Taste response profiles of the labellar chemosensilla of the medfly Ceratitis capitata (Diptera: Tephritidae)

Abstract

The medfly (Ceratitis capitata Wiedemann, 1824) is a widespread pest for horticulture, mainly targeting fruits of pomaceous and citrus cultivars. Scanty data is available about its chemosensory responses to contact salty, sweet, sour and bitter stimuli. High Resolution Scanning Electron Microscopy (HRSEM) and Transmission Electron Microscopy (TEM) observations indicate the presence of six pairs of Long type and about 40 pairs of Intermediate type labellar chemosensilla, each containing four different neurons. Spike activity was electrophysiologically recorded from the labellar sensilla in response to sodium chloride (NaCl), fructose, acids (citric and malic) and bitter compounds (quinine, quercetin and 6-n-propyl-thiouracil (PROP)). By analysis of spike waveforms in the responses, four neurons were identified responding to different stimuli in both sensillum types of C. capitata. The response specificity of three of these cells (named M1, M2 and S) was determined on the basis of their dose-response profiles. “M1” was specific to NaCl, “M2” to fructose. Both acids appeared to excite the “S” cell and to inhibit the “M1” cell. Citric acid also partly inhibited the response of the “M2” cell. No cell clearly responded to any of the bitter stimuli.

Keywords:

• Taste
• medfly
• electrophysiology
• chemoreception
• Tephritidae

Introduction

The Mediterranean  fruit fly or medfly, Ceratitis capitata (Wiedemann 1824) (Diptera: Tephritidae), is a polyphagous pest that infests numerous plant species of agronomic interest, such as pomaceous (apples and pears) and citrus (oranges) cultures, and easily adapts to novel host plants according to their availability (Malacrida et al. 2007).

The reproductive behaviour of the medfly is relatively well studied (Eberhard 2000; Sivinski et al. 2000; Yuval & Hendrichs 2000; Marchini & Del Bene 2006; Gomulski et al. 2012; Harwood et al. 2013). Receptive females are attracted to gatherings (leks) of males emitting sex pheromone (Papadopoulos et al. 2004). The after-mating behaviour of females diverges from that of males, switching to host-fruit locating for oviposition (Jang et al. 1998). Medfly females use different sensory modalities such as sight, smell and taste to identify potential host-plant fruits on which to lay eggs (Fletcher & Prokopy 1991). Ovipositing females mark fruits with a deterrent pheromone to prevent further egg-laying by other females (Arredondo & Díaz-Fleischer 2006). The behavioural strategies underlying male aggregation, mating, selection and localization of host plants are largely conditioned by the neural input arising from olfactory and taste sensilla (Jang et al. 1998). Moreover, the gustatory system represents a model to study the mechanism of food intake, and its chemosensory pathways are considered ideal targets for the development of novel and useful control methods for this species.

Typically, taste neurons in Diptera are bipolar cells: somata are located at the base of the sensillum and dendrites extend into the sensillum shaft up to the tip, whereas axons project into the subesophageal ganglion (Ishimoto & Tanimura 2004; Wang et al. 2004; Amrein & Thorne 2005; Sellier et al. 2011).

On the basis of the morphology of their cuticular surface, supported by ultrastructure and electrophysiology, chemosensilla have been classified in a variety of morphological types (Schneider 1964; Zacharuk 1985) possessing a chemosensory function (Zacharuk & Shields 1991). Taste chemosensilla of flies are typically uniporous and house a variable number (i.e. two to six) of sensory neurons (Adams et al. 1965; Dethier 1976).

In tephritid flies, information about size and distribution of the labellar taste sensilla is limited to a few species (Gothilf et al. 1971; Crnjar & Prokopy 1982; Crnjar et al. 1989). In C. capitata, Gothilf and colleagues (1971) showed that tarsal and labellar chemosensilla respond to salts and sugars.

Pest management of the medfly has mainly being done by the Sterile Insect Technique (SIT) that uses infertile matings of sterile males to wild females with the aim of decreasing its population distribution (Ogaugwu et al. 2013). Tephritid fruit flies have been recently used as laboratory models for aging and biodemography research (Papadopoulos et al. 2010; Carey 2011). Obtaining new information on the chemosensory profiles of the medflies could be helpful in designing novel pest control strategies.

The aim of the present study is to provide information about the taste chemosensory system in the labellum of the medfly, in both sexes, and to describe the response profiles of the labellar sensilla to salty, sweet, sour and bitter taste stimuli.

Materials and methods

Insects

Morphological and electrophysiological experiments were performed on adult male and female medflies 3–5 days old, obtained from a colony reared at the Department of Animal Biology, University of Pavia (Italy). Males and females were kept separate to avoid reciprocal exposure, under the following conditions: 22 ± 1°C and 70–80% relative humidity.

Electron microscopy

Sixteen females and males of C. capitata were treated for transmission electron microscopy (TEM) and high resolution scanning electron microscopy (HRSEM). For HRSEM analysis, C. capitata flies were fixed with a mixture of 0.5% glutaraldehyde and 0.5% paraformaldehyde in 0.1 M cacodylate buffer, pH 7.2, at room temperature for 2 hours. Some specimens were fixed up to 24 hours in order to prevent collapse of sensilla. The medflies were sonicated (3 × 20 s) in 2% Triton X-100 at room temperature, in order to remove any material that would prevent observation of details from the cuticular surface. Finally, insects were washed in a phosphate buffered saline (PBS) (3 × 30 min) at room temperature; specimens were dehydrated through a graded acetone series, and subjected to Critical Point Drying (CPD) with carbon dioxide (CO₂). Samples were immobilized on aluminium electron microscopy stubs and coated with platinum (2 nm) by an Emitech 575 turbo sputtering apparatus. Observations were carried out with a Hitachi S-4000 field emission HRSEM, operated at 15–20 kV. Image acquisitions were obtained by the software Quartz PCI v. 5 (Quartz Imaging Corporation, Vancouver, BC, Canada). Over 500 images were collected by HRSEM at different magnifications. Measurements were obtained on HRSEM images and carried out with Image Tool v. 3.0 for Windows (University of Texas Health Science Centre in San Antonio). For TEM analysis, insects were fixed with a mixture of 1.25% glutaraldehyde and 1% paraformaldehyde in 0.1 M cacodylate buffer, pH 7.2, at room temperature for 2 hours. Specimens were included in epoxy resin and ultrathin sections, stained with uranyl acetate and bismuth subnitrate (Riva 1974), and observed by a JEOL 100S TEM operated to 80 kV.

Electrophysiology

Experiments were performed on medflies of both sexes deprived of food and water for 12 hours before taste stimulation. Room temperature and humidity conditions (22 ± 1°C and 70–80% relative humidity) were kept constant during electrophysiological experiments.

Spike activity from the chemosensory cells of the two types of sensilla, identified through the morphological observations, was recorded by means of the tip-recording technique (Hodgson et al. 1955). All recording operations were carried out by means of micromanipulators under the field of a stereomicroscope. A thin silver/silver chloride (Ag/AgCl) wire serving as a ground electrode was inserted into an isolated head through the “foramen magnum”. The recording electrode, a glass micropipette (tip diameter = 40 μm) filled with the stimulating solution, was placed over the sensillum tip.

A second Ag/AgCl wire connected the stimulating/recording electrode to the input of the electrometer (WPI Duo 773). Recordings were band-pass filtered (0.1–3 KHz), digitized by means of the Axon Digidata 1442A A/D converter (sampling rate: 10 KHz) and stored on a PC for further analyses. Stimulating solutions were applied in increasing concentrations to the sensilla for 2–3 s, and a 3-min interval was allowed between consecutive solutions to minimize adaptation phenomena. Spike firing frequency was analysed in the first second, starting 30 ms after stimulation onset, in order to skip any baseline shift induced by the contact artefact.

In order to avoid any drift in solution concentration due to evaporation, a piece of filter paper was used to draw a small amount of solution from the electrode tip just before each stimulation.

Stimuli

The following stimuli were chosen on the basis of preliminary experiments and/or previous studies (Gothilf et al. 1971): sodium chloride (NaCl) (10, 100, and 500 mM) and fructose (1, 10, and 50 mM), the latter being present in citrus fruits (Rodriguez-Saona et al. 2001). Two acids were selected: malic and citric (0.1, 1 and 20 mM); these are present in medfly host-plants fruits (Bernays et al. 1998).

The following bitter stimuli were also assayed: 6-n-propyl-thiouracil (PROP), a compound present in cruciferous plants, which contains a chemical thiocyanate moiety N − C = S responsible for their bitter taste for humans (Fox 1932; Harris & Kalmus 1949; Guo & Reed 2001); quercetin, a flavonol present in citrus fruits (Drewnowski & Gomez-Carneros 2000; Masek & Scott 2010), and quinine, a natural alkaloid, bitter for humans and a deterrent for insects (Liscia & Solari 2000). All bitter stimuli were tested at the following concentrations: 0.01, 0.1 and 1 mM (Sellier et al. 2011; Weiss et al. 2011).

All these substances were purchased by Sigma-Aldrich and their solutions were prepared using distilled water. NaCl 25 mM was used as a control stimulus and was added to all solutions in order to ensure adequate electrical conductance.

Data analysis and statistics

Active cells were sorted out by performing analysis of spike discharges obtained from the same sensilla with the SAPID Tools software (Smith et al. 1990). Responding cells were identified as higher (“H”), middle 1 (“M1”), middle 2 (“M2”) and small (“S”) units, on the basis of the shape and amplitude of their spike responding to salt, sugar and bitter solutions as described in previous papers (Dethier 1976; Liscia et al. 1998; Liscia & Solari 2000).

Statistical differences among data obtained on HRSEM images were determined by a non-parametric test (Mann-Whitney U-test). Differences in spike frequency were evaluated by means of “two-sided” Student “t” (STATISTICA software package) test with a 95% confidence level (p ≤ 0.05).

Results

Morphology

On the basis of HRSEM observations, mouthpart apparatus of C. capitata were formed by two hemilabella. The labellar external surface shows different types of sensilla trichodea (Figure 1A, B).

Sensilla trichodea have a prominent socket at the base and present a few grooves along the sensillum shaft that converge near the tip. As reported in Drosophila (Shambhag et al. 2001), we have identified three sensillum types: Long (L-type), Intermediate (I-type a and b) and Short, according to their length.

Figure 1. A, scanning electron micrograph of the proboscis labella in Ceratitis capitata showing the chemosensilla on the aboral surfaces. B, schematic diagram of a labellum showing the localization of Long, Intermediate and Short type sensilla. The sensilla outlined in white (Long) and black (Intermediate a) are those used for electrophysiological experiments. Other types of sensilla (Intermediate b and Small) are indicated by solid circles and empty circles, respectively. Scale bar: 66 μm.

Long sensilla trichodea have a length of 174 ± 47 µm (mean length ± S.D.) and a base width of 5.3 ± 0.39 µm and they are located near the base of the labellum. We have consistently counted six Long sensilla in each hemilabellum (Figure 1A, B).

Taking length and position into account, Intermediate sensilla trichodea can be classified in two subtypes: a and b. “a” sensilla show a mean length of 64 ± 17 µm and a base width of 4.3 ± 0.67 µm, and they are located in the region spanning from the Long type area to the labellar inner surface. “b” sensilla are shorter (mean length of 48 ± 10 µm) than the previous ones, and are located both immediately beyond the margin of the labellar lobe protruding in front of the oral cavity and among Long type sensilla. About 40 Intermediate type sensilla are observed on each hemilabellum. No variations in the total number of Intermediate and Long type are observed in either sex or between the two lobes.

Short type sensilla trichodea are scattered among Long and Intermediate type; they number about 20 and have a mean length of 17 ± 4 µm and a base width of 1.95 ± 0.22 µm.

By TEM, cross sections of Long and Intermediate type sensilla trichodea show an external cuticle and a circular lumen, which contains four dendrites of different diameter (n = 26) (Figure 2). On the basis of their morphology, Long and Intermediate type sensilla have been assigned to the uniporous type.

Figure 2. Transmission electron microscopy (TEM) image of a cross section of a labellar sensillum trichodeum in Ceratitis capitata. Four dendrites with similar diameters are visible within the sensillum shaft. Scale bar: 1 μm.

Electrophysiology

By the electrophysiological approach, salty, sweet, sour and bitter taste stimuli were tested in Long and Intermediate type sensilla, both in males and females. Figure 3 shows a sample discharge from the two types of sensilla in response to NaCl and fructose. In response to these stimuli, we noted that spike activity is produced by different neurons. By analysing spike waveforms, we were able to distinguish four types of neurons that were named by their apparent amplitude as high (H), middle 1 (M1), middle 2 (M2), and small (S) in response to salt and sweet (Figure 3).

Figure 3. Samples of spike discharges from the (a) Intermediate and (b) Long type sensilla in Ceratitis capitata responding to 50 mM fructose and 500 mM sodium chloride (NaCl) respectively. Portion of the discharges enlarged in the trace underscored. High (H), middle 1 (M1), middle 2 (M2), and small (S) spike types are shown.

During stimulation with different concentrations (10, 100 and 500 mM) of NaCl, in both Long and Intermediate type sensilla of females and males, in response to increasing salt concentration, only “M1” cell enhances its activity. In particular, Long type sensilla elicit the highest spike activity of “M1” cells with a peak of response of 28.14 ± 3.54 and 18.30 ± 2.72 spikes/s at 500 mM NaCl, respectively, in females and males (Figure 4A, C). Similarly, Intermediate type sensilla increased the activity of “M1” cells with a peak of response of 40.88 ± 4.24 and 34.59 ± 4.42 spikes/s at 500 mM NaCl, respectively, in females and males (Figure 4B, D). For the “M1” cells of both sensillum types, all frequency mean values were statistically different from one another (p ≤ 0.05), with the exception of the responses in Long type sensilla of both sexes to NaCl between 100 and 500 mM (Figure 4A, C).

Figure 4. Dose-response histograms following stimulation with 10, 100 and 500 mM sodium chloride (NaCl) in the Long and Intermediate type sensilla of Ceratitis capitata. A–B, spike firing frequencies (spikes/1^st s) from high (H), middle 1 (M1), middle 2 (M2) and small (S) cells in Long and Intermediate type sensilla of females. Mean values (26–40 sensilla) ± SE (vertical bars); C–D, spike firing frequencies (spikes/1^st s) from high (H), middle 1 (M1), middle 2 (M2) and small (S) cells in Long and Intermediate type sensilla of males. Mean values (31–59 sensilla) ± SE (vertical bars); “a” indicates significance difference (p < 0.05) in “M1” cell following stimulation with 10 and 100 mM NaCl of Long type sensilla in females; “b” and “c” indicate significant difference (p < 0.05) in “M1” cell of Intermediate type sensilla in females following stimulation with 10, 100 and 500 mM NaCl; “d” indicates significant difference (p < 0.05) in “M1” cell of Long type sensilla in males following stimulation with 10 and 100 mM NaCl; “e” and “f” indicate significant difference (p < 0.05) in “M1” cell of Intermediate type sensilla in males following stimulation with 10, 100 and 500 mM NaCl.

When the same sensilla are stimulated with 1, 10 and 50 mM fructose, a different cell, “M2”, shows an activity increase (Figure 5A–D). In Long type sensilla of females and males, the results show that following stimulation with fructose, “M2” cells give the largest contribution with mean values (± SE) of 44.58 ± 3.64 and 39.62 ± 3.34 spikes/s, in females and males respectively (Figure 5A, C). Instead, in Intermediate type in female and males, increasing the fructose concentration enhances significantly (p ≤ 0.05) the activity of “M2” cells in a dose-dependent manner (from 8.32 ± 1.84 to 29 ± 2.95 spikes/s for females and from 6.57 ± 2.59 to 40.57 ± 5.07 spikes/s for males) (Figure 5B, D). In response to fructose, as observed for salt, all frequency values are statistically (p ≤ 0.05) different from one another for “M2” cells for both Long and Intermediate sensillum types, with the exception of the responses to 10 and 50 mM fructose of Long type of both sexes (Figure 5A, C).

Figure 5. Dose-response histograms following stimulation with 1, 10 and 50 mM fructose in the Long and Intermediate type sensilla of Ceratitis capitata. A–B, spike firing frequencies (spikes/1^st s) from high (H), middle 1 (M1), middle 2 (M2) and small (S) cells in Long and Intermediate type sensilla of females. Mean values (77–86 sensilla) ± SE (vertical bars); C–D, spike firing frequencies (spikes/1^st s) from high (H), first intermediate (M1), second intermediate (M2), and small (S) cells in Long and Intermediate type sensilla of male. Mean values (23–70 sensilla) ± SE (vertical bars); “a” and “b” indicate significant difference (p < 0.05) in “M2” cell of Long type sensilla in both sexes following stimulation with 1 and 10 mM fructose; “c” and “d” indicate significant difference (p < 0.05) in “M2” cell of Intermediate type sensilla in female following stimulation with 1, 10, 50 mM fructose; “e” and “f” indicate significant difference (p < 0.05) in “M2” cell of Intermediate type sensilla in males following stimulation with 1, 10, 50 mM fructose.

Following stimulation with malic acid at 0.1, 1 and 20 mM in Long type sensilla, the “S” cell increases its activity with respect to the control (25 mM NaCl), from 1.92 ± 0.72 to 4.87 ± 1.03 spikes/s in females and from 2.52 ± 0.90 to 10.53 ± 1.76 spikes/s in males (Figure 6A, B). In addition, malic acid, at both 1 mM and 20 mM, produces an activity decrease in “M1” cells with respect to the control (p < 0.05). On the other hand, citric acid at 0.1, 1 and 20 mM enhances the spike firing frequency of “S” cells from 1.70 ± 0.92 to 14.60 ± 1.75 spikes/s in females and from 1.66 ± 0.45 to 10.17 ± 1.22 spikes/s in males (Figure 7A, B). Besides, citric acid (1 mM and 20 mM) decreases the activity of “M1” cells with respect to the control (p < 0.05). Moreover, 20 mM citric acid decreases the activity of “M2” cells.

Figure 6. Dose-response histograms following stimulation with 0.1, 1 and 20 mM malic acid in the Long type sensilla of Ceratitis capitata. A–B, spike firing frequencies (spikes/1^st s) from high (H), middle 1 (M1), middle 2 (M2) and small (S) cells in Long type sensilla of female and males respectively. Mean values (15–40 sensilla) ± SE (vertical bars); “a” and “d” indicate significant differences (p < 0.05) in “S” cell between 0.1 mM malic acid and the control 25 mM sodium chloride (NaCl) in females and males respectively; “b” and “e” indicate significant differences (p < 0.05) in “S” cell 1 mM malic acid and the control 25 mM NaCl in females and males respectively; “c” and “f” indicate significant differences (p < 0.05) in “S” cell 20 mM malic acid and the control 25 mM NaCl in females and males respectively; “*” indicates significant difference (p < 0.05) in “M1” cell between 1 mM malic acid and the control 25 mM NaCl in female and male; “**” indicates significant difference (p < 0.05) in “M1” cell between 20 mM malic acid and the control 25 mM NaCl in females and males.

Figure 7. Dose-response histograms following stimulation with 0.1, 1 and 20 mM citric acid in the Long type sensilla of Ceratitis capitata. A–B, spike firing frequencies (spikes/1^st s) from high (H), middle 1 (M1), middle 2 (M2) and small (S) cells in Long type sensilla of female and males respectively. Mean values (40–60 sensilla) ± SE (vertical bars); “a” and “c” indicate significant differences (p < 0.05) in “S” cell between 1 mM citric acid and the control 25 mM sodium chloride (NaCl) in females and males respectively; “b” and “d” indicate significant differences (p < 0.05) in “S” cell 20 mM citric acid and the control 25 mM NaCl in females and males respectively; “y” and “z” indicate significant differences (p < 0.05) in “M2” cell 20 mM citric acid and the control 25 mM NaCl in females and males respectively; “*” indicates significant difference (p < 0.05) in “M1” cell between 1 mM citric acid and the control 25 mM NaCl in females and males; “**” indicates significant difference (p < 0.05) in “M1” cell between 20 mM citric acid and the control 25 mM NaCl in females and males; “***” indicates significant difference (p < 0.05) in “M1” cell between 0.01mM citric acid and the control 25 mM NaCl in females.

Bitter stimulation was performed, in Long and Intermediate sensilla types of both sexes, using quinine, PROP and quercetin at the same concentrations (0.01, 0.1 and 1 mM). Our data show that no significant responses were obtained from any of the four neurons with respect to the control (25 mM NaCl) (Figure 8).

Figure 8. Samples of spike discharges from the Long type sensilla in Ceratitis capitata responding to 25 mM sodium chloride (NaCl), 0.1 mM 6-n-propyl-thiouracil (PROP), 0.1 mM quinine and 0.1 mM quercetin, respectively. Traces show spikes during the first second after stimulation.

Discussion

Our data provide a morpho-functional description of the taste chemosensory system located in the labellum of C. capitata, that plays a role in mediating food ingestion and the choice of oviposition sites (Gothilf et al. 1971; Arredondo & Díaz-Fleischer 2006; Coronado-Gonzales et al. 2008).

The labellum is characterized by a large population of sensilla trichodea housing taste receptor neurons and located in specific areas. In calliphorid flies, the labellar apparatus presents several subtypes of sensilla trichodea with a specific morphometry and distribution on the aboral surfaces: in Phormia regina, Wilczek (1967) described six types. In Tephritid flies, available information on sensillum types and distribution is limited to a few species (Gothilf et al. 1971; Crnjar & Prokopy 1982; Crnjar et al. 1989).

Drosophila melanogaster is instead a valuable reference species, as a large number of studies exist describing its morphology and taste physiology (Hiroi et al. 2002; Montell 2009; Sellier et al. 2011; Weiss et al. 2011). In Drosophila, the labellar apparatus shows three types of chemosensory sensilla, classified into Long, Intermediate and Short (Shambhag et al. 2001; Hiroi et al. 2002; Montell 2009; Sellier et al. 2011). Long and Short types contain one mechanoreceptor and four gustatory neurons that respond to water, sugar, low salt or high salt concentration (Rodrigues & Siddiqi 1981; Hiroi et al. 2002; Sellier et al. 2011). On the other hand, Intermediate type sensilla contain only two neurons: one responds to sugar and low salt concentration, while the other responds to high salt concentration and bitter substances (Hiroi et al. 2004). Weiss et al. (2011) described five classes of sensilla in Drosophila, on the basis of response to bitter stimuli; four were activated by bitter and one did not elicit any physiological response to bitter.

In the medfly, labellar sensilla trichodea are numerous and, as in Drosophila melanogaster, we have assigned them to four different groups: Long, Intermediate (a and b) and Short (Shambhag et al. 2001). According to HRSEM images, each sensillum type is characterized by its own length range and distribution on labellar surface. Moreover, TEM images show the presence of four neurons associated with each sensillum. The number of neurons in each sensillum may vary in different insects: in Drosophila, Long type sensilla house four gustatory receptor neurons and one mechanoreceptor, and Intermediate type sensilla only have two gustatory receptor neurons and one mechanoreceptor (Shambhag et al. 2001); labellar and tarsal hairs in Phormia regina contain four gustatory neurons and a mechanoreceptor (Dethier 1976).

In order to analyse the taste responses of the labellar chemosensilla in both sexes of the medfly, we described the spike activity of Long and Intermediate type sensilla evoked by salty, sweet, sour and bitter stimuli; we overlooked Short type sensilla as their short size and curved shape make electrophysiological access difficult. By cross-comparing spike amplitudes and shapes, our data confirm the presence of four neurons in C. capitata.

NaCl stimulation evoked, in Long and Intermediate type sensilla of both sexes, an increase of “M1” neuron activity. In particular, “M1” spike activity, in Intermediate type, increases significantly after stimulation with 100 and 500 mM NaCl; on the contrary, Long type spike frequency does not change between these concentrations. For fructose stimulation, a different cell (M2) increases its activity. In fact, in Long type sensilla, “M2” cell spike frequency does not change after stimulation with 10 and 50 mM fructose while, in Intermediate type, this value increases significantly.

Electrophysiological responses, after NaCl and fructose stimulation, indicate that Long type sensilla, in both sexes alike, probably reached response saturation at the highest value tested. In contrast, no saturation was observed in Intermediate type sensilla. This difference could provide space-related information on the localization of stimuli by using differential information from sensilla with different sensitivity profiles and distributions. In fact, fructose is important for the medfly not only as a high energy source, but also for recognizing suitable oviposition sites in host plants (Gothilf et al. 1971). Fructose is the most common monosaccharide present in medfly host plants (Rodriguez-Saona et al. 2001). On the other hand, salts are essential to regulate electrolyte homeostasis (Lindemann 1996; Hiroi et al. 2004) and have a phagostimulant effect (Smedley & Eisner 1995). Differences in the sensitivity threshold of the various chemosensillum types in response to the same stimuli have already been shown in other insects. In Protophormia terraenovae, Long and Intermediate sensilla have different response profiles to NaCl and sucrose. This suggests that different types of sensilla located on the rostral and caudal ends of the labellum provide a spatial representation of the taste stimuli (Liscia et al. 1995, 1998).

Citric and malic acid evoked both an increase in “S” cell response and a decrease in the response of “M1” cells in the two sexes. In this respect, acids have been shown to inhibit the activity of salt taste neurons in the blowfly Phormia regina (McCutchan 1969). The “S” cell was activated by both acids, but we do not exclude that it may respond to other stimuli as well, in agreement with Goldrich (1973) who reported that blowflies can detect acids, but there is no evidence of the presence of a specific “acid” receptor: the responses to acids are mediated by the activity of the other cells (“salt”, “sugar” and “water”). The fact that the two acids inhibit the “M1” cell also supports what was suggested in Drosophila melanogaster, where acid stimuli both activate the bitter neurons and inhibit the sweet neurons (Charlu et al. 2013). In fact, citric acid also seems to inhibit to some extent the “M2” cell. Charlu et al. (2013) suggest that a specific population of  “sour” cells may not exist in flies, with acidity evoking excitatory responses in a defined subset of bitter cells.

The juice of apples and citrus fruits, common host plants in the medfly, contains fructose, sucrose, glucose, sorbitol, malic, citric, fumaric acids, sodium and calcium; malic acid is the major acid of apple juice at a concentration of 1738.2 mg/100 mL (Eisele & Drake 2005). Moreover, citric acid has been detected in different citrus fruits: at the highest concentration in grapefruit juice (64.7 mmol/L), and in lower concentrations in lemon juice (47.66 mmol/L), orange juice (47.36 mmol/L) and pineapple juice (41.57 mmol/L) (Haleblian et al. 2008). Acid stimuli are commonly found in overripe or rotting fruits that are common food sources for medflies.

To study the response to bitter stimuli in C. capitata, we used three substances with different molecular structures: PROP (a thiouracil-derived bitter compound), quinine (a natural alkaloid) and quercetin (a bitter flavonol present in citrus fruits). None of them elicited any response from each of the four neurons in both Long and Intermediate type sensilla. In the light of our data, we assume that labellar sensilla do not respond to deterrent stimuli. This fact is not so unusual as previous studies in Drosophila revealed that Long type sensilla respond to salt and sweet but not to bitter stimuli, while Intermediate and Short type sensilla are activated by only a few bitter substances (Weiss et al. 2011). However, one cannot exclude that bitter sensitivity may be confined to Short type sensilla, which we did not test, or to sensilla located in other body regions, such as the tarsi (Chapman 2003; Ishimoto & Tanimura 2004), the ovipositor (Crnjar et al. 1989) or the interpseudotracheal papillae (Coronado-Gonzales et al. 2008).

The input from the tarsal sensilla is known to regulate the first phase of feeding behaviour (Dethier 1976; Singh 1997). In Drosophila, the sensitivity to bitter stimuli is different in labellar and tarsal sensilla: tarsi are known to respond preferentially to quinine or berberine, while strychnine is detected by the labellar sensilla (Meunier et al. 2003). This effect suggests that the different input from the labellar and tarsal apparatus could control feeding behaviour. In our experiments, we only tested three bitter stimuli; given the great variety of bitter compounds, characterized by very different chemical structures and specific transduction mechanisms, we cannot rule out the possibility that other bitter substances may stimulate the labellar chemosensilla of C. capitata.

Besides, during specific steps of the medfly life cycle such as oviposition and mating, so-called bitter compounds may not have a toxic effect or even an attractive significance, as reported in caterpillars and in Drosophila (Yang et al. 2008; Singer et al. 2009; Weiss et al. 2011). Indeed, citrus fruits should be little infested by the medfly due to toxic effects of compounds in the flavedo and the chemical properties of essential oils (Salvatore et al. 2004). Furthermore, limonene, a common compound in all citrus fruits, has no deterrent effects but stimulates oviposition, while linalool, a chemical compound of immature citrus fruits, has a deterrent effect (Ioannou et al. 2012).

Finally, we identified the response profiles for three (“M1”, “M2” and “S”) of the four cells housed in both chemosensillum types of both sexes, on the basis of their dose-response curves: “M1” is specific to NaCl and “M2” to fructose. The “S” cell is somewhat activated by acids, but this does not necessarily imply a specificity of response to “sour”. Indeed, the sensitivity to acids seems to be somewhat “distributed” as excitation and inhibition of different cells, as suggested for other flies (McCutchan 1969; Goldrich 1973; Charlu et al. 2013). Instead, no response specificity was detected for the remaining unit, the “H cell”: it could be tuned on such compounds as the oviposition-deterring pheromone. Further electrophysiological and behavioural studies are necessary to elucidate the sensitivity to bitter stimuli in the medfly.

Acknowledgements

The authors wish to thank Dr. Marco Melis for his helpful cooperation and technical support in the experimental part of this study. We thank Mr. Alessandro Cadau for photography and image treatment assistance and the staff of Prof. Anna Malacrida, Department of Animal Biology, University of Pavia for the help in rearing the medflies. This work was supported by the Italian “Ministero dell’Istruzione, dell’Università e della Ricerca” (MIUR).