cell types [71][72][73][74][75], it is likely that Gbc acts on L-type VDCCs in a similar manner.
Gustducin bc subunits inhibit L-type VDCCs to cause bronchodilation, highlighting the importance of these channels in mediating bronchoconstriction and their potential as a target for bronchodilators. Indeed, L-type VDCCs are expressed in ASM cells and their activation causes these cells to fully contract (Figures 4, S4, S5, S6) [17,27,76,77]. Also, activation of these channels is a major mechanism underlying bronchoconstrictorinduced contraction of different species including airway and human ( [25,27,30,32], but see [24]). Moreover, three classes of organic L-type VDCC blockers (i.e., dihydropyridines, phenylalkylamines, and benzothiazepines) are effective in relieving airway spasm in animal models of asthma and in exercise-induced asthmatic patients [78][79][80][81]. A long-standing puzzle regarding Ltype channels in ASM is that clinical trials in the 1980s suggested that antagonists for this channel were of limited use treating asthma in the population as a whole [81,82]. A potential reason for this enigma may be to some extent related to the mode of action of these blockers. It is known that these classic organic blockers exert their inactivation of L-type VDCCs in a voltage, stimulation, and frequency dependent manner [34,36,37]. Interestingly, allergen sensitized guinea-pig and rabbit ASM cells have a more hyperpolarized membrane potential than normal cells [83]. This implies that L-type Ca 2+ channel blockers (for example, dihydropyridines) would bind more weakly with these channels, thus decreasing the efficacy of these agents to inhibit these channels, should ASM cells from asthma patients have a more negative membrane potential. Bitter tastants, by their ability to inhibit Ltype Ca 2+ channels via activation of gustducin Gbc, perhaps could circumvent the drawbacks of the currently available L-type Ca 2+ channel blockers, and thus be a more effective asthma treatment. This is likely given that bitter tastants induce a stronger bronchodilation in both in vitro and in vivo asthmatic mouse models than do b2 agonists [6,11], the most commonly used bronchodilators for treating asthma and COPD.
Although bitter tastants are promising candidates to be developed as a new class of bronchodilators, and the findings in the present study provide the cellular and molecular rationale for this line of inquiry, we would caution that chloroquine and denatonium examined in this study may not be ideal candidates because of the high concentration (i.e., on the order of 100 mM) needed to fully relax precontracted ASM. This caveat, however, should not dampen enthusiasm for this endeavor as there are many thousands of bitter tastants available from plants and animals, and numerous bitter small molecules synthesized by research laboratories and a variety of companies over the years. In fact, bitter tastants can stimulate bitter taste receptors at concentrations in the nanomolar range: strychnine activates human TAS2R46 with an EC 50 of 430 nM and aristolochic acid activates human TAS2R43 with an EC50 of 81 nM [84,85]. Therefore, it is highly likely that bitter tastants with a highly potent bronchodilating action can be discovered. Searching for these bitter tastants is of clinical significance because the current bronchodilators are insufficient for treating severe asthma and many COPD patients. A critical step in identifying highly potent bitter tastants is developing reliable and highly effective screening methodologies. Simultaneous measurements of cell shortening and the [Ca 2+ ] i signal (i.e., a decrease of elevated [Ca 2+ ] i ), as developed in the present study, are robust and quantitative and provide a powerful paradigm for identifying potential bronchodilators from among the many bitter tastants available. . Ba 2+ was used as a charge carrier, and the peak current was used to construct the I-V relationship. Note that the high voltage threshold for activation seen in the I-V relationship, and its sensitivity to FPL and nifedipine [17] indicate these Ca 2+ currents resulted from the opening of L-type VDCCs. (B) Bitter tastants relaxed KCl-induced contraction of mouse airways. The left panel shows representative force recordings in response to 60 mM KCl followed by 1 mM chloro in the presence of 100 nM atropine and in its absence. The right panel shows the mean values of the relaxation of KCl-induced contraction by 1 mM chloro in the control (n = 9), in the presence of 100 nM atropine (n = 16), or in the presence of 1 mM tetrodotoxin (TTX) and 1 mM indomethacin (Indom) (n = 6). Note that the inhibition of chloro in the presence of 100 nM atropine was calculated relative to the atropine-resistant contraction in response to KCl. Not shown: 1 mM denatonium relaxed airways precontracted by 60 mM KCl by 97%64% (n = 5 independent experiments).

Animal Tissue Handling
Experimental protocols for animal research were approved by the Institutional Animal Care and Use Committees at the University of Massachusetts Medical School (protocol A-1473 to RZG).

Isolation of Mouse Airway Smooth Muscle Cells
C57BL/6 mice from 7 to 12 wk of age were anesthetized with intraperitoneally injected pentobarbitone (50 mg kg 21 ), and the trachea and mainstem bronchi were quickly removed and placed in a pre-chilled dissociation solution consisting of (in mM): 135 NaCl, 6 KCl, 5 MgCl 2 , 0.1 CaCl 2 , 0.2 EDTA, 10 HEPES, and 10 Glucose (pH 7.3). Tracheas and mainstem bronchi were dissected free from the surface of the connective tissue. The airway tissue was incubated in the dissociation medium containing papain 30 unit/ml, 1 mM DTT, and 0.5 mg/ml BSA, at 35uC for 30 min, and then transferred to a dissociation medium containing 3 unit/ ml collagenase F and 0.5 mg/ml BSA, and incubated at 35uC for another 15 min to produce isolated ASM cells. Finally, the tissue was agitated with a fire polished wide-bore glass pipette to release the cells.

Mouse Airway Smooth Muscle Contraction Bioassay
C57BL/6 mice at 7-12 wk of age were sacrificed and the entire respiratory trees were rapidly removed and immersed in Krebs physiologic solution containing (in mM) 118.07 NaCl, 4.69 KCl, 2.52 CaCl 2 ,1.16 MgSO 4 , 1.01 NaH 2 PO 4 , 25 NaHCO 3 , and 11.10 glucose. Trachea and mainstem bronchi were isolated and cut into rings (4 mm in length). The rings were mounted on a wire myograph chamber (Danish Myo Technology), and a PowerLab recording device (AD Instruments) was used to record isometric tension. The ring preparations with zero tension were immersed in 5 ml of Krebs physiologic solution, bubbled with 95% O 2 and 5% CO 2 at 37uC. After 10 min equilibration, three stretches (each 2.5 mN) at 5 min intervals were applied to the rings. After these stretches the basal tones of the rings were usually settled at approximately 2 mN. To test the contractile response, each ring was stimulated twice with KCl (60 mM), separated by 20 min, before proceeding to other treatments. The order and treatment time of agonists and antagonists are indicated in the figure captions. In the experiments in Figures 4B, S4B, and S4D, 1 mM tetrotodoxin was added to prevent action potentials of neurons and 1 mM indomethacin to inhibit cyclooxygenase. The force in response to 60 mM KCl in the presence of tetrodotoxin and indomethacin was 95%67% (n = 6) of that in their absence.

Airway Smooth Muscle Permeabilization
Mouse bronchi rings (4 mm in length) free of connective tissues were incubated for 5 min in HEPES-Tyrode (H-T) buffer which contained 137.0 mM NaCl, 2.7 mM KCl, 1.0 mM MgCl 2 , 1.8 mM CaCl 2 , 10 mM HEPES, 5.6 mM glucose (pH 7.4). The rings were then transferred to and incubated in Ca 2+ -free H-T buffer for 5 min followed by another 5 min in buffer A (30 mM TES, 0.5 mM DTT, 50 mM KCl, 5 mM K 2 EGTA, 150 mM  Figure 4E. Compared to the control (i.e., chloro alone after KCl, black filled bar), p,0.0001 for PTX, gallein, and anti-bc; and p.0.05 for U73122 and 2-APB. Data are shown as mean 6 SEM (n = 12-38 cells). (C) A model for TAS2R signaling and bitter tastant-induced bronchodila-tion. We propose that bitter tastants activate the canonical TAS2R signaling cascade; this modestly increases [Ca 2+ ] i in resting cells but exerts no significant effect on resting tone. On the other hand, activation of TAS2Rs activates gustducin and release Gbc, which turns off L-type VDCCs that are pre-activated by bronchoconstrictors, leading to bronchodilation. doi:10.1371/journal.pbio.1001501.g005 sucrose [pH 7.4]). To skin ASM, the rings were incubated for 45 min with a-toxin (16,000 units/ml) in buffer A at room temperature. After the permeabilization, the rings were treated with 10 mM ionomycin for 10 min to deplete intracellular Ca 2+ stores.
Skinned airway rings were mounted on the wire myograph chamber and washed two times with pCa 9 solution (20 mM TES, 4 mM K 2 EGTA, 5.83 mM MgCl 2 , 7.56 mM potassium propionate, 3.9 mM Na 2 ATP, 0.5 mM dithioerythritol, 16.2 mM phosphocreatine, 15 units/ml creatine kinase [pH 6.9]). The viability of the skinned muscle rings was examined by stimulation with pCa 4.5 solution (20 mM TES, 4 mM CaEGTA, 5.66 mM MgCl 2 , 7.53 mM potassium propionate, 3.9 mM Na 2 ATP, 0.5 mM dithioerythritol, 16.2 mM phosphocreatine, 15 units/ml creatine kinase [pH 6.9]) followed by the pCa 9.0 solution. The muscles that could generate sustained contraction in response to the pCa 4.5 solution and fully relax when exposed to the pCa 9.0 solution were used for subsequent experiments. To test the effect of bitter tastants, the viable muscle rings were induced to contract by exposure to pCa 5.5 solution followed by the administration of bitter tastants at the concentrations indicated in Figure 3.

Measurement of Global [Ca 2+ ] i and Ca 2+ Sparks
Fluorescence images using fluo-3 as a calcium indicator were obtained using a custom-built wide-field digital imaging system. The camera was interfaced to a custom made inverted microscope, and the cells were imaged using either a 206 Nikon 1.3 NA for global [Ca 2+ ] measurement or a 606Nikon 1.4 NA oil for Ca 2+ spark measurement. The 488 nm line of an Argon Ion laser provided fluorescence excitation, with a shutter to control exposure duration, and emission of the Ca 2+ indicator was monitored at wavelengths .500 nm. The images were acquired at the speed of either 1 Hz for global [Ca 2+ ] measurement or 50 Hz for Ca 2+ spark measurement. Subsequent image processing and analysis was performed off line using a custom-designed software package, running on a Linux/PC workstation. [Ca 2+ ] i was represented as DF/F 0 6100 with F calculated by integrating fluo-3 over entire cells for global [Ca 2+ ] after background correction with areas free of cells, or just the value at the brightest pixel (i.e., epicenter pixel) for Ca 2+ sparks.

Patch-Clamp Recording
Membrane currents were recorded with an EPC10 HEKA amplifier under perforated whole-cell patch recording configuration. The extracellular solution contained (in mM): NaCl 126, tetraethylammonium Cl 10, BaCl 2 2.2, MgCl 2 1, Hepes 10, and glucose 5.6 (pH adjusted to 7.4 with NaOH). The pipette solution contained (in mM): CsCl 139, MgCl 2 1, Hepes 10, MgATP 3, Na 2 ATP 0.5 (pH adjusted to 7.3 with KOH); amphotericin B was freshly made and added to the pipette solution at a final concentration of 200 mg/ml. Whole-cell Ba 2+ currents were evoked by step depolarization with 300 ms duration every 10 s from a holding potential of 270 mV at a 10 mV increment ( Figure 4A) or with protocol as described in the caption of figure caption ( Figure S4C). Currents were leak corrected using a P/4 protocol.

Measurement of Cell Shortening
Myocytes were placed into a recording chamber superfused with the bath solution for patch clamp experiments at room temperature. Cells loaded with Fluo-3 were imaged using a custom-built wide-field digital imaging system and their lengths were determined using custom software to manually trace down the center of the cell [17].

Reverse Transcription-PCR to Detect mRNA
The connective tissues in trachea and mainstem bronchi were carefully removed and the ASM were then quickly frozen in dry ice. The total RNA of the ASM was isolated with the TRIzol (Invitrogen) method following the manufacturer's guidelines; and cDNA was synthesized using extracted RNA with an Omniscript Reverse Transcription kit (Qiagen). The specific primers, synthesized by Invitrogen, are listed in Table S1. b-actin was used as a positive control and the absence of DNA as a negative control, and the PCR reactions were carried out in a PCR mastercycler.

Immunocytochemistry
Mouse ASM cells, isolated as described above and plated onto poly-L-lysine coated coverslips were fixed and permeabilized (0.1 M ethanolamine in PBS plus 0.1% triton X-100 [pH 8]) and then immunolabeled as described previously [86]. Anti-TAS2R107, an affinity purified rabbit polyclonal antibody raised against a peptide mapping within an extracellular domain of mouse TAS2R107, was purchased from Santa Cruz Biotechnology (sc-139175), and purified IgG was used as control.
3D fluorescence imaging was performed on an inverted wide field microscope (Nikon Diaphot 200) with excitation by a 100 W mercury lamp. Images were obtained through a 606objective and digitally recorded on a cooled, back-thinned CDD camera (Photometrics), with an effective pixel size at the specimen of 83 nm in x-y and a z spacing of 100 nm. This resulted in a 3D stack of approximately 100 image planes for each cell.
The fluorescence images were deconvolved with a constrained, iterative approach [87] originally designed for UNIX systems. The algorithm was rewritten using FFTW, a free, fast Fourier transform library and implemented as a multiuser client/server system on computers running the Fedora operating system (Red Hat), either stand-alone or configured in a Beowulf cluster. Each image was dark current and background subtracted, flat-field corrected, and then deconvolved. After deconvolution images were thresholded to eliminate non-specific binding. Voxels that fell below a threshold were considered to be non-specific bindings and were set to zero; all other voxels remained unchanged. This threshold was derived from analysis of control images containing purified IgG. The intensity which eliminated 99% of the voxels in the control images became the threshold intensity.

Reagents and Their Application
All chemicals, except fluo-3 (Invitrogen Co), gallein (Tocris Bioscience), anti-bc blocking peptide (AnaSpec), anti-TAS2R107 (Santa Cruz Biotechnology), and purified IgG (Jackson Immu-noResearch Laboratories) were purchased from Sigma-Aldrich Co. For single cell studies, agonists and antagonists were applied locally to cells via a picospritzer at a constant pressure, so that the duration of its action and concentration could be controlled easily.

Statistics
Unless stated otherwise, data are reported as mean 6 standard error of the mean (SEM) and n represents the number of cells or trachea and mainstem bronchi. Statistical analysis of differences was made with Student's paired or unpaired t-test and the significance level was set at p,0.05. . % relaxation = tension decrease due to diltiazem divided by tension increase due to Mch, times 100. The tension decrease at each concentration of diltiazem is measured once the tension stabilizes. The tension decrease at each increased concentration is always measured relative to the peak tension (i.e., it is total decrease, not the incremental decrease due to the additional diltiazem which was added). (C) Chloroquine (1 mM) and diltiazem (100 mM) relaxed human intrapulmonary bronchi precontracted by 60 mM KCl (n = 3-5 independent experiments). % relaxation = tension decrease due to chloroquine divided by tension increase due to Mch, times 100. Bar charts are mean 6 SEM. Human lung tissues were obtained (with informed consent) from patients undergoing surgery (lobectomy) for lung cancer at the Department of Surgery and the Department of Pathology at the University of Massachusetts Memorial Medical Center (Worcester). The tumors were identified as non-small cell carcinoma (adenocarcinoma or squamous cell carcinoma). Intrapulmonary airways were dissected out and cleaned free of the connective tissues. These airways were either cut into the rings (4 mM in length) for force measurements the same as for mouse airway tissues, or digested with the same enzymes, dissociation medium, and isolation procedure as for single mouse ASM cells.  Movie S1 This clip shows 1 mM chloro reversed the 100 mM Mch-induced increase in [Ca 2+ ] i and cell shortening; the first 120 images of this clip were analyzed and plotted in Figure 2A. The images are displayed as fluorescence intensity (rather than DF/F 0 ) because the cell changes its shape dramatically in response to stimuli (and changing thickness makes DF/F 0 measures misleading). (MOV)