G-Protein-Dependent and -Independent Pathways

in Denatonium Signal Transduction

Shoko SAWANO,1 Eri SETO,1 Tomohiko MORI,2 and Yukako HAYASHI1;y

1Graduate School of Agriculture, Kyoto University, Uji, Kyoto 611-0011, Japan2Course of Health and Nutrition, Kio University, Nara 635-0832, Japan

Received November 15, 2004; Accepted May 17, 2005

To clarify the involvement of G protein in denato-

nium signal transduction, we carried out a whole-cell

patch-clamp analysis with isolated taste cells in mice.

Two different responses were observed by applying

GDP-�-S, a G-protein inhibitor. One response to

denatonium was reduced by GDP-�-S (G-protein-de-

pendent), whereas the other was not affected (G-

protein-independent). These different patterns were also

observed by concurrently inhibiting the phospho-

lipase C �2 and phosphodiesterase pathways via G

protein. These data suggest dual, G-protein-dependent

and -independent mechanisms for denatonium. More-

over, the denatonium responses were not attenuated by

singly inhibiting the phospholipase C �2 or phospho-

diesterase pathway, implying that both pathways were

involved in G-protein-dependent transduction. In the G-

protein-independent cells, the response was abolished by

the depletion of calcium ions within the intracellular

store. These results suggest that Ca2þ release from the

intracellular store is an important factor. Our data

demonstrate multiple transduction pathways for dena-

tonium in mammalian taste cells.

Key words: bitter; mouse; patch-clamp; taste; denatoni-

um signal transduction

A bitter taste is unpleasant, although bearable whenweak, but repulsive when strong. A variety of chemicalcompounds induce bitter taste responses via differentmechanisms. Bitter taste effectively warns us not toingest potentially harmful compounds, and studies onthe bitter transduction mechanisms have been re-viewed.1)

Denatonium, a synthetic bitter substance, is widelyused in bitter signal transduction. Many studies haveimplicated second-messenger-mediated mechanisms in-volving the activation of G protein in denatonium signaltransduction. One possible mechanism involves theactivation of phospholipase C �2 (PLC �2) via G-

protein-coupled receptors.2–8) Activated PLC �2 catal-yses an increase in diacylglycerol (DAG) and IP3, whichactivates IP3R in the endoplasmic reticulum, resulting inthe release of Ca2þ pools stored in this compartment.This Ca2þ from the internal store may lead to theactivation of the transient receptor potential channel,TRPM5, or may suffice to cause neurotransmitterrelease. Another possible mechanism is the activationof PDE via its receptor and G protein subunit �-gustducin, a taste-enriched G-protein.9–11) The activationof PDE decreases the level of the intracellular cyclicnucleotide (cNMP) that may lead to opening/closing ofthe cNMP-gated (CNG) channels in taste cells.12,13) Bothmechanisms generally rely on signaling cascades ini-tiated by G-protein-coupled receptors.1,14–17)

However, aside from these G-protein-mediated path-ways, there is the opinion that several bitter compoundsdo not require G-protein-coupled receptors for trans-duction. For instance, quinine and CaCl2 have directlyblocked voltage-gated K channels in mudpuppy tastecells,18,19) leading to membrane depolarization andtransmitter release. Membrane permeant bitter com-pounds such as caffeine and theophylline increaseintracellular cGMP by directly inhibiting PDE.20) Caf-feine has been proposed to use the G-protein-independ-ent mechanism. Caffeine signal transduction may be notmediated through gustducin, because gustducin knock-out mice were normal in their caffeine-tasting ability,and caffeine failed to activate transducin or gustducin inan in vitro assay.11) It is known that caffeine is anagonist of the ryanodine-sensitive calcium store and maydirectly induce Ca2þ release. The response to dextro-methorphan (DXM), a bitter antitussive, did not involveG proteins and PLC in mudpuppy taste cells and hasbeen suggested to enter taste cells and interact directlywith the Ca2þ store.21) These G-protein-independentresults were also different from the results that TRPM5and PLC�2 null mice could not recognize quininehydrochloride or denatonium.22) However, Tsunenari

y To whom correspondence should be addressed. Fax: +81-774-38-3749; E-mail: yhayashi@kais.kyoto-u.ac.jp

Abbreviations: PLC �2, phospholipase C �2; PDE, phosphodiesterase; IP3, inositol 1,4,5-triphosphate; IP3R3, IP3 receptor type III; cNMP, cyclic

nucleotide; CNG channels, cNMP-gated channels; DAG, diacylglycerol; DXM, dextromethorphan; 8-Br-cGMP, 8-bromo-guanosine-30,50-cyclic-

monophosphate; IBMX, 3-isobutyl-1-methylxanthine

Biosci. Biotechnol. Biochem., 69 (9), 1643–1651, 2005

et al. have suggested that quinine hydrochloride directlyaffected the cationic channel independently of secondmessenger systems.23) The choice of strategy adopted forthe experiments (in vivo/vitro assays) might havecaused some difference in the results. Caicedo et al.24)

have shown that the incidence of denatonium-responsivecells was significantly reduced in �-gustducin-null mice.Nevertheless, some taste cells lacking �-gustducin stillresponded to denatonium, suggesting the involvement ofan other G protein � subunit, G�i2, which was expressedin more taste cells than �-gustducin-positive taste cells.Bitter taste receptor candidates, T2Rs, could couple withG�i2 other than �-gustducin.25) However, T2Rs were co-expressed with a subset of �-gustducin in taste cells.5)

These data indicate that there are no T2Rs in �-gustducin- or G�i2-negative cells. These findings sug-gest that there is non-�-gustducin, non-T2R-mediatedtransduction in bitter taste. G-protein-independent signaltransduction has also been reported for denatonium.Denatonium may directly block the delayed-rectifier Kþ

channels in mammalian taste cells.26) Another study hasshown that �-gustducin-null mice reduced the aversionto denatonium and reduced gustatory nerve responses,whereas they could recognize a high concentration(more than a few millimolar) of denatonium.10) Gustdu-cin- and transducin-double-null mice were also able torespond to denatonium at a high concentration.27) Thesedata imply the existence of G-protein-independentdenatonium transduction mechanisms. Furthermore,our previous studies on mouse taste cells have shownthat denatonium induced outward current responses,including either an increase or decrease in the membraneconductance, while reversal of the potential did notchange during the stimulation.28) Our data suggest thatthere are two types of mechanism in the signal trans-duction of denatonium in taste cells. The existence ofmultiple denatonium transduction mechanisms is there-fore implied in not nonspecific cells but taste cells.In the present study, we conducted an electrophysio-

logical analysis to investigate the possibility of Gprotein-independent denatonium signal transduction inmouse taste cells. Our results indicate the presence ofdual, G-protein-dependent and -independent mecha-nisms for denatonium in mouse taste cells.

Materials and Methods

Reagents. Collagenase and elastase were purchasedfrom Worthington Biochemical Corporation (Lakewood,NJ, U.S.A.). Amiloride, DNase, denatonium benzoate,8-bromo-guanosine-30,50-cyclic-monophosphate (8-Br-cGMP), 3-isobutyl-1-methylxanthine (IBMX), and thap-sigargin were purchased from Sigma Chemical Co. (St.Louis, MO, U.S.A.). U73122 was purchased fromNacalai Corporation (Kyoto, Japan). All other reagentswere of analytical grade and obtained from standardsuppliers.

Solutions. Two solutions were used for cell prepara-tion. The normal Ringer solution contained (in mM) 140NaCl, 5 KCl, 1 CaCl2, 10 MgCl2, 10 glucose, 10 pyruvateNa, 0.014 phenol red, and 10 Na-HEPES at pH 7.4. Thedivalent-cation-free Ringer solution contained (in mM) 65NaCl, 20 KCl, 20 glucose, 26 NaHCO3, 2.5 NaH2PO4,0.014 phenol red, and 1 EDTA-4Na.

In respect of the basic patch pipette solutions and theextracellular solution, either 0.5mM or 1.0mM GDP-�-S,1.0mM U73122, 1.0mM 8-Br-cGMP, or 1.0mM IBMXwas added to the pipette solution as appropriate. Similarly,1.0mM 8-Br-cGMP, 10mMU73122 or 2.0mM thapsigarginwas added to the bath solution as appropriate.

Pipette solution: The solution used to fill the Kþ

pipette contained (in mM) 140 KCl, 1 CaCl2, 2 MgCl2, 5ATP-Na2, 11 EGTA, and 10 K-HEPES at pH 7.2. Thesolution used to fill the Csþ pipette contained (in mM)140 CsCl, 1 CaCl2, 2 MgCl2, 5 ATP-Na2, 11 EGTA, and10 NMDG-HEPES at pH 7.2.

Bath solution: The extracellular solution contained (inmM) 140 NaCl, 5 KCl, 1 CaCl2, 10 MgCl2, 0.014 phenolred, and 10 Na-HEPES at pH 7.4.

Preparation of isolated taste cells. Taste cells wereisolated from 8-10-week-old female C57BL/6J Jms Slcmice. The mice were sacrificed by cervical dislocation.Their tongues were immediately removed and washedwith the divalent-cation-free Ringer solution that hadbeen equilibrated with 95% O2–5% CO2. The normalRinger solution containing 2.5mg/ml of collagenase,2.0mg/ml of elastase, 1.0mg/ml of trypsin inhibitor,14 mM DNase, and 10 mM amiloride was injected into thetongue from the cut end with a fine needle. The tonguewas incubated for 10min at 31.5 �C in the divalent-cation-free Ringer solution equilibrated with 95% O2–5% CO2, and the tongue epithelium was peeled off byusing tweezers. The epithelium was rinsed with thedivalent-cation-free Ringer solution, and fixed upside-down with pins in the normal Ringer solution. Thisepithelium was stored at 4 �C until needed.

The isolated taste cells were sucked from the circum-vallate and foliate papillae with a glass capillary under adissecting microscope, and placed on to a patch-clamprecording dish coated with concanavalin A. The glasscapillary was made from a borosilicate capillary(GC150T-10; Clark Electromedical Instruments, UK)coupled with a patch electrode puller (P-97; ShutterInstruments Co., U.S.A.). The tip of this capillary wasfire-polished with a microforge (MF-83; Narishige,Japan), and its opening diameter was maintained ataround 20 mm.

Electrophysiological recording. The whole-cell modeof the patch-clamp was used for electrical recording ofthe taste response. A dish containing isolated cells wasplaced on the stage of an inverted microscope providedwith Hoffman modulation contrast. Patch pipettes weremade from the borosilicate capillary (GC150F-10; Clark

1644 S. SAWANO et al.

Electromedical Instruments, UK) and fire-polished to afinal resistance of 6–10M� with a microforge. Thepipette was connected to a patch-clamp amplifier andoperated with a 3-D micromanipulator. The signal waslow-pass-filtered at 1 kHz and sampled at 0.05–10 kHzby a 12-bit analogue-to-digital converter connected to apersonal computer controlled by laboratory-made pro-gram. The command voltage was generated by a 12-bitdigital-to-analogue converter with a time resolution of0.1ms, the command voltage resolution at the mem-brane being 0.098mV. Series resistance was notcompensated, and the input resistance was checked tobe >1:0G� in the tested cells. An Ag–AgCl indifferentelectrode was connected to the bath solution via an agarbridge containing the Ringer solution from which CaCl2and MgCl2 had been removed. All experiments wereperformed at room temperature (20–26 �C).

Taste stimulation. Denatonium benzoate (1mM) wasdissolved in the extracellular solution to fill the stimuluspipette (3 mm opening). The pipette for stimulation was

connected to a microinjector (IM-200; Narishige, Japan)and positioned within about 25–30 mm of the taste cellby a micromanipulator (WR-90; Narishige, Japan). Afterthe whole-cell configuration had been established, theexcitability of the cell was confirmed. Denatoniumstimuli were then applied to the taste cell by pressureejection (20 kPa) for 30 s. The response of the taste cellwas measured in the voltage-clamp mode (holdingpotential at �80mV).The cells were superfused at a rate of 1ml/min with

the extracellular solution. Kþ and Csþ internal solutionswere used to fill the pipette as appropriate. A Csþ pipettewas used to eliminate the voltage-dependent Kþ current.The characteristics of the patched cells were checked

before recording the denatonium-induced current. Cellsisolated from the epithelium of the tongue often includeboth taste cells and non-taste cells. We selected spindle-shaped cells as the taste cells (Fig. 1A). Moreover, weassumed that they had a voltage-dependent fast transientinward current of Naþ ions and slow outward current ofKþ ions (Figs. 1B and C), since mouse taste cells are

Fig. 1. Morphological and Electrophysiological Characteristics of an Isolated Mouse Taste Cell.

A, Photograph of the whole-cell patch clamp configuration. Spindle-shaped cells were morphologically selected as taste cells for the

experiments. B, Voltage-dependent Naþ (transient inward) and Kþ (transient and sustained outward) currents at each command voltage.

Electrical excitability was confirmed by depolarizing the voltage after the whole-cell configuration had been established. Voltage commands

were from �70mV to 50mV in 10mV steps. C, Current-voltage relationship of voltage-dependent Naþ and Kþ currents. Peaks of the two

currents are plotted against the command potentials.

Denatonium Transduction Pathways in Mouse Taste Cells 1645

excitable. To distinguish the taste cells from nonsensoryepithelial cells, their electrical excitability was con-firmed by depolarizing voltage pulses after the whole-cell configuration had been established. The excitablecells, which had a voltage-dependent channel, wereregarded as healthy taste cells and used for the experi-ments. During recording, the resting potential was stableat around �40mV, and a depolarization step confirmedthat the tested taste cells had excitability.

Results

Dual, G-protein-dependent and -independent re-sponses for denatoniumTo investigate the involvement of G protein in

denatonium signal transduction, the effect of GDP-�-S(a non-specific G-protein inactivator) was analyzed byadding it to the patch pipette solution.

Immediately after establishing the whole-cell config-uration, denatonium stimulation was applied to a tastecell. Most of tested cells induced an outward current.This was regarded as the inherent response to denato-nium before GDP-�-S had diffused into the cells. After10–15 minutes, when GDP-�-S had completely diffusedthroughout the taste cells, the cells were stimulatedagain with denatonium. Interestingly, the GDP-�-Streatment showed two different response patterns. Insome cells, the response was significantly inhibited byGDP-�-S (n ¼ 12 of 35 cells tested; p < 0:001, Figs. 2Aand 2C), whereas other cells were not affected by GDP-

Fig. 2. Two Different Responses Observed after GDP-�-S Application to MouseTaste Cells.

Current responses to 1.0mM denatonium were recorded at �80mV before (left panels) and after (right panels) the diffusion of GDP-�-S, a

non-selective G-protein inhibitor. Solid lines denote the period of denatonium stimulation for 30 s. A, G-protein-dependent responses (0.5mM

GDP-�-S, n ¼ 3 of 6; 1.0mM GDP-�-S, n ¼ 9 of 29). B, G-protein-independent responses (0.5mM GDP-�-S, n ¼ 3 of 6; 1.0mM GDP-�-S,n ¼ 20 of 29). C, Pharmacological effects of GDP-�-S on the denatonium-induced current. The current amplitude after GDP-�-S application has

been normalized by that of before the application. Each relative ratio is presented as the mean� SEM. The responses shown in (A) were

significantly suppressed (unfilled bar, ���p < 0:001, paired t-test), but the responses shown in (B) were not significantly suppressed (filled bar,

p ¼ 0:526, paired t-test). The relative ratio of the responses shown in (A) was significantly lower than that in (B) (##p < 0:0001, unpaired t-test).

1646 S. SAWANO et al.

�-S (n ¼ 23 of 35; Figs. 2B and 2C). This unaffectedoutward response to denatonium stimulation was ob-served even after 40 minutes when GDP-�-S hadcompletely diffused throughout the cell. A Statisticalanalysis showed that the two types of cell showedthe different responses to denatonium (Fig. 2C, p <0:0001).

G-Protein-independent denatonium signal transduc-tion requires intracellular Ca2þ release

We studied the mechanism for the G protein-inde-pendent denatonium-induced response in more detail. Ithas been reported that several bitter tastants directlyinteracted with the Ca2þ store21) or directly inhibitedPDE.20) Therefore, the involvement of the intracellularCa2þ store and PDE was examined in the G-protein-independent denatonium pathway.

We first investigated the role of the intracellular Ca2þ

store in the G-protein-independent current. After con-firmating a G-protein-independent response in the tastecells, the extracellular solution was replaced with thealternate extracellular solution containing 2.0 mM thap-sigargin to deplete the intracellular Ca2þ store. Thedenatonium-induced outward response was abolished bythapsigargin (n ¼ 4; Figs. 3A and 3D, p < 0:002). Sincethere was a possibility that Ca2þ release from the Ca2þ

store had been caused by PLC �2 activation, the PLC-mediated pathway was blocked by U73122. U73122(10 mM, a PLC inhibitor) at a dose sufficient to affectPLC8,21,29) had no effect on the G-protein-independentdenatonium response (n ¼ 5; Figs. 3B and 3D).We also investigated the involvement of the PDE-

mediated pathway in the G-protein-independent denato-nium current. Applying 8-Br-cGMP (1.0mM; a mem-

Fig. 3. Effects of Thapsigargin, 8-Br-cGMP, and U73122 on the G-Protein-Independent Denatonium Current.

These experiments were carried out on the GDP-�-S-insensitive cells. The bathing solution was replaced with the extracellular solution

containing 2 mM thapsigargin (A), 10 mM U73122 (B), or 1mM 8-Br-cGMP (C). A, The G-protein-independent current was abolished by

thapsigargin (n ¼ 4). B and C, The G-protein-indepedent current was not attenuated by 8-Br-cGMP or U73122 (n ¼ 3, n ¼ 5, respectively). D,

Pharmacological effects of 8-Br-cGMP, U73122, and thapsigargin on the denatonium-induced current. The responses shown in (A) were

significantly suppressed by thapsigargin (��p < 0:002, paired t-test).

Denatonium Transduction Pathways in Mouse Taste Cells 1647

brane-permeating cGMP analogue) did not affect the G-protein-independent outward current (n ¼ 3; Figs. 3Cand 3D).

G-Protein-dependent denatonium signal transductionneeds both PLC and PDE activationsSince second-messenger-mediated pathways such as

PDE and PLC are induced by the activation of G protein,the denatonium-induced responses were recorded whenthe PDE and PLC pathways were blocked in order toelucidate G-protein-dependent denatonium signal trans-duction.

To block both the PDE and PLC pathways, 8-Br-cGMP (1.0mM) and U73122 (1.0mM) were added to theKþ pipette solution. Interestingly, similar to the appli-cation of GDP-�-S, the co-application of 8-Br-cGMPand U73122 showed two different response patterns. Insome cells, the response was significantly inhibited bythese modulators (n ¼ 6 of 10; Figs. 4A-1 and 4D,p < 0:001), whereas other cells were not affected (n ¼ 4

of 10, Figs. 4A-2 and 4D). The inhibition ratios weresignificantly different among these cells (Fig. 4D,p < 0:0001).

We examined the independent effect of the PLC and

Fig. 4. Effects of Second Messengers on the Denatonium-Induced Response in Taste Cells.

The taste cell response to denatonium was recorded with a Kþ pipette containing (A) 1mM 8-Br-cGMP and 1mM U73122, (B) 1mM 8-Br-

cGMP only, or (C) 1mM U73122 only. All the left-hand panels show the denatonium-induced response before diffusion of the additives into the

taste cells, while the right-hand panels show the response after diffusion. Solid lines denote the period of denatonium stimulation for 30 s. A, Two

types of response were observed after combined 8-Br-cGMP and U73122 application. Some cells were significantly suppressed by these agents

(A-1; n ¼ 6=10Þ, whereas others were not suppressed (A-2; n ¼ 4=10Þ. B, 8-Br-cGMP did not affect the denatonium-induced current (n ¼ 3). C,

U73122 did not affect the denatonium-induced current (n ¼ 3). D, Pharmacological effects of 8-Br-cGMP and/or U73122 on the denatonium-

induced current. The responses shown in (A-1) were significantly suppressed by 8-Br-cGMP and U73122 (shaded bar, ���p < 0:001, paired t-

test), and the responses in (A-2) were not (filled bar, p ¼ 0:203, paired t-test). These two responses shown in (A) were significantly different

(##p < 0:0001, unpaired t-test).

1648 S. SAWANO et al.

PDE pathways on the respective G-protein-dependentdenatonium responses. The denatonium outward re-sponse was not impaired by 1.0mM 8-Br-cGMP (n ¼ 3;Figs. 4B and 4D). We also examined the effect ofIBMX, as a direct inhibitor of PDE, on the denatoniumresponse. The addition of 1.0mM IBMX to the Csþ

pipette solution had no effect on the current response(n ¼ 5, data not shown).

Finally, U73122 (1.0mM) was added to the Kþ pipettesolution to test whether a PLC- and IP3-related mech-anism was involved in denatonium signal transduction.Even with such an excessive concentration of U73122(1.0mM), denatonium repeatedly induced an outwardcurrent, which wasn’t impaired even 10–20min after thediffusion of U73122 (n ¼ 3; Figs. 4C and 4D).

Discussion

We found in the present study two types of denato-nium-induced response by blocking the activation of Gprotein. One response was abolished by GDP-�-S (G-protein-dependent), while the other was not diminishedby GDP-�-S (G-protein-independent). Similarly, twotypes of response were also obtained by co-blocking thesecond-messenger-related PDE and PLC pathways via Gprotein. These findings suggest that there are dual, G-protein-dependent and -independent mechanisms in theresponse to denatonium in mammalian taste cells.

Wong et al.10) have shown that, in behavioral experi-ments, while a low concentration of denatonium was notrecognized, a high concentration of more than a fewmillimolar was recognized by �-gustducin-null mice.Even gustducin- and transducin-double-knockout micewere able to respond to a high concentration ofdenatonium.23) Previous studies have suggested thepossibility of backup systems through other G pro-teins.30) In this study, however, the denatonium responsewas not abolished by even a non-specific G-proteininactivator, suggesting a possible novel G-protein-independent pathway in denatonium signal transduction.

We have subsequently examined the G-protein-de-pendent and -independent mechanisms in detail. In G-protein-independent cells, the depletion of calcium ionsin the intracellular Ca2þ store diminished the denato-nium response, while treatment with U73122 or 8-Br-cGMP had no effect on the denatonium-induced re-sponse. These results suggest that denatonium induceddirect Ca2þ release from the intracellular Ca2þ store inG-protein-independent transduction.

In our study, upon the application of 8-Br-cGMP, thelevel of intracellular cNMP was kept high. An excessivelevel of cNMP blocks transmission through the PDEpathway and leads to desensitization of the CNGchannels. As a result, 8-Br-cGMP blocks not onlytransmission through the PDE pathway, but also theopening/closing of the CNG channel. The sensitivity tocGMP has been found to be about 100 times greater thanthat to cAMP in the CNG channels expressed in taste

cells,13) and moreover, Rosenzweig et al.20) havesuggested that bitter taste signal transduction may bemediated by cGMP through the direct inhibition of PDE,independent of a G-protein-mediated pathway. Hence,we selected 8-Br-cGMP as the modulator in this study.Subsequently, in order to elucidate the G-protein-

dependent and -independent mechanisms, the blockingeffect of the PDE and PLC pathways on the denatoniumresponse was examined. Interestingly, when the PDEand PLC pathways were concurrently blocked, twopatterns of denatonium response were observed, similarto the case of GDP-�-S; a diminished pattern meant G-protein-dependent, and an unaffected pattern meant G-protein-independent. On the other hand, there was noeffect on the denatonium-induced response when thePDE or PLC pathway was solely blocked. Other studieshave also suggested that both a transient decrease incNMP and increase in IP3 in mice were needed in G-protein-dependent denatonium transduction.31,32) To-gether with our data, both the PDE and PLC pathwaysmay play an important role in G-protein-dependentdenatonium transduction.The experimental results show that the percentage of

the U73122- and 8-Br-cGMP-unaffected patterns waslower than the GDP-�-S-unaffected pattern. There mayalso be other pathways: G protein-dependent, G protein-and PDE/PLC-independent, and G protein-independentbut PDE/PLC-dependent. Another possibility is that theeffect of the mixture of U73122 and 8-bromo-cGMPmay occur more broadly than the effect of GDP-�-s dueto an adverse drug action. Recent studies have indicatedthat the action of U73122 may not be specific, asoriginally thought, since the agent also inhibits phos-pholipase D activity,33) and since U73122 elicits intra-cellular calcium release.34) Furthermore, the inhibitionby U73122 of the KACh channels occurs downstream ofthe G�� or Naþ channel activation.35) IntracellularcGMP has been demonstrated to modulate the activitiesof ion channels in various species and tissue types, asreviewed by Lucas et al.36) Intracellular cGMP has beenfound to either stimulate or depress the L-type Ca2þ

current not only by altering the cAMP concentrationthrough the regulation of PDEs, but also by activatingcGMP-dependent protein kinase.37) cGMP also regulatesthe activating component of the delayed rectifier Kþ

current.38) Activation of the delayed rectifier Kþ currentplays an important role in initiating the depolarizationprocess of the action potential. In this present study,however, there is no evidence to eliminate the possibilityof these adverse effects coupling to suppress theresponse.Taste cells are electrically excitable and can synapse

with nerve fibers. The cells in taste buds are classed asproliferative basal and three mature types: type I cells,type II cells and type III cells.39,40) The physiologicaland biochemical differences of the cell types are poorlyunderstood. Currently, type II cells expressing gustducinare known to be associated with the transducin of bitter,

Denatonium Transduction Pathways in Mouse Taste Cells 1649

sweet and umami taste,41,42) although how these cellscommunicate with the nervous system is unknown.Type III cells synapse with afferent nerve fibers,40,43,44)

but whether these cells can transduce taste stimuli or notis unclear. Electrophysiologically, types I, II and III cellshad voltage-gated inward Naþ and outward Kþ currents,and a subset of blood antigen A-IR cells (type II) andNCAM-IR cells (type III) had voltage-gated Ca2þ

currents, while gustducin-expressing cells (type II)lacked the voltage-gated Ca2þ current.45) In our study,the cells were confirmed to have a voltage-dependentfast transient inward current of Naþ ions and slowoutward current of Kþ ions before denatonium applica-tion. However, the existence of a voltage dependentCa2þ current was not examined. Different transductionmechanisms for denatonium might be generated indifferent types of cell. Additional studies are required toclarify the relationship between the mechanism fordenatonium response and the cell type.In conclusion, we identified not only the G-protein-

dependent pathway, but also the G-protein-independentpathway to be involved in denatonium signal trans-duction. It would make sense for animals to havemultiple mechanisms for denatonium signal transductionas a means of survival. It is of benefit for animals tohave as many transduction mechanisms as possibleagainst a specific and potentially harmful substance. Ourfindings may reflect the complexity of these mecha-nisms.

Acknowledgments

This research was supported in part by grants inAid for Scientific Research (C)(2)(13660123) and(B)(2)(16380088) from the Japan Society for thePromotion of Science (JSPS). We thank Dr. Andrew I.Spielman of New York University for helpful commentson the manuscript.

References

1) Lindemann, B., Receptors and transduction in taste.Nature, 413, 219–225 (2001).

2) Nakashima, K., and Ninomiya, Y., Increase in inositol1,4,5-triphosphate levels of the fungiform papilla inresponse to saccharin and bitter substances in mice. Cell.Physiol. Biochem., 8, 224–230 (1998).

3) Clapp, T. R., Stone, L. M., Margolskee, R. F., andKinnamon, S. C., Immunocytochemical evidence forco-expression of type III IP3 receptor with signalingcomponents of bitter taste transduction. B.M.C. Neuro-sci., 2, 6 (2001).

4) Ogura, T., Margolskee, R. F., and Kinnamon, S. C.,Taste receptor cell responses to the bitter stimulusdenatonium involve Ca2þ influx via store-operatedchannels. J. Neurophysiol., 87, 3152–3155 (2002).

5) Asano-Miyoshi, M., Abe, K., and Emori, Y., Co-expression of calcium signaling components in verte-brate taste bud cells. Neurosci. Lett., 283, 61–64 (2000).

6) Hwang, P. M., Verma, A., Bredt, D. S., and Snyder, S.H., Localization of phosphatidylinositol signaling com-ponents in rat taste cells: Role in bitter taste transduction.Proc. Natl. Acad. Sci. U.S.A., 87, 7395–7399 (1990).

7) Spielman, A. I., Huque, T., Nagai, H., Whitney, G., andBrand, J. G., Generation of inositol phosphates in bittertaste transduction. Physiol. Behav., 56, 1149–1155(1994).

8) Ogura, T., Mackay-Sim, A., and Kinnamon, S. C., Bittertaste transduction of denatonium in the mudpuppyNecturus maculosus. J. Neurosci., 17, 3580–3587(1997).

9) Ruiz-Avila, L., McLaughlin, S. K., Wildman, D.,McKinnon, P. J., Robichon, A., Spickofsky, N., andMargolskee, R. F., Coupling of bitter receptor tophosphodiesterase through transducin in taste receptorcells. Nature, 376, 80–85 (1995).

10) Wong, G. T., Gannon, K. S., and Margolskee, R. F.,Transduction of bitter and sweet taste by gustducin.Nature, 381, 796–800 (1996).

11) Ming, D., Ruiz-Avila, L., and Margolskee, R. F.,Characterization and solubilization of bitter-responsivereceptors that couple to gustducin. Proc. Natl. Acad. Sci.U.S.A., 95, 8933–8938 (1998).

12) Kolesnikov, S. S., and Margolskee, R. F., A cyclic-nucleotide-suppressible conductance activated by trans-ducin in taste cells. Nature, 376, 85–88 (1995).

13) Misaka, T., Kusakabe, Y., Emori, Y., Gonoi, T., Arai, S.,and Abe, K., Taste buds have a cyclic nucleotide-activated channel, CNGgust. J. Biol. Chem., 272, 22623–22629 (1997).

14) Gilbertson, T. A., Damak, S., and Margolskee, R. F., Themolecular physiology of taste transduction. Curr. Opin.Neurobiol., 10, 519–527 (2000).

15) Margolskee, R. F., Molecular mechanisms of bitter andsweet taste transduction. J. Biol. Chem., 277, 1–4 (2002).

16) Montmayeur, J. P., and Matsunami, H., Receptors forbitter and sweet taste. Curr. Opin. Neurobiol., 12, 366–371 (2002).

17) Ozeck, M., Brust, P., Xu, H., and Servant, G., Receptorsfor bitter, sweet and umami taste couple to inhibitory Gprotein signaling pathways. Eur. J. Pharmacol., 489,139–149 (2004).

18) Cummings, T. A., and Kinnamon, S. C., Apical Kþ

channels in Necturus taste cells. Modulation by intra-cellular factors and taste stimuli. J. Gen. Physiol., 99,591–613 (1992).

19) Kinnamon, S. C., and Roper, S. D., Membrane propertiesof isolated mudpuppy taste cells. J. Gen. Physiol., 91,351–371 (1988).

20) Rosenzweig, S., Yan, W., Dasso, M., and Spielman, A.I., Possible novel mechanism for bitter taste mediatedthrough cGMP. J. Neurophysiol., 81, 1661–1665 (1999).

21) Ogura, T., and Kinnamon, S. C., IP3-Independent releaseof Ca2þ from intracellular stores: A novel mechanism fortransduction of bitter stimuli. J. Neurophysiol., 82,2657–2666 (1999).

22) Zhang, Y., Hoon, M. A., Chandrashekar, J., Mueller, K.L., Cook, B., Wu, D., Zuker, C. S., and Ryba, N. J.,Coding of sweet, bitter, and umami tastes: Differentreceptor cells sharing similar signaling pathways. Cell,112, 293–301 (2003).

23) Tsunenari, T., Kurahashi, T., and Kaneko, A., Activation

1650 S. SAWANO et al.

by bitter substances of a cationic channel in membranepatches excised from the bullfrog taste receptor cell. J.Physiol., 519, 397–404 (1999).

24) Caicedo, A., Pereira, E., Margolskee, R. F., and Roper,S. D., Role of the G-protein subunit �-gustducin in tastecell responses to bitter stimuli. J. Neurosci., 23, 9947–9952 (2003).

25) Ueda, T., Ugawa, S., Yamamura, H., Imaizumi, Y., andShimada, S., Functional interaction between T2R tastereceptors and G-protein � subunits expressed in tastereceptor cells. J. Neurosci., 23, 7376–7380 (2003).

26) Spielman, A. I., Mody, I., Brand, J. G., Whitney, G.,MacDonald, J. F., and Salter, M. W., A method forisolating and patch-clamping single mammalian tastereceptor cells. Brain Res., 503, 326–329 (1989).

27) He, W., Yasumatsu, K., Varadarajan, V., Yamada, A.,Lem, J., Ninomiya, Y., Margolskee, R. F., and Damak,S., Umami taste responses are mediated by �-transducinand �-gustducin. J. Neurosci., 24, 7674–7680 (2004).

28) Seto, E., Hayashi, Y., and Mori, T., Patch clamprecording of the responses to three bitter stimuli inmouse taste cells. Cell. Mol. Biol., 45, 317–325 (1999).

29) Nakamura, T., Akiyoshi, T., Tanaka, N., Shinozuka,K., Matzno, S., Nakabayashi, T., Matsuyama, K.,Kashiwayanagi, M., and Uchida, T., Effect of quininesolutions on intracellular Ca2þ levels in Neuro-2a cells:conventional physiological method for the evaluation ofbitterness. Biol. Pharm. Bull., 26, 1637–1640 (2003).

30) Ruiz-Avila, L., Wong, G. T., Damak, S., andMargolskee, R. F., Dominant loss of responsiveness tosweet and bitter compounds caused by a single mutationin �-gustducin. Proc. Natl. Acad. Sci. U.S.A., 98, 8868–8873 (2001).

31) Spielman, A. I., Nagai, H., Sunavala, G., Dasso, M.,Breer, H., Boekhoff, I., Huque, T., Whitney, G., andBrand, J. G., Rapid kinetics of second messengerproduction in bitter taste. Am. J. Physiol., 270, C926–C931 (1996).

32) Yan, W., Sunavala, G., Rosenzweig, S., Dasso, M.,Brand, J. G., and Spielman, A. I., Bitter taste transducedby PLC-�2-dependent rise in IP3 and �-gustducin-dependent fall in cyclic nucleotides. Am. J. Physiol.Cell Physiol., 280, C742–C751 (2001).

33) Bosch, R. R., Patel, A. M., Van Emst-de Vries, S. E.,Smeets, R. L., Depont, J. J., and Willems, P. H., U73122and U73343 inhibit receptor-mediated phospholipase Dactivation downstream of phospholipase C in CHO cells.Eur. J. Pharmacol., 346, 345–351 (1998).

34) Muto, Y., Nagano, T., and Urushidani, T., The putativephospholipase C inhibitor U73122 and its negative

control, U73343, elicit unexpected effects on the rabbitparietal cell. J. Pharmacol. Exp. Ther., 282, 1379–1388(1997).

35) Cho, H., Youm, J. B., Shin, Ryu, Y., Earm, Y. E., andHo, W.-K., Inhibition of acetylcholine-activated Kþ

currents by U73122 is mediated by the inhibition ofPIP2-channel interaction. British J. Pharmacol., 134,1066–1072 (2001).

36) Lucas, K. A., Pitarri, G. M., Kazerounian, S., Ruis-Stewarti, I., Park, J., Schulz, S., Chepenk, K. P., andWaldman, S. A., Guanylyl cyclases and signaling bycyclic GMP. Pharmacol. Rev., 52, 375–413 (2000).

37) Vandecasteele, G., Verde, I., Martin, C. R., Gouge, P.D., and Fischmeister, R., Cyclic GMP regulation of theL-type Ca2þ channel current in human atrial myocytes.J. Physiol. (Lond.), 533, 329–340 (2001).

38) Shimizu, K., Shintani, Y., Ding, W.-G., Matsuura, H.,and Bamba, T., Potentiation of slow component ofdelayed rectifier Kþ current by cGMP via two distinctmechanisms: inhibition of phosphodiesterase 3 andactivation of protein kinase G. British J. Pharmacol.,137, 127–137 (2002).

39) Murray, R. G., and Murray, A., Relations and possiblesignificance of taste bud cells. Contrib. Sens. Physiol., 5,47–95 (1971).

40) Kinnamon, J. C., Taylor, B. J., Delay, R. J., and Roper,S. D., Ultrastructure of mouse vallate taste buds. I. Tastecells and their associated synapses. J. Comp. Neurol.,235, 48–60 (1985).

41) Seta, Y., and Toyoshima, K., Three-dimensional struc-ture of the gustatory cell in the mouse fungiform tastebuds: A computer-assisted reconstruction from serialultrathin sections. Anat. Embryol., 191, 83–88 (1995).

42) Yang, R., Tabata, S., Crowley, H. H., Margolskee, R. F.,and Kinnamon, J. C., Ultrastructural localization ofgustducin immunoreactivity in microvilli of type II tastecells in the rat. J. Comp. Neurol., 425, 139–151 (2000).

43) He, W., Yasumatsu, K., Varadarajan, V., Yamada, A.,Lem, J., Ninomiya, Y., Margolskee, R. F., and Damak,S., Umami taste responses are mediated by alpha-transducin and alpha-gustducin. J. Neurosci., 24, 7674–7680 (2004).

44) Royer, S. M., and Kinnamon, J. C., Ultrastructure ofmouse foliate taste buds: Synaptic and nonsynapticinteractions between taste cells and nerve fibers. J.Comp. Neurol., 270, 11–24, 58–59 (1988).

45) Medler, K. F., Margolskee, R. F., and Kinnamon, S. C.,Electrophysiological characterization of voltage-gatedcurrents in defined taste cell types of mice. J. Neurosci.,23, 2608–2617 (2003).

Denatonium Transduction Pathways in Mouse Taste Cells 1651