Sugar-regulated cation channel formed by an insect gustatory receptor

Abstract

Insects sense the taste of foods and toxic compounds in their environment through the gustatory system. Genetic studies using fruit flies have suggested that putative seven-transmembrane gustatory receptors (Grs) expressed in gustatory sensory neurons are required for responses to specific tastants. We reconstituted sugar responses of Bombyx mori Gr-9 (BmGr-9), a silkworm Gr, in two heterologous expression systems. Xenopus oocytes or HEK293T cells expressing BmGr-9 selectively responded to d-fructose with an influx of extracellular Ca²⁺ and a nonselective cation current conductance in a G protein-independent manner. Outside-out patch-clamp recording of BmGr-9–expressing cell membranes provides evidence supporting the hypothesis that BmGr-9 constitutes a ligand-gated ion channel. The fructose-activated current associated with BmGr-9 was suppressed by other hexoses, including glucose and sorbose. The activation and inhibition of insect Gr ion channels may be the molecular basis for the decoding system that discriminates subtle differences in sweet taste. Finally, Drosophila melanogaster Gr43a (DmGr43a), a BmGr-9 ortholog, also responded to d-fructose, suggesting that DmGr43a relatives appear to compose the family of fructose receptors.

In insects, taste and the gustatory systems play a critical role in multiple behaviors, including feeding, toxin avoidance, courtship, mating, and oviposition. Gustatory organs are widely distributed over the entire surface of the body, enabling insects to efficiently detect nonvolatile chemosensory information such as potential foods or toxic compounds. Taste substances are recognized by gustatory sensory neurons that express putative seven-transmembrane proteins in the gustatory receptor (Gr) family. Like the odorant receptor (Or) family, the Gr family is encoded by many related but diverse genes, and genome projects have revealed 68, 13, 76, and 65 Gr genes in the fruit fly (1), honeybee (2), mosquito (3), and silkworm moth (4), respectively.

Molecular genetic studies using the fruit fly Drosophila melanogaster support the role of the Gr family in taste perception. DmGr5a is expressed in gustatory organs and is an essential Gr for a sugar trehalose perception (5–7). Multiple Grs have been implicated in the detection of sweet (8, 9) and bitter tastes (10–12) and CO₂ vapors (13, 14). Genetic studies have suggested that G proteins are involved in a Gr-mediated signaling cascade (15–18), but recognizable sequence motifs that support this hypothesis are not present in Gr sequences. Thus, the existing evidence has led to debate about the molecular mechanisms underlying insect taste perception.

Similar to the Gr, insect Or genes also encode proteins with seven-transmembrane domains, but lack the apparent G protein-coupled receptor motifs (19). Evidence for atypical signal transduction characteristics of insect Ors came from studies of silkmoth Bombyx mori Ors (20), and later it was found that insect Ors had the capacity to act as ligand-gated ion channels directly activated by odorants (21, 22). Thus, it seems reasonable to propose that insect Grs may also share an ionotropic coupling mechanism with the insect Ors (23). In this study, we examined the detailed molecular response of a single Gr to fructose in heterologous expression systems and propose that this Gr constitutes a nonselective, fructose-activated cation channel.

Results

d-Fructose Is a Specific Ligand for a B. mori Gustatory Receptor, BmGr-9.

We examined 14 of the 65 B. mori Grs (BmGr) that were most closely related to Gr genes in other insects. RT-PCR experiments demonstrated that 11 of these 14 BmGrs were transcribed in gustatory organs (Fig. S1); 8 of the 11 that were functional Grs (3 were pseudogenes) were expressed in Xenopus laevis oocytes for functional analysis. Using the two-electrode voltage-clamp technique, oocytes injected with Gr cRNA were stimulated with tastants, including compounds that have been reported to elicit feeding behavior or electrical responses in silkworms (24). At a holding potential of –80 mV, an inward current was recorded from oocytes expressing BmGr-9 in response to d-fructose, a monosaccharide hexose (Fig. 1A). On the basis of the dose–response curve, the EC₅₀ value of d-fructose was 56 mM for BmGr-9 (Fig. 1B). The threshold concentration (∼0.3 mM) was slightly lower than that necessary for an electrophysiological response to d-fructose in silkworms in vivo (5–10 mM) (24). BmGr-9 recognized d-fructose very selectively and precisely discriminated between molecules (e.g., an enantiomer and stereoisomers) on the basis of the stereo-configuration of a hydroxyl group on the hexose ring (Fig. 1A).

Fig. 1.

d-Fructose is a ligand for a B. mori gustatory receptor, BmGr-9, and a D. melanogaster gustatory receptor, DmGr43a. (A) The current traces recorded from BmGr-9–expressing Xenopus oocytes or control oocyte (no injection) with sequential application of various sweeteners at the holding potential of −80 mV. Tastants were applied for 3 s at the time indicated by arrowheads. All compounds were tested at 100 mM. The data are representative of recordings from ten oocytes. (B) The BmGr-9 current was dependant on the dose of d-fructose. (Upper) The current responded to application of the indicated concentrations of d-fructose. d-Fructose was sequentially applied for 3 s to the same oocyte at the time points indicated by the red arrowheads. (Lower) Dose–response curve of BmGr-9. The curve was fitted to the Hill equation (n = 6; EC₅₀ = 55.5 mM, Hill coefficient = 1.02). (C) Current-voltage (I–V) relationships resulting from treatments with several concentrations of d-fructose. The data are representative of recordings from four oocytes. (D) Dose-dependent Ca²⁺ responses of HEK293T cells expressing BmGr-9 to d-fructose. The pseudocolored images demonstrate the changes in Fura-2 fluorescence with increasing concentrations of d-fructose where red indicates a cell that shows the greatest response. The chart shows average Fura-2–based Ca²⁺ responses to increasing concentrations of d-fructose (n = 34), and the bar indicates the timing of stimulation. (E) Dose–response curve of BmGr-9 to d-fructose based on quantitative analysis of Ca²⁺ imaging. The curve was fitted to the Hill equation (n = 159; EC₅₀ = 35 mM, Hill coefficient = 1.62). (F) Average Fura-2–based Ca²⁺ responses of COS-7 cells expressing DmGr43a with application of 1 and 10 mM d-fructose (n = 10). The shaded region around the trace represents SEM.

Interestingly, there were dose-dependent changes in current-voltage (I–V) relationships; at near-threshold concentrations of d-fructose, the current showed strong inward rectification (Fig. 1C, Left), whereas the rectification was abolished at concentrations over the EC₅₀ without a shift in reversal potential (Fig. 1C, Right). These results suggested that d-fructose was a ligand for BmGr-9 and that activation of BmGr-9 leads to the generation of a depolarizing receptor potential in oocytes.

We also expressed BmGr-9 in HEK293T cells and performed Ca²⁺ imaging. HEK293T cells expressing BmGr-9 exhibited Ca²⁺ responses to d-fructose in a dose-dependent manner with an EC₅₀ value of 35 mM and a threshold concentration of 3 mM (Fig. 1 D and E). A specific response to d-fructose was also observed in the Ca²⁺ imaging experiment (Fig. S2).

BmGr-9 belongs to one of the functionally unknown members of the insect Gr family that forms a highly confident single lineage on the phylogenetic tree (4). We then tested the function of Drosophila melanogaster Gr43a (DmGr43a), a BmGr-9 ortholog. COS-7 cells expressing DmGr43a exhibited Ca²⁺ responses to d-fructose in a dose-dependent manner (Fig. 1F).

Characterization of d-Fructose–Stimulated Ca²⁺ and Electric Signals in BmGr-9–Expressing HEK293T Cells.

We next characterized the source of increase in intracellular Ca²⁺ when BmGr-9–expressing cells are stimulated by d-fructose. Chelating extracellular Ca²⁺ with EGTA eliminated the d-fructose–activated Ca²⁺ increase, indicating that the Ca²⁺ increase resulted from the influx of extracellular Ca²⁺ (Fig. 2 A and B). The basal Ca²⁺ levels in BmGr-9–expressing cells significantly decreased in the presence of EGTA (Fig. 2 A and B), suggesting that BmGr-9 mediates a spontaneous Ca²⁺ influx in a manner similar to an insect Or complex (21).

Fig. 2.

Characterization of d-fructose–stimulated Ca²⁺ and electric signals in BmGr-9–expressing HEK293T cells. (A) Average Ca²⁺ responses of BmGr-9–expressing HEK293T cells (n = 29; Left) or vector (n = 29; Right) to 100 mM d-fructose (20-s stimulation) in the presence or absence of 10 mM EGTA (red bar). (B) Quantification of response amplitudes of d-fructose–stimulated Ca²⁺ responses in A. (C) Average Ca²⁺ responses of HEK293T cells expressing BmGr-9 plus α₁-adrenergic receptor (AR) to 100 mM d-fructose (20-s stimulation) or to 100 nM norepinephrine (5-s stimulation) before and after application of 10 μM U73122 (blue bar) (n = 7). (D) Quantification of cAMP in HEK293T cells expressing mOR-EG or BmGr-9 stimulated with eugenol (1 mM) or various concentrations of d-fructose (from 1 to 300 mM), respectively (n = 4, mean ± SEM). (E) Average Ca²⁺ responses of HEK293T cells expressing BmGr-9 to d-fructose (100 mM), 8-bromo-cAMP (1 mM), 8-bromo-cGMP (1 mM), and forskolin (5 μM) (20-s application) (n = 18). (F) Effect of GDP-βS on ligand-induced whole-cell currents in HEK 293T cells (each: n = 11–13). (Left panels) Arrowheads indicate the timing of the 20-s carbachol application to vector-transfected HEK293T cells. (Right panels) Arrows indicate the timing of the 1-s d-fructose application to BmGr-9–transfected HEK293T cells. Recording was performed in the presence (blue trace) or absence (green trace) of 2 mM GDP-βS in the electrode. Red bar indicates the timing of GDP-βS application (whole-cell mode configuration). Significance was assessed by t test or ANOVA and Fisher's protected least squares difference: *P < 0.05, ***P < 0.001. Error bars and shaded regions around the Ca²⁺ response traces represent SEM.

We performed pharmacological experiments to determine whether the d-fructose–evoked Ca²⁺ influx required a G protein-mediated intracellular signaling cascade. U73122, an inhibitor of phospholipase C, abolished α₁-adrenergic receptor-mediated Ca²⁺ responses, but did not affect the BmGr-9–evoked Ca²⁺ responses (Fig. 2C). Although eugenol stimulation resulted in cAMP production in HEK293T cells expressing mouse eugenol olfactory receptor (mOR-EG), no cAMP increase resulted from d-fructose stimulation of BmGr-9–expressing HEK293T cells (Fig. 2D). Application of a cell membrane-permeable cyclic-nucleotide analog (either 8-bromo-cGMP or 8-bromo-cAMP) or an adenylyl cyclase activator (forskolin) failed to produce Ca²⁺ responses in BmGr-9–expressing HEK293T cells (Fig. 2E).

To examine whether G protein signaling was necessary for BmGr-9 activation more directly, we loaded 2 mM GDP-βS, a non-hydrolyzable form of GDP that inhibits G protein-coupled signaling, into the HEK293T cells and performed whole-cell patch-clamp experiments. GDP-βS significantly inhibited endogeneous muscarinic receptor-mediated current responses induced by carbachol, but GDP-βS had no effect on the current response of BmGr-9 to d-fructose (Fig. 2F). These results suggest that G protein-signaling cascades are not involved in the response of BmGr-9 to d-fructose.

d-Fructose–Activated Cation Channel Activity Produced by BmGr-9.

We then hypothesized that BmGr-9 formed a ligand-gated cation channel, similar to an insect Or complex (21, 22). Inward currents were observed in whole-cell voltage-clamp recordings of BmGr-9–expressing HEK293T cells at a holding potential of –60 mV (Fig. 3A). The average latency of the BmGr-9–mediated current was 73 ± 6.2 ms, ranging from 22 to 201 ms (n = 35) (Fig. 3B). The slopes of the initial BmGr-9 currents were superimposable regardless of the duration of stimulation (Fig. 3B), unlike the case of G protein-mediated odorant responses in vertebrates (21, 25). In a manner similar to the results from the oocyte recordings (Fig. 1C), the BmGr-9 current showed inward rectification at near-threshold concentrations of d-fructose (Fig. 3C). COS-7 cells expressing DmGr43a also exhibited the current responses to 100 mM d-fructose without current rectification (Fig. 3D). Reversal potentials changed depending on the cation composition in the external and internal solution at 100 mM d-fructose stimulation, suggesting that BmGr-9 elicited a nonselective cation conductance (Fig. 3E). The 100 mM d-fructose–evoked inward current was inhibited by a calcium channel blocker, ruthenum red (Fig. 3F).

Fig. 3.

BmGr-9 forms a d-fructose–activated cation channel. (A) A whole-cell current recorded from a HEK293T cell expressing BmGr-9 upon stimulation with 1 or 100 mM d-fructose for 200 ms. (B) Onset of BmGr-9–mediated inward current in response to d-fructose stimulus of increasing duration as indicated by the different colors. (C) Current response of a HEK293T cell expressing BmGr-9 (10 or 100 mM d-fructose at various holding potentials). Top traces indicate onset of stimulation. (Left) Blue shows a cell with strong inward rectification, and (Right) green shows a cell without rectification. (Right) I–V relationship with the respective peak current of each response in Left panel plotted. (D) A whole-cell current recorded from a COS-7 cell expressing DmGr43a upon stimulation with 100 mM d-fructose for 2 s (first application) or 5 s (second and third applications). (Right) I–V relationship. (E) Reversal potentials of whole-cell current responses of HEK293T cells expressing BmGr-9 [100 mM d-fructose resulting from different cation compositions in the recording solution (each: n = 7–24)]. Ion and reagents in the external (Ext.) and internal (Int.) solutions are indicated at the bottom. (F) Effect of ruthenium red at a holding potential of +80 and −60 mV on the whole-cell current response of a HEK293T cell expressing BmGr-9. The timing of 100 mM d-fructose and 50 μM ruthenium red application is indicated by white and red bars, respectively. (G) Excised outside-out patch-clamp recording of BmGr-9 currents measured in a HEK293T cell membrane. The top trace shows the timing of the addition of 100 mM d-fructose. All-point current histograms (bottom) were obtained from the region indicated on a trace of first stimulation by the green bar before stimulation (green histogram, left) and by the blue bar during stimulation (blue histogram, right). (Right) Recording of an untransfected cell membrane. Mean peak current levels were obtained from the fitted Gaussians.

We further characterized the source of BmGr-9–evoked current induced by d-fructose. The macroscopic currents recorded from the outside-out cell membrane of a BmGr-9–expressing HEK293T cell showed electrical characteristics similar to those of whole-cell currents (Fig. S3A). Finally, outside-out patch-clamp recordings from HEK293T cell membrane expressing BmGr-9 showed unitary currents of 1.29 ± 0.083 pA [response index (RI) ranging from 0.07 to 0.76; n = 5] with a conductance of 17.1 picosiemens (pS) when cells were held at –60 mV upon stimulation with d-fructose (Fig. 3G). The activation of channel opening by d-fructose was not observed from vector-transfected cell membranes (n = 10). This measured conductance was consistent with a single-channel conductance obtained by a noise analysis (26): 14.9 ± 0.88 pS by the linear fit (n = 4) (Fig. S4). Taken together, these results provide evidence supporting the hypothesis that BmGr-9 forms a d-fructose–activated cation channel.

Inhibition of BmGr-9–Mediated Currents by Hexoses.

We next examined whether BmGr-9 channel activity was modulated by other sugars. No enhancement or suppression was observed when 300 mM of pentose, disaccharide, or trisaccharide (i.e., arabinose, sucrose, trehalose, or raffinose) was added to 100 mM d-fructose in BmGr-9–expressing oocytes (Fig. 4A). However, some hexoses, including d-glucose, d-galactose, and l-sorbose, suppressed the d-fructose–evoked inward current in BmGr-9–expressing oocytes (Fig. 4A). The inhibitory effect was dose-dependent (Fig. 4 B and C), and the current recovered with increasing amounts of d-fructose, suggesting that the hexoses competed with d-fructose at the ligand binding site on BmGr-9 (Fig. 4D). Similarly, the BmGr-9–mediated d-fructose–dependent current in HEK293T cells was inhibited by d-glucose (Fig. 4E, Lower). Spontaneous BmGr-9–mediated activity was also suppressed by d-glucose, and a reversal current response was observed (Fig. 4E, Upper). The latency of the d-fructose–evoked current was longer in the presence of d-glucose (Fig. 4F). The macroscopic current activity of the outside-out cell membrane of a BmGr-9–expressing HEK293T cell was also inhibited by d-glucose (Fig. S3B). Together, these results demonstrated that the BmGr-9 channel is positively and negatively regulated by monosaccharide hexoses.

Fig. 4.

Inhibition of BmGr-9–mediated currents by hexoses. (A) The d-fructose–evoked whole-cell currents recorded from BmGr-9–expressing Xenopus oocytes were inhibited by coapplication of 300 mM sweeteners (n = 6, mean ± SEM). (B) Dose-dependent inhibition of d-fructose–evoked inward currents in BmGr-9–expressing Xenopus oocytes by d-glucose, d-galactose, or l-sorbose. Current traces upon 3 s application of a mixture of 10 mM d-fructose and the indicated concentration of d-glucose (magenta trace), d-galactose (green trace), or l-sorbose (blue trace); the inhibitors were added at the time points indicated by the arrowheads. (C) Quantification of dose-dependent inhibition (n = 7, mean ± SEM). (D) The inhibition associated with 300 mM d-glucose or l-sorbose can be overcome in BmGr-9–expressing oocytes by increasing the concentration of d-fructose to 30 mM (blue bar) or 100 mM (brown bar) (n = 4, mean ± SEM). (E) Current traces of a BmGr-9–expressing HEK293T cell upon application of 100 mM d-fructose (open bar) or 300 mM d-glucose (magenta bar) (F) Onset of inward current in a BmGr-9–expressing HEK293T cell to d-fructose in the presence (magenta trace) or absence (black trace) of 300 mM d-glucose; the competitor was added at the time point indicated by the open bar. Significance was assessed by the t test: *P < 0.05, **P < 0.01, ***P < 0.001. The holding potential was −80 mV (Xenopus oocytes) and −60 mV (HEK293T).

Discussion

In this study, we propose that BmGr-9, a silkworm Gr, constitutes a nonselective, fructose-activated cation channel that is inactivated by several hexoses, including glucose, galactose, and sorbose. Although genetic studies using Drosophila have suggested that coexpression of multiple Grs is necessary to respond to sugars such as d-glucose, trehalose, and sucrose (8, 9), to bitter compounds such as caffeine and theophylline (10–12), and to CO₂ (13, 14), BmGr-9 appears not to require the expression of other Grs to show the responsiveness to d-fructose in vitro. BmGr-9–expressing neurons, however, may coexpress other Grs among 65 BmGr genes, and therefore, we cannot exclude the possibility that BmGr-9 exhibits different ligand response properties in vivo.

BmGr-9 and its orthologs, including DmGr43a, form a distinct Gr subfamily that is not categorized in the sugar or bitter receptor subfamilies but are conserved within endopterygote insects (4). We demonstrated that both BmGr-9 and DmGr43a responded to d-fructose, suggesting that the DmGr43a ortholog family may represent a Gr subfamily that plays a role in d-fructose perception. Although the physiological function of d-fructose in insect species has not been well characterized, the fact that the DmGr43a family is separated from other sugar receptor gene families on the phylogenetic tree indicates that d-fructose may play a distinct physiological role in insects. In this regard, it is intriguing to note that BmGr-9 is also expressed in the gut (Fig. S1), suggesting the involvement in intestinal absorption or some metabolism.

As is the case for insect Ors, the insect gustatory system has long been thought to use a G protein-mediated signaling pathway. Genetic ablation of Gα_s, Gα_q, Gγ1, or phospholipid signaling resulted in partial reduction in trehalose responses in Drosophila (15, 16, 18). Additionally, Gα_o is evidently involved in the detection of sucrose, glucose, and fructose in Drosophila (27). However, there was no evidence for a second messenger-mediated pathway in the BmGr-9 response to d-fructose. Our results were consistent with previous electrophysiology experiments that suggested the presence of d-fructose-driven ion channel transduction in the flesh-fly sugar receptor neurons (28), and these sugar-activated currents showed nonselective cation conductance in vivo (29).

In conclusion, we provide several lines of evidence supporting the hypothesis that an insect Gr is an ionotropic channel regulated by taste substances. Similar to the olfactory system, a channel that is both positively and negatively regulated plays a role in taste perception in insects. Although the insect Ors and Grs do not resemble each other at the amino acid sequence level, our findings suggest that these insect chemosensory systems have a common mechanism for decoding a variety of chemical signals in the external environment. Our success in reconstituting the sugar responsiveness of an insect Gr and in performing pharmacology in vitro paves the way for characterizing the remaining insect Grs and for a better understanding of the contribution of many Grs to taste perception and the regulation of feeding behaviors in insects.

Materials and Methods

Insects.

Eggs of the silkworm B. mori (hybrid strain, Kinshu × Showa) were purchased from Ueda Sanshu Ltd.. Larvae were reared in plastic containers at 25 °C with 70% relative humidity and long-day lighting conditions (16 h light/8 h dark) on a SILKMATE 2S artificial diet (Nippon Nosan Co. Ltd.). Larvae were provided with fresh food on a daily basis, and all larvae were staged to synchronize growth.

RT-PCR.

Total RNA was isolated from the antennae of 10 male adult moths, 10 female adult moths, and 10 larvae maxilla; from the labrum, mandible, labium, thoracic, and proleg of 10 larvae; and from the gut of one larva using a Microto-Midi Total RNA Purification System (Invitrogen). Following treatment with DNase I (Promega), cDNA was produced using SuperScript III (Invitrogen). cDNA was amplified using Ex Taq DNA polymerase (Takara) under the following reaction conditions: 94 °C for 5 min and then 40 cycles at 94 °C for 30 s, * °C for 30 s, and 72 °C for 2 min, followed by 72 °C for 7 min [asterisk (*) indicates BmGr-1, -2, -5, -9, -66, and -67 primers at 60 °C; BmGr-4, -6, -7, -8, -53, and -68 primers at 54 °C; and BmGr-3 and -10 primers at 52 °C]. The following primer pairs were used for the PCR: BmGr-3 (5′-ATGTCCTTCGAAATAAAAAATAATTTC-3′ and 5′-TCAATCATTTTTTCTTTTCGCAAAAGC-3′); BmGr-2 (5′-ATGATTCCGGACCATCTTTTTGAAG-3′ and 5′-TTATTGACCGGTGCCATGTG-3′); BmGr-1 (5′-ATGAACAGACACGACCATAGATTC-3′ and 5′-TCATTCTTGATCTTCATCACTTCC-3′); BmGr-66 (5′-ATGAAACGTAAATTAAAGAAGTTTTTTCCG-3′ and 5′-TTACGAACGACGTTGTTCTTG-3′); BmGr-67 (5′-ATGAGAGAAAGAAAAAAAAAATTTAACAAA-3′ and 5′-TTATCGCCCGCGTC-3′); BmGr-4 (5′-ATGTCACGGATCTTCTCGATG-3′ and 5′-TTATATAGGTACCTCTAACAACAATTC-3′); BmGr-53 (5′-ATGGCTCAAATAAAAGATGAAAATCAATC-3′ and 5′-TTAGACAAAATGAGAGAGTTGAATAATC-3′); BmGr-8 (5′-ATGGCTCCTCGATCAGTTC-3′ and 5′-TTAAATTTGAAGTAATACTATTTCGTACGT-3′); BmGr-9 (5′-ATGCCTCCTTCGCCAG-3′ and 5′-TTAACTATCATATCGCTGGAATTGAATG-3′); BmGr-5 (5′-ATGAGTAAAATTCTAAAATTCCTGAC-3′ and 5′-TTAATCTTCATTGCTGAATTGAAGC-3′); BmGr-6 (5′-ATGGTAACACAGTTCCTTAACATTC-3′ and 5′-CTACGAGTAGTTGTAAAATGTTTC-3′); BmGr-10 (5′-ATGATCAAAATTCGACTAAAAACATTAAG-3′ and 5′-TTATTGCATATTTTTAATTTCCAGCTG-3′); BmGr-7 (5′-ATGTGTTGTTTGGGAGAAACCAG-3′ and 5′-TCATATAAGATGTTTCGGCAAATAATAATC-3′); and BmGr-68 (5′-ATGCGTTTCGGTTTGAAGGC-3′ and 5′-TTAATTTCTTTTATCAAACTGAACTAAAAT-3′).

Patch-Clamp Experiments in Mammalian Cell Lines and Outside-Out Current Data Analysis.

A full-length cDNA for BmGr-9 and DmGr43a were cloned into the pME18S vector. This Gr expression vector was transiently transfected into COS-7, HeLa, or HEK293T cells with lipofectamine 2000 reagent (Invitrogen). GFP or monomeric red fluorescent protein were cotransfected as a control. Whole-cell currents were amplified with a patch-clamp amplifier (Axopatch 200B; Molecular Devices) and were digitized with PowerLab (AD Instruments). The extracellular buffer solution contained (in mM) 140 NaCl, 5.6 KCl, 5 Hepes, 2.0 pyruvic acid sodium salt, 1.25 KH₂PO₄, 2.0 CaCl₂, and 2.0 MgCl₂, (pH 7.4). The electrode solution contained (in mM: 140 KCl, 10 Hepes, 5 EGTA-2K; pH 7.4). To record the shift in reversal potential and in equilibrium potential for a specific ion, the following external (bath) and internal (electrode) solutions were used: N-methyl-d-glutamine (NMDG) plus Ca²⁺ external solution (in mM: 190 NMDG, 40 Hepes, 5.6 KCl, 2.0 pyruvic acid sodium salt, 1.25 KH₂PO₄, 2.0 CaCl₂, and 2.0 MgCl₂; pH 7.4); NMDG external solution (in mM: 190 NMDG, 40 Hepes, 5.6 KCl, 2.0 pyruvic acid sodium salt, 1.25 KH₂PO₄, and 2.0 MgCl₂; pH 7.4); K⁺plus Ca²⁺ external solution (in mM: 145 KCl, 5 Hepes, 2.0 pyruvic acid sodium salt, 1.25 KH₂PO₄, 2.0 CaCl₂, and 2.0 MgCl₂; pH 7.4); and Na⁺ internal solution (in mM: 140 NaCl, 10 Hepes, and 5 EGTA-2Na, pH 7.4). For the outside-out recording, both the external and electrode sodium solutions contained (in mM) 140 NaCl, 5 Hepes, and 2.0 pyruvic acid sodium salt, pH 7.4. The data were sampled at 20 kHz and low-pass-filtered at 2 kHz.

For the analysis of an outside-out current recording, the channel conductance was first determined by the peak of fitted multiple Gaussian function. Next, the responsiveness of a channel was assessed by the RI that was defined as the inverted ratio of d-fructose–activated channel activity to baseline: RI = integration of currents during [t = −10, t = 0] (pA·s)/integration of currents during [t = 0, t = 10] (pA·s), where t = −10, t = 0, and t = 10 correspond to 10 s before d-fructose stimulation, onset of stimulation, and 10 s after stimulation, respectively. The values of RI < 1 and RI > 1 were expected to obtain upon channel activation and inhibition, respectively. The responsiveness of each cell membrane was confirmed by reproducible responses to repeated stimulations.

Xenopus Oocyte Electrophysiology.

Oocytes were microinjected with 50 ng of BmGr-9 cRNA. Injected oocytes were then incubated for 3 d at 18 °C in Barth's solution [(in mM): 88 NaCl, 1 KCl, 0.3 Ca(NO₃) ₂, 0.4 CaCl₂, 0.8 MgSO₄, 2.4 NaHCO₃, and 15 Hepes, pH 7.6, supplemented with 10 μg/mL penicillin and streptomycin]. Whole-cell currents were recorded with a two-electrode voltage clamp filled with 3 M KCl and were amplified with an OC-725C amplifier (Warner Instruments), low-pass-filtered at 50 Hz and digitized at 1 kHz. Tastants were applied through the superfusing bath solution [(in mM): 115 NaCl, 2.5 KCl, 1.8 MgSO₄, 2.4 NaHCO₃, and 10 Hepes; pH 7.2]. Solutions were switched using handmade, Peripheral Interface Controller-driven (16F877, Microchip Technology Inc.) solenoid valves (Cole Parmer).

Ca²⁺ Imaging.

BmGr-9 or DmGr43a was transfected into HEK293T or COS-7 cells, which were loaded with 2.5 μM Fura-2/AM for 30 min. Fluorescence was measured with an Aquacosmos Ca²⁺ imaging system (Hamamatsu Photonics). Tastants and drugs were delivered through the superfusing exracellular buffer solution.

cAMP Assay.

HEK293T cells transfected with BmGr-9 were incubated with 1 mM 3-isobutyl-1-methylxanthine (IBMX) for 30 min. The cells were then exposed to the indicated concentration of d-fructose for 15 min. cAMP levels were determined with an ELISA kit (Applied Biosystems) in accordance with the manufacturer's directions.