A Drosophila Gustatory Receptor Required for Strychnine Sensation

Abstract

Strychnine is a potent, naturally occurring neurotoxin that effectively protects plants from animal pests by deterring feeding behavior. In insects, such as the fruit fly, Drosophila melanogaster, bitter-tasting aversive compounds are detected primarily through a family of gustatory receptors (GRs), which are expressed in gustatory receptor neurons. We previously described multiple GRs that eliminate the behavioral avoidance to all bitter compounds tested, with the exception of strychnine. Here, we report the identity of a strychnine receptor, referred to as GR47a. We generated a mutation in Gr47a and found that it eliminated strychnine repulsion and strychnine-induced action potentials. GR47a was narrowly tuned, as the responses to other avoidance compounds were unaffected in the mutant animals. This analysis supports an emerging model that Drosophila GRs fall broadly into two specificity classes—one class is comprised of core receptors that are broadly required, whereas the other class, which includes GR47a, consists of narrowly tuned receptors that define chemical specificity.

Introduction

Plants produce nonvolatile repellent compounds, such as alkaloids, to ward off pests. Among the best-characterized alkaloids is strychnine. In mammals, strychnine is detected through taste receptors, which are G-protein coupled receptors (Bufe et al. 2002; Meyerhof et al. 2010). However, in insects, the taste receptors that enable the behavioral decision to avoid strychnine-containing foods are not known.

Most taste receptors in Drosophila belong to the 68-member gustatory receptor (GR) family (Clyne et al. 2000; Dunipace et al. 2001; Scott et al. 2001; Robertson et al. 2003). GRs are unrelated to mammalian taste receptors but are distantly related to Drosophila olfactory receptors. Thirty-one hair-like bristles (sensilla) are distributed on each of two bilaterally symmetrical labella and are grouped based on length: short (S), intermediate (I), and long (L) (Liman et al. 2014). Each sensillum contains two or four gustatory receptor neurons (GRNs). Sugar sensation occurs through one GRN per sensillum (Hiroi et al. 2002). Detection of bitter compounds is also mediated primarily through one GRN in I-type and S-type sensilla (Meunier et al. 2003; Weiss et al. 2011).

Based on reporter expression in I- and S-type sensilla on the labellum, >33 genes encode GRs that are likely to function in bitter sensation, and they appear to compartmentalize into groups corresponding to the functional classes (Weiss et al. 2011). Currently, only five bitter-responsive GRs have been characterized through analyses of loss-of-function mutations. Three (GR32a, GR33a, and GR66a) are required broadly for responding to most aversive compounds and may be coreceptors (Moon et al. 2006; Moon et al. 2009; Lee et al. 2010). In addition to these “core-bitter GRs” (Weiss et al. 2011), two GRs (GR93a and GR8a) are narrowly tuned and are required for the responses to caffeine and the toxic amino acid, L-canavanine, respectively (Lee et al. 2009; Lee et al. 2012).

In insects, the receptor requirement for strychnine is enigmatic. With the exception of strychnine, the behavioral repulsion and electrophysiological responses to other deterrent tastants depend on GR32a, GR33a, and GR66a (Moon et al. 2006, 2009; Lee et al. 2010). However, the effects on the strychnine response resulting from mutations in the core-bitter Grs are complex. Loss of any of these receptors virtually eliminates strychnine-induced action potentials. Surprisingly, strychnine repulsion is normal in either Gr33a or Gr66a mutants and only moderately reduced in Gr32a mutants. Thus, no GR has been defined that is essential for strychnine rejection.

Here, we found that mutation of Gr47a profoundly reduced both the behavioral and electrophysiological responses to strychnine. Gr47a ¹ mutants responded normally to other repellent compounds tested, indicating that GR47a is a relatively narrowly tuned strychnine receptor. We propose that GR47a is a receptor that imparts strychnine specificity.

Materials and methods

Drosophila stocks

We reported the following mutants previously and deposited them in the Bloomington Stock Center: Gr33a ¹, UAS-Gr33a, UAS-Gr66a, and Gr66 ^ex83 (Moon et al. 2006, 2009). H. Amrein provided the ΔGr32a and UAS-Gr32a (Miyamoto and Amrein 2008) and the P[Gr66a-GAL4] flies (Thorne et al. 2004). K. Scott provided the P[Gr47a-I-GFP] and P[Gr47a-GAL4] flies (Wang et al. 2004). J.Y. Kwon provided the P[Gr8a-GAL4], P[Gr36a-GAL4], P[Gr39b-GAL4], P[Gr59c-GAL4], P[Gr93a-GAL4], and P[Gr98b-GAL4] flies (Weiss et al. 2011). We obtained the Gr47a deficiency line, Df(2R)12, from the Bloomington Stock Center (stock 5425). We used w ¹¹¹⁸ as the “wild-type” control.

Generation of mutant and transgenic fly lines

We used ends-out homologous recombination (Gong and Golic 2003) to generate the deletion in Gr47a ¹, which removed residues −166 to +291. To obtain the knockout construct, we amplified two 3-kb genomic fragments by polymerase chain reaction (PCR), and then subcloned the DNAs into the pw35 vector (Gong and Golic 2003). We generated the original transgenic insertion lines using germline transformation (BestGene Inc.) and obtained homologous recombinants as described (Gong and Golic 2003). We confirmed the homologous insertions by PCR using primer pairs P1/P2 and P3/P4 (Figure 1A): P1, 5′-TGGCCTGACCCAAAGGCCTATAAA-3′; P2, 5′-TCAGAA CAGTCACACTCACACGCA-3′; P3, 5′-TGAACTGGAATATGGGC GAACCCT-3′; P4, 5′-GCGCTTGTTTGTTTGCTCAGCTTG-3′.

Figure 1.

Gr47a is required for behavioral avoidance to strychnine. (A) Physical map of the Gr47a genomic region. Gr47a ¹ was generated by ends-out homologous recombination. The white box indicates the mini-white gene. The arrows indicate the primers used for the PCR analyses in B. (B) PCR analyses of genomic DNA. The PCR products, which were generated using DNA prepared from control or Gr47a ¹ flies and the indicated PCR primer pairs, were fractionated on an agarose gel. The 3-kb product produced using the P3 and P4 primers indicated successful targeting. The 300-bp product using the P1 and P2 primers confirms the genomic deletion. (C) Avoidance of noxious compounds. The flies were given a choice between 1-mM sucrose and 5-mM sucrose plus the following aversive compounds: 10-mM caffeine (CAF), 10-mM theophylline (TPH), 5-mM umbelliferone (UMB), 1-mM quinine (QUI), 0.2-mM denatonium (DEN), 1-mM papaverine (PAP), 0.3-mM lobeline (LOB), 0.05-mM berberine (BER), and 0.3-mM strychnine (STR). n = 4–8. The control flies used here and throughout this work were w ¹¹¹⁸. (D) Concentration-dependent avoidance of caffeine in control and Gr47a ¹ flies. n = 4. (E) Concentration-dependent avoidance of strychnine in control and Gr47a ¹ flies. n = 4. (F) Strychnine-avoidance behavior exhibited by Gr47a ¹ homozygous flies, Gr47a ¹ placed in trans with a deficiency (Df) that uncovered the Gr47a ¹ mutation, and rescue of the strychnine sensation defect in Gr47a ¹ using a Gr47a ⁺ genomic rescue transgene. n = 4–7. The error bars represent SEMs. The asterisks indicate significant differences from control flies (P < 0.01) using single factor ANOVA with Scheffe’s analysis as a post hoc test to compare two sets of data.

To obtain the UAS-Gr47a transgene, we amplified the full-length Gr47a cDNA by reverse transcription polymerase chain reaction (RT-PCR) using fly labellar mRNA. We subcloned the cDNA into the pUAST vector (Brand and Perrimon 1993) and verified the cDNA by DNA sequencing. The transformation vector was injected into w ¹¹¹⁸ embryos (BestGene Inc.). The Gr47a genomic transgene, P[gGr47a], was 19639-bp long and extended to 3330 and 14799bp 5′ and 3′ of the predicted transcribed region. To generate this transgene, we subcloned the genomic region from P[acman] CH322-152A22 (www.pacmanfly.org) into the insertion site of the attP154 on the 3rd chromosome (BestGene Inc.).

RT-PCR analyses

We used TRIzol (Invitrogen) to extract mRNA from the labella, wings, abdomens, and legs and AMV reverse transcriptase to generate the cDNAs (Promega). To perform the RT-PCR, we used the following Gr47a primers: 5′-ATGGCCTTTACCAGCTCGCA-3′; 5′- GCGAACATGGAGAGCAAACG-3′. The following were the tubulin primers: 5′-TCCTTCTCGCGTGTGAAACA-3′; 5′- CCGAA CGAGTGGAAGATGAG-3′. The Gr47a RT-PCR products were obtained after 40 cycles.

Immunohistochemistry

The labella of Gr66a-GAL4/Gr47a-I-GFP;UAS-DsRed/+ flies were dissected and fixed using 4% paraformaldehyde with 0.2% Triton X-100 in phosphate buffered saline (PBS-T) for 15min at room temperature. We then washed the labella three times with PBS-S (1× PBS and 0.2% saponin), cut them in half with a razor blade, and blocked the samples with blocking buffer (1× PBS, 0.1% saponin and 5mg/ml bovine serum albumin) for 4h at 4 °C. Primary antibodies, anti-GFP green fluorescent protein (GFP) (Molecular Probe, mouse anti-GFP, cat. no. A11120, 1:1000) and anti-DsRed (Clontech, rabbit anti-DsRed, catalogue no. 632496, 1:1000), were added to fresh blocking buffer and incubated with the labella overnight at 4 °C. The labella were washed three times with PBS-T at 4 °C and incubated with secondary antibodies (goat anti-mouse Alexa488 and goat anti-rabbit Alexa568, 1:200) for 4h at 4 °C. The tissues were washed three times with PBS-T and mounted in mounting buffer (37.5% glycerol, 187.5mM NaCl, 62.5mM Tris pH8.8). The samples were viewed using a Zeiss LSM700 confocal microscope.

Chemicals

Sucrose, caffeine, denatonium, lobeline, papaverine, quinine, strychnine, theophylline, umbelliferone, sulforhodamine B, and KCl were purchased from Sigma-Aldrich Co. Berberine sulfate trihydrate and Brilliant Blue FCF were obtained from Wako Pure Chemical Industries Ltd.

Two-way choice behavioral assays

We performed the two-way choice assays as described previously (Meunier et al. 2003; Moon et al. 2006). Briefly, we starved 50–70 flies (3–6 days old) for 18h in a humidified chamber, and then introduced the animals into 72-well microtiter dishes. We filled alternating wells with 1% agarose combined with one of two types of test mixtures: 1-mM sucrose or 5-mM sucrose plus an avoidance chemical. To monitor food intake, one test mixture contained a blue dye (brilliant blue FCF, 0.125mg/ml), whereas the other contained a red dye (sulforhodamine B, 0.2mg/ml). We allowed the flies to feed for 90min at room temperature in the dark and froze the animals at –20 °C. The numbers of flies that were blue (N^B), red (N^R), or purple (N^P) were determined in a blind fashion based on the colors of the abdomen. The preference index (PI) values were calculated according to the following equation: (N^B + 0.5N^P) / (N^R + N^B + N^P) or (N^R + 0.5N^P)/(N^R + N^B + N^P), depending on the dye/tastant combinations. PIs equal to 1.0 and 0 indicated complete preferences for either 1 or 5mM sucrose plus an avoidance chemical, respectively. A PI = 0.5 indicated no bias between the two food alternatives.

Proboscis extension response assays

We performed the proboscis extension response (PER) assays as described (Shiraiwa and Carlson 2007) with minor modifications. We starved the flies by placing them in an empty vial for 18–24h with a piece of Kimwipe soaked with water. We removed the fly from the vial with an aspirator, placed it in a 200-µl yellow tip, and moved it to the end of the tip by applying air. Using a razor blade, we increased the opening of the 200-µl yellow tip. The proboscis extended out of the tip opening. Before testing any bitter tastants, we offered a 2% sucrose solution to the fly. If the fly did not respond to the 2% sucrose, we discarded it. We also tested the flies’ response to water, which served as a negative control to make sure that the flies were not responding to water alone. We then offered each fly 2% sucrose (1st exposure), 2% sucrose plus 1-mM strychnine (1st exposure), 2% sucrose (2nd exposure), and 2% sucrose plus 1-mM strychnine (2nd exposure). We applied water to the flies between each for the preceding food offers.

Electrophysiological responses of GRNs to tastants

We performed tip recordings as described (Moon et al. 2006) using 10-mM caffeine, 10-mM theophylline, 10-mM umbelliferone, 1-mM denatonium, 1-mM lobeline, 1-mM papaverine, 1-mM quinine, 1-mM strychnine, and 0.1-mM berberine. We immobilized newly eclosed flies by inserting a glass capillary filled with Ringer’s solution into the abdomen, which we extended into the head. This electrode also functioned as the indifferent electrode. We stimulated the sensilla with a recording pipet (10- to 20-µm tip diameter) containing the tastants dissolved in 1-mM KCl, which served as the electrolyte in all recordings. The recording electrode was connected to a preamplifier (TastePROBE, Syntech, Hilversum, The Netherlands), and we collected and amplified the signals 10× using a signal connection interface box (Syntech) in conjunction with a 100- to 3000-Hz band-pass filter. Recordings of action potentials were acquired using a 12-kHz sampling rate and analyzed using Autospike 3.1 software (Syntech).

Statistical analyses

All error bars represent standard error of the means (SEMs). Single factor ANOVA with Scheffe’s analysis as a post hoc test was used to compare multiple sets of data. Asterisks indicate statistical significance (*P < 0.05, **P < 0.01).

Results

Strychnine behavioral avoidance depended on Gr47a ¹

To dissect the roles of GRs, we generated a mutation in the Gr47a gene, which encodes a protein (GR47a) that belongs to one of the remaining uncharacterized clades within the GR phylogenetic tree. This branch contains GRs that are distantly related to any of the functionally analyzed GRs, including receptors required for the responses to sugars and bitter compounds. To disrupt Gr47a, we used ends-out homologous recombination. We created Gr47a ¹ by deleting ~500 base pairs extending 166 base pairs 5′ of the predicted transcription start site through the region coding the N-terminal 97 residues of the 361 amino acid protein (Figures 1A,B).

To identify a deficit in Gr47a ¹, we performed two-way choice tests. Normally, flies choose 5- over 1-mM sucrose (Figure 1C). However, addition of bitter compounds eliminates the preference of control animals for the higher concentration of sugar (Figure 1C). The mutant flies also exhibited normal repulsion to most bitter compounds, including caffeine, quinine, and six others (Figures 1C,D). Among the compounds tested, Gr47a was specifically required for inhibiting consumption of strychnine (Figures 1C,E). The impairment in avoiding strychnine was due to mutation of Gr47a because we recapitulated the phenotype when we placed Gr47a ¹ in trans with a deficiency that removed Gr47a (Figure 1F). Furthermore, we rescued the phenotype with a wild-type Gr47a genomic transgene demonstrating that the defect was due to mutation of Gr47a (Figure 1F). Thus, Gr47a was narrowly required for strychnine sensation.

Elimination of strychnine-induced action potentials in Gr47a ¹ flies

We performed tip recordings to assess bitter compound–induced action potentials in Gr47a ¹ flies. We recorded from S6 sensilla, which responded to most aversive tastants including caffeine and strychnine (Weiss et al. 2011). Consistent with the behavior, Gr47a ¹ displayed normal frequencies of action potentials to all bitter compounds tested, except for strychnine (Figures 2A,B).

Figure 2.

Gr47a was indispensible for strychnine-induced nerve firings. (A) Tip recordings were performed on S6 bristles on the labella. Average frequencies of action potentials (spikes/s) to 10-mM caffeine, 10-mM theophylline, 10-mM umbelliferone, 1-mM quinine, 1-mM denatonium, 1-mM papaverine, 1-mM lobeline, 0.1-mM berberine, and 1-mM strychnine are shown. n = 7–12. (B) Representative traces of strychnine-evoked action potentials from control (w ¹¹¹⁸) and Gr47a ¹ flies. (C) Responses of different sensilla to 1-mM strychnine in control and Gr47a ¹ labella. n = 6–17. Two nomenclature systems are used to identify the sensilla: Tanimura study (T) (Hiroi et al. 2002) and Carlson study (C) (Weiss et al. 2011). (D) Tip recordings performed using the indicated concentrations of strychnine on S5, S6, and S10 bristles. n = 7–22. (E) Average frequencies of action potentials (spikes/s) induced by 10-mM caffeine upon application to the indicated sensilla. n = 8–11. (F) Average frequencies of action potentials (spikes/s) induced by 1-mM strychnine on S6 sensilla using the indicated fly strains. n = 8–20. The error bars represent SEMs. The asterisks indicate significant differences from control flies (**P < 0.01, *P < 0.05) using single factor ANOVA with Scheffe’s analysis as a post hoc test.

We surveyed the strychnine responses of S-type sensilla in control and the Gr47a ¹ flies because S-type but not I- or L-type sensilla are activated by strychnine (Weiss et al. 2011). As previously reported, the L- and I-type sensilla were unresponsive to strychnine, whereas multiple S-type sensilla were activated by strychnine (Figure 2C; note that two nomenclature systems are reported and we employ the classical one from the Tanimura group) (Hiroi et al. 2002; Weiss et al. 2011). S1, S3, and S5 were moderately responsive to strychnine, whereas S6, S9, and S10 were most robustly activated. However, S9 were present in only a subset of labella. Of significance here, we found that strychnine-induced action potentials were severely reduced in Gr47a ¹ mutant flies (Figures 2A–C), even at the highest concentrations tested (Figure 2D). Three other Gr47a ¹ sensilla surveyed also showed normal caffeine responses (Figure 2E). We fully rescued this electrophysiological defect with a genomic transgene (Figure 2F).

Rescue of strychnine behavior

There are two Gr47a gene reporters, one of which stains just three sensilla (Gr47a-GAL4, S12, I8, and I9), whereas the other (Gr47a-I-GFP) is detected in many more sensilla (Wang et al. 2004; Weiss et al. 2011). We recapitulated these observations, because we found that 3.2±0.6 (n = 12) GRNs were labeled using the Gr47a-GAL4 driver in combination with the UAS-GFP (Weiss et al. 2011), whereas the Gr47a-I-GFP line stained 14.8±0.2 (n = 7) GRNs per labellum (Figure 3A). Gr47a-I-GFP overlapped partially with the Gr66a-GAL4 reporter, which was ubiquitously expressed in bitter responsive GRNs, and labeled all I-type and half of S-type sensilla. Thus, the limited expression of the Gr47a-GAL4 driver might not reflect the bona fide cellular distribution of Gr47a. Consistent with this conclusion, we could not rescue the strychnine deficit in Gr47a ¹ flies using Gr47a-GAL4 in combination with UAS-Gr47a (Figure 3B; Supplementary Figure 1), and the three Gr47a-GAL4 positive sensilla (S12, I8, and I9) were unresponsive to strychnine (Figure 2C).

Figure 3.

Testing for rescue of strychnine repulsion by expression of Gr47a in different sensilla. (A) The Gr47a-I-GFP reporter was expressed in a subset of Gr66a-GAL4 positive GRNs. Left panel, Gr47a-I-GFP labeled a subset of bitter GRNs (anti-GFP, green). Middle panel, DsRed was expressed using the Gr66a-GAL4 and the UAS-DsRed transgenes (anti-DsRed, red). The Gr66a-GAL4 reporter labeled all bitter GRNs. Right panel, merged image of left and middle panels. (B) Recovery of the strychnine-induced action potentials in the indicated sensilla in Gr47a ¹ after expression of UAS-Gr47a under control of the indicated GAL4 drivers. n = 10. (C) Rescue of the strychnine taste defect in Gr47a ¹ after expression of UAS-Gr47a under control of GAL4 drivers, which label different sensilla (Weiss et al. 2011). n = 5. The error bars represent SEMs. The asterisks indicate significant differences from Gr47a ¹ (*P < 0.05) using single factor ANOVA with Scheffe’s analysis as a post hoc test.

Because Gr47a-I-GFP is a direct fusion of the Gr47a promoter to GFP, we could not use this line for rescue experiments. Therefore, we drove UAS-Gr47a using other GAL4 lines that were expressed in different subsets of bitter-responsive sensilla. We restored strychnine-induced action potentials and strychnine aversion in the Gr47a ¹ mutant flies using Gr36a-GAL4, Gr39b-GAL4, or Gr8a-GAL4 (Figures 3B,C; Supplementary Figure 1). These drivers were detected exclusively in sensilla (S3, S5, and S10) (Weiss et al. 2011) that respond to strychnine (Figure 2C). We obtained robust rescue in S6 sensilla only, as a result of expressing UAS-Gr47a under the control of Gr59c-GAL4 and Gr98b-GAL4 (Figure 3B; Supplementary Figure 1). However, when we used the Gr93a-GAL4 to perform rescue experiments, we observed few action potentials in S6 sensilla (Figure 3B; Supplementary Figure 1), even though the Gr93a-GAL4 expression pattern was similar to Gr98b-GAL4 (Weiss et al. 2011). These findings suggested that the Gr93a-GAL4 was a weak driver relative to the Gr98b-GAL4.

Strychnine avoidance dependent on Gr47a in the labellum

Drosophila include taste sensilla distributed on multiple body parts other than labella, including the legs, wings, and female oviposter, the latter of which is located at the tip of the female abdomen. To address whether Gr47a was expressed in any of these extralabellar portions of the fly, we manually dissected various body parts and performed RT-PCR. As expected, we detected a signal in the proboscis, which included the two bilaterally symmetrical labella (Figure 4A). In addition, we found that Gr47a RT-PCRs were produced in the abdomen and wings (Figure 4A). However, we did not detect a signal in the legs (Figure 4A). Consistent with this latter result, the Gr47-I-GFP reporter did not stain the tarsi.

Figure 4.

Gr47a expression in various body parts and PER assays. (A) Gr47a expression in the proboscis, legs, abdomen, and wings analyzed by RT-PCR. We also amplified tubulin products by RT-PCR, which served as a quality control for the quality of the RNA. (B) PER assays performed using the indicated flies. Either 2% sucrose alone or 2% sucrose and 1-mM strychnine were applied to the labella, and the fraction of flies that elicited a PER were determined. Ten flies/experiment were used (n = 4). The error bars represent SEMs. The asterisks indicate significant differences from Gr47a ¹ (**P < 0.01) using single factor ANOVA with Scheffe’s analysis as a post hoc test to compare two sets of data.

The preceding data suggested that the behavioral avoidance to strychnine resulted from Gr47a-dependent sensation in the labellum. However, taste sensilla in both the legs and labella influence responses using two-way choice assays. Therefore, to address whether strychnine avoidance depended on Gr47a expression in the labellum specifically, we assayed PERs, by applying sucrose alone or sucrose plus strychnine to the labellum. Nearly all control flies extended their probosci in response to application of 2% sucrose, and this response was diminished only slightly upon presentation of sucrose a second time (Figure 4B). Addition of 1-mM strychnine greatly suppressed the PER (Figure 4B). The Gr47a ¹ flies also exhibited robust PERs to 2% sucrose (Figure 4B). However, the suppression by 1-mM strychnine was severely impaired in the mutant animals (Figure 4B). We rescued this defect by expressing UAS-Gr47a under control of the Gr36a-GAL4 (Figure 4B).

Behavioral requirement for Gr47a is distinct from broadly required Grs

We showed previously that three core-bitter Grs (Gr32a, Gr33a, and Gr66a) are required for strychnine-induced action potentials (Moon et al. 2009; Lee et al. 2010). However, Gr33a mutant flies show normal behavioral repulsion to strychnine over a range of concentrations (Moon et al. 2009). Loss of Gr66a also does not appear to cause defects in strychnine rejection, whereas mutation of Gr32a results in a modest impairment (Lee et al. 2010).

Because our previous studies on Gr66a ^ex83 and ΔGr32a tested only one concentration of strychnine, we reexamined these latter mutants using a range of strychnine levels. We found that only ΔGr32a, but not Gr33a ¹and Gr66a ^ex83, had a relatively small defect in strychnine avoidance (Figure 5A). Thus, Gr47a ¹ was unique as it was the only mutant that displayed a strong deficit in strychnine aversion. Nevertheless, because four GRs contribute to strychnine-induced action potentials, we considered whether these GRs are sufficient to elicit a strychnine response. We misexpressed the four Grs in sugar-responsive GRNs using the Gr5a-GAL4 in combination with UAS-Gr66a, UAS-Gr33a, UAS-Gr32a, and UAS-Gr47a. However, this manipulation did not induce action potentials in sugar-responsive GRNs upon application of strychnine (Figures 5B,C). Because Gr66a, Gr33a, and Gr32a were present in S12, I8, and I9 sensilla (Weiss et al. 2011), we ectopically expressed UAS-Gr47a using the Gr66a-GAL4. However, introduction of Gr47a in these sensilla did not result in significant strychnine-induced action potentials (Figures 5D,E).

Figure 5.

Analyses of strychnine repulsion exhibited by ΔGr32a, Gr33a ¹, and Gr66a ^ex83 flies. (A) Binary food choice assays using 1-mM sucrose versus 5-mM sucrose laced with different concentrations of strychnine. n = 4–7. (B) Ectopic expression of Gr32a, Gr33a, Gr47a, and Gr66a in Gr5a-expressing GRNs using the GAL4/UAS system was insufficient to produce strychnine-induced action potentials. Shown are sample traces from L4 sensilla in response to 1-mM strychnine. (C) Average number of spikes/s in L4 sensilla of control flies (lane 1) or flies expressing UAS-Gr32a, UAS-Gr33a, UAS-Gr47a, and UAS-Gr66a under the control of the Gr5a-GAL4 (lane 2; n = 11). (D) Expression of UAS-Gr47a in I8, I9, and S12 using the Gr66a-GAL4 was not sufficient to generate strychnine-induced action potentials. Sample traces from the indicated sensilla are shown. (E) Average number of spikes/s in the indicated sensilla of flies expressing UAS-Gr47a under the control of the Gr66a-GAL4 (n = 10).

Discussion

We found that Gr47a was required for the response to strychnine and not any of eight other chemicals tested. This observation, combined with prior genetic analyses of other bitter Grs, supports the emerging model that Drosophila GRs fall into two general specificity classes. The first class is comprised of GRs that are broadly required, and this group of core-bitter receptors includes GR32a, GR33a, and GR66a (Moon et al. 2009; Lee et al. 2010). The second specificity class consists of GRs that are narrowly tuned and now includes three members: (1) GR47a, (2) GR93a, which is required for sensing caffeine only (Lee et al. 2009), and (3) GR8a, which is required for the response to the toxic amino acid derivative, L-canavanine (Lee et al. 2012). Nevertheless, it seems very likely that the narrowly tuned receptors, including GR47a, are responsive to more than just one tastant, as the number of GRs expressed in bitter responsive GRNs is limited.

The identification of GRs with broad and narrow specificities is consistent with a comprehensive electrophysiological analysis of the activities of bitter-responsive sensilla (Weiss et al. 2011). According to this study, there are two broadly tuned and two narrowly tuned specificity classes. Moreover, the expression patterns of the Gr reporters largely support the existence of these functional categories, as expression of many Gr reporters is limited to one of the four classes. Exceptions are Gr32a, Gr33a, and Gr66a, which are expressed in all four classes and are broadly required for sensing nearly all bitter compounds.

We conclude that introduction of Gr47a in bitter-responsive GRNs in S3, S5, and S10 was sufficient to restore normal strychnine responsiveness, because we rescued both the behavioral and electrophysiological impairments in the Gr47a ¹ mutant by expression of the wild-type Gr47a transgene in these sensilla. The Gr47a-GAL4 did not appear to reflect the normal expression pattern of Gr47a because it was not expressed in strychnine-activated sensilla and was not effective in rescuing the Gr47a ¹ phenotype in combination with UAS-Gr47a.

With the exception of strychnine, the behavioral avoidance to every deterrent compound tested is greatly impaired by single mutations in at least two, and in most cases, in any of the three core-bitter Grs (Gr32a, Gr33a, and Gr66a). However, strychnine avoidance was normal in flies missing either Gr33a or Gr66a and only moderately reduced in Gr32a mutant flies. Furthermore, misexpression of GR47a with the core receptors, GR32a, GR33a, and GR66a was insufficient to recapitulate a strychnine response in sugar-responsive GRNs. Moreover, introduction of Gr47a in bitter-sensing GRNs that express Gr32a, Gr33a, and Gr66a, but do not normally respond to strychnine, was ineffective at conferring strychnine responsiveness. These latter findings further highlight the complexity of most Drosophila GR complexes.

Supplementary Material

Supplementary material can be found at http://www.chemse.oxfordjournals.org/

Funding

This work was supported by the National Institute on Deafness and Other Communication Disorders (DC007864 to C.M.); the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning [2012M3A9B2052525 to Y.L.]; and the Basic Science Research Program of the NRF of Korea funded by the Ministry of Education [2012R1A1A2003727 to Y.L.].

Conflict of interest

The authors have no conflicts of interest to declare.