Background odour induces adaptation and sensitization of olfactory receptors in the antennae of houseflies


Dr F. J. Kelling, Department of Animal Physiology, University of Groningen, P.O. Box 14, 9750 AA Haren, the Netherlands. E-mail: or


Abstract The presence of background odour was found to have a small but significant effect on the sensitivity of the antennal olfactory system of houseflies, Musca domestica Linnaeus (Diptera: Muscidae), to new pulses of odour. We show that cross-adaptation and cross-sensitization between a background odour of (±)-1-octen-3-ol and pulses of (±)-1-octen-3-ol, 2-pentanone and R-(+)-limonene can occur, confirming that olfactory receptor cells are sensitive to different odours. Background odour can increase the responses to low concentration odour pulses and decrease the responses to higher concentration odour pulses. It is suggested that background odour has a larger effect on olfactory receptor cells that respond with a tonic increase of spike frequency, giving information about the level of odour concentration, i.e. the ‘static’ environment. Cells that respond in a phasic way only provide information on the dynamics of the olfactory environment.


The housefly, Musca domestica L., is one of the most familiar nuisance pests of human and livestock habitations and constitutes major hygienic and economic problems in a variety of industries, such as cattle and poultry farms and food processing industries (Hansens, 1963). Flying between various food sources, houseflies can act as vectors of many diseases, such as dysentery, gasteroenteritis, cholera and tuberculosis (Grübel et al., 1997; Tan et al., 1997; Fotedar, 2001). Therefore, methods have been developed to control housefly populations. Some methods of control are based on olfactory cues, luring houseflies into traps with attractive odours, or repelling them with odours that are unpleasant to them (Mulla et al., 1977). To identify chemical substances that can be used as baits or repellents for houseflies, several behavioural and electrophysiological studies have been carried out (Frishman & Matthysse, 1966; Mulla et al., 1977; Warnes & Finlayson, 1986; Cossé & Baker, 1996). Most of these studies were performed in the laboratory in which test chemicals were added to clean air to observe the responses of the flies to these chemicals. However, in nature, air always contains some ambient background odour. For attractants or repellents to be effective, flies have to be able to distinguish these chemicals from the ambient odours. Studies on other arthropods have shown that the electrophysiological responses to mixtures of chemicals may be suppressed compared to the sum of the responses to the individual compounds (Derby et al., 1985; Voskamp et al., 1999). These studies on mixtures were also performed in an odour-free environment.

In the present study we investigated the influence of continuously present ambient odours on the electrophysiological responses to known attractive and repellent odours. We compared the sensitivities of antennal olfactory cells to odours added to clean air and applied in the presence of either a synthetic ((±)-1-octen-3-ol) or a natural (chicken manure) background odour. We recorded electroantennograms (EAGs) and spike responses of individual antennal cells on stimulation with (±)-1-octen-3-ol, 2-pentanone and R-(+)-limonene. (±)-1-Octen-3-ol is known to be an attractant for several dipteran species (Hall et al., 1984) and R-(+)-limonene is repellent to houseflies (J. R. Moskal, Denka International, pers. comm.). In preliminary studies, 2-pentanone was found to be electrophysiologically active (Kelling, 2001).

Materials and Methods


Pupae of a Musca domestica WHO strain Ij2 were obtained from the Danish Pest Infestation Laboratory (Lyngby, Denmark). Cultures of the strain were kept in the laboratory at 25°C and 75% r.h. Larvae were reared in an aqueous jelly of agar, yeast and skim milk powder (1 : 5 : 5 by weight) and allowed to pupate in wood curls on top of the jelly. Flies were fed a mixture of milk powder, sugar and autolysed yeast (5 : 5 : 1 by weight) and had access to tap water. Experimental flies were 4–20-day-old (mature) females that had mated.


An intact fly was immobilized in a plastic pipette, with its head protruding from the tip, so that its antennae were accessible for recording (Den Otter et al., 1988). Glass micropipette/Ag–AgCl electrodes filled with Beadle–Ephrussi saline were used. For both electro-antennogram (EAG) and single cell (SC) recordings the tip (c. 25 µm) of the indifferent glass electrode was inserted into the head of the fly between the eyes, near the base of the antennae. EAG recordings were made by placing the tip (25–100 µm) of a recording electrode over the distal end of the antenna. Single cell recordings were obtained using the surface-contact technique (Den Otter et al., 1980). By gently pressing the tip of the recording electrode (< 5 µm) against the cuticle of the funiculus, spikes from individual receptor cells could be recorded. When electrical contact with the antenna was made, the preparation was left for 5 min before starting an experiment.

The electrodes were connected to a high-impedance AC/DC amplifier (Syntech, Hilversum, the Netherlands). The EAG-signal was directly fed into a computer. The SC-signal was displayed on an oscilloscope, made audible via a speaker and stored on tape with a DAT-recorder (Sony, Japan) for later analysis.


Chemicals used to prepare test stimuli were obtained from Fluka ((±)-1-octen-3-ol, > 98%, 2-pentanone, > 98%; Fluka, Zwijndrecht, The Netherlands) and Denka (R-(+)-limonene, > 98%; Denka, Barneveld, The Netherlands). The chemicals were dissolved in silicon oil, 0.4 g/ml solutions and four decadic dilutions were made. Aliquots of 25 µL of each solution were pipetted onto a filter paper (6 × 35 mm2) in a Pasteur pipette. The pipettes served as odour cartridges and thus contained 0.001, 0.01, 0.1, 1 or 10 mg doses of the chemicals.

A constant flow (A: 3 mL/s) of charcoal-filtered, humidified air was led through a clean empty bottle or through a bottle with the background odour source. The latter bottle contained 5 g fresh chicken manure, or a filter paper loaded with 250 µl silicon oil containing 1, 10, 100 mg (±)-1-octen-3-ol, or loaded with 1250 µL (1000 mg) pure (±)-1-octen-3-ol. This clean or background odour-loaded air was added to an airstream (B: 10 ml/s) continuously passing over the antennae through a stainless steel tube (i.d. 7 mm), the outlet of which was about 5 mm from the preparation. Stimulation with test odour occurred by injecting, in 0.2 s, 2 mL vapour (C: 10 mL/s) from an odour-loaded pipette into the airstream through an aperture in the tube (3 mm diameter), 8 cm from its outlet. The complementary airstream (B) was switched off during the stimulus by a stimulus controller (Syntech) to maintain the total flow over the antennae at 13 mL/s. Concentrations of the background odour in the bottles and the test stimuli in the odour cartridges were determined using a Shimazu GC-7A gas chromatograph, and the resulting diluted concentrations in the air at the site of the antennae were calculated. Odours were presented at 60 s intervals in random sequence, in (1) clean air, (2) background odour-loaded air and (3) again in clean air. Between (1) and (2) the preparation was flushed with background odour for at least 15 min, and between (2) and (3) with clean air for at least 15 min.


EAGs were analysed using the software package EAG V2.6c (Syntech). To compare the EAG-responses of different experiments, the EAG amplitudes were normalized. First, the responses were calculated as a fraction of the mean response to 10 mg (±)-1-octen-3-ol for each experiment; then this fraction was multiplied by the mean EAG-response (in mV) to 10 mg (±)-1-octen-3-ol of all experiments.

Action potentials were analysed using the software package AutospikeTM (Syntech). Spikes from different cells were distinguished by their amplitudes. It was uncertain at what moment the bulk of the stimulus reached the recording site after injection. Therefore, the response magnitude was calculated from the maximum number of spikes in a sliding 0.1 s period of the response, and expressed as spikes/s. The mean non-stimulated activity (spikes/s) during the 3 s period prior to the beginning of the stimulation was determined and subtracted.

Only cells that responded to the test odours in a dose-dependent manner were used for further analysis. A cell was considered to respond in a dose-dependent manner when the responses to the highest two doses of a chemical tested were larger than the responses to the lower doses applied.


Figures 1–3 show the EAG dose–response series of the three test odours when applied in clean air and in four different background concentrations of (±)-1-octen-3-ol. For a quick screening of the effect of a background odour on the sensitivity of the antennal olfactory system, repeated measurements anova tests were performed. No significant effects of the two lower background concentrations of 2.2 × 10−8 m and 9.5 × 10−8 m (±)-1-octen-3-ol were found. In both the absence and presence of these concentrations the responses to the test odours were the same (Figs 1, 2 and 3A, B); however, when background concentrations of 1.7 × 10−7 m and 2.4 × 10−7 m (±)-1-octen-3-ol were present, the shape of the dose–response curves differed significantly from those in clean air. Repeated measurements anova showed significant interactions of dose with the presence or absence of 1.7 × 10−7 m (±)-1-octen-3-ol background odour for (±)-1-octen-3-ol (P = 0.008, F8,6 = 2.9) and for 2-pentanone (P = 0.001, F8,6 = 4.0) and with the presence or absence of 2.4 × 10−7m (±)-1-octen-3-ol background odour for (±)-1-octen-3-ol (P < 0.001, F8,4 = 5.6) and for R-(+)-limonene (P = 0.014, F8,4 = 2.9). Therefore, we considered the effects of these background concentrations in more detail.

Figure 1.

 EAG dose–response curves of (±)-1-octen-3-ol when flushed with clean air (‘before’), with background (±)-1-octen-3-ol odour (‘test’) and again with clean air (‘after’). The concentrations of background (±)-1-octen-3-ol are indicated in each graph. Error bars represent SEM. Significant differences (P < 0.05) between responses during background odour and clean air ‘before’ (*) and ‘after’ (#) are indicated.

Figure 2.

 Dose–response curves of 2-pentanone when flushed with clean air (‘before’), with background (±)-1-octen-3-ol odour (‘test’) and again with clean air (‘after’). See Fig. 1 for details.

Figure 3.

 Dose–response curves of R-(+)-limonene when flushed with clean air (‘before’), with background (±)-1-octen-3-ol odour (‘test’) and again with clean air (‘after’). See Fig. 1 for details.

A background concentration of 1.7 × 10−7 m (±)-1-octen-3-ol (Figs 1, 2 and 3C) increased the responses to the lowest two concentrations of (±)-1-octen-3-ol and to the lowest three concentrations of R-(+)-limonene compared to the responses in clean air. In contrast to this, the background concentration decreased the responses to the higher concentrations of (±)-1-octen-3-ol and 2-pentanone. After switching back to clean ambient air, some recovery occurred, the responses to the test substances becoming higher again. Interestingly, a 1.7 × 10−7 m background of (±)-1-octen-3-ol did not inhibit the responses to the higher concentrations of R-(+)-limonene.

Flushing the preparation with a background concentration of 2.4 × 10−7 m (±)-1-octen-3-ol, the EAG responses elicited by the highest two concentrations of (±)-1-octen-3-ol and the highest concentration R-(+)-limonene were decreased (Figs 1, 2 and 3D). After switching back to clean air, no significant recovery of the EAG responses occurred.

In single cell experiments we only tested the effects of the two intermediate concentrations of background (±)-1-octen-3-ol: 9.5 × 10−8 m and 1.7 × 10−7 m. In addition, we used the odour of chicken manure as a natural background odour. This odour had been proven to evoke responses in antennal cells (Fig. 4).

Figure 4.

 Example of a spike response to manure. Horizontal bar indicates the duration of the stimulus (0.5 s).

It appeared that on application of background (±)-1-octen-3-ol two different types of single cell responses could be observed: some cells showed a short phasic, others a sustained tonic increase of spike activity (Figs 5A, B). In the latter cells, spike activity returned to the initial level after switching back to clean air (Fig. 5C). No changes in spike frequency were observed in phasic cells while switching back to clean air. Most cells responding to (±)-1-octen-3-ol also responded to the test substances 2-pentanone and R-(+)-limonene. Some cells were only sensitive to (±)-1-octen-3-ol and 2-pentanone, others to (±)-1-octen-3-ol and R-(+)-limonene, and one cell responded to (±)-1-octen-3-ol only.

Figure 5.

 Examples of 50 s spike activity during onset and offset of background odour. Horizontal bars indicate the presence of background odour. Each column in the lower graphs shows the number of spikes in the corresponding second. (A) Cell responding with a short phasic response to the onset of 1.7 × 10−7 m (±)-1-octen-3-ol. (B) Cell responding to the onset of 9.5 × 10−8 m (±)-1-octen-3-ol with a tonic response, that lasted all the time (0.5 h) that background odour was present, after which (C) the spike frequency returned to the initial level.

Six cells showed a tonic response and two cells a phasic response during continuous stimulation with 9.5 × 10−8 m (±)-1-octen-3-ol background odour. The initial non-stimulated activity of all eight cells did not differ (tonic: 9.0 ± 6.6 spikes/s; phasic: 9.1 ± 0.7 spikes/s, Table 1). During the presence of the background odour, the spike frequency of the tonic cells increased significantly to 20.1 ± 9.0 spikes/s (P = 0.03, Paired Wilcoxon test), whereas the spike frequency of the phasic cells stayed at 7.3 ± 4.8 spikes/s after a short burst at the onset of the background odour.

Table 1.   Average spike frequencies ± SD (spikes/s) of tonically and phasically responding cells in clean air (‘before’), during background odour (‘during’), and after switching off the background odour, again in clean air (‘after’). 1o3 = (±)-1-octen-3-ol. The number of cells tested is indicated between brackets.
Background odourBeforeDuringAfter
  • *

    Indicates a significant difference (P=0.031, paired Wilcoxon test).

9.5·10−8m 1o3
 Tonic cells (6)9.0±6.6*20.1±9.0*14.3±9.2
 Phasic cells (2)9.1±0.77.3±4.87.4±3.9
1.7·10−7m 1o3
 Phasic cells (8)11.6±6.58.9±6.59.0±7.4
5 g manure
 Phasic cells (10)8.8±8.08.8±6.88.4±6.7

On continuous application of 1.7 × 10−7 m (±)-1-octen-3-ol or the odour of chicken manure as background odour, no tonic responses were observed: all cells tested showed a phasic response to these odours, after which the spike frequencies returned to initial values (Table 1).

The 9.5 × 10−8 m (±)-1-octen-3-ol background odour affected the responses of the two phasic cells to the test (±)-1-octen-3-ol in a manner comparable to the six tonic cells: by lowering the responses of the cells to the higher two test doses of (±)-1-octen-3-ol compared to the responses in clean air. The two phasic cells did not respond to 2-pentanone or R-(+)-limonene. Figure 6 shows the responses of the tonic cells only.

Figure 6.

 Single cell measured dose–response curves of (±)-1-octen-3-ol (A, B), 2-pentanone (C, D) and R-(+)-limonene (E, F) when flushed with clean air (‘before’), with background (±)-1-octen-3-ol odour (‘test’) and again with clean air (‘after’). See Fig. 1 for details.

Repeated measurements anova tests indicated no significant differences between the dose–response series in clean air and during background odour for any test odour (Figs 6 and 7). An inhibitory effect of the 9.5 × 10−8m (±)-1-octen-3-ol on the responses to the higher test doses can be seen, but this was not significant for any single concentration in clean air compared with the response in background odour. The 1.7 × 10−7 m (±)-1-octen-3-ol and the odour of chicken manure as background odour did not inhibit the responses to the test odours.

Figure 7.

 Dose–response curves of (±)-1-octen-3-ol (A), 2-pentanone (B) and R-(+)-limonene (C) when flushed with clean air (‘before’), the odour of 5 g chicken manure (‘test’) and again with clean air (‘after’). See Fig. 1 for details.


The presence of background odour may affect the sensitivities to additional pulses of odour. This effect of background odour was more pronounced in the EAG experiments than in the single cell experiments. The EAG experiments showed no effect of the 2.2 × 10−8 m and 9.5 × 10−8m background (±)-1-octen-3-ol on the responses to the test odours (±)-1-octen-3-ol, 2-pentanone and R-(+)-limonene. However, in the presence of 1.7 × 10−7m background (±)-1-octen-3-ol the responses to the lower doses of (±)-1-octen-3-ol and R-(+)-limonene were enhanced, whereas the responses to some of the higher doses of (±)-1-octen-3-ol and 2-pentanone were inhibited, compared to the responses in clean air. During stimulation with 2.4 × 10−7 m background (±)-1-octen-3-ol the responses were lower than before in clean air and did not reach their original level after the background odour was switched off and replaced by clean air again.

These effects may be explained by considering the stimulus transduction processes of odours as summarized by Kaissling (1996). Odour molecules diffuse through the cuticular pores into the sensillum lymph. Odorant binding proteins (OBPs) may transport the odour molecules through the sensillum lymph towards the odourant receptors in the dendrites of the olfactory cells (Vogt & Riddiford, 1986; Van den Berg & Ziegelberger, 1991). Upon binding of the odourant-OBP complex to the receptors, the latter are activated. Then ion channels open, depolarizing the dendrite of the olfactory cell. The ion channels may be directly coupled to the receptor molecules, or may be activated via the IP3 or cAMP second messenger system (Krieger et al., 1997). Upon depolarization of the dendrite, the soma of the receptor cell generates action potentials that are sent via the axon to the brain. For deactivation of the odour, it was proposed that the odourant-OBP complex is rapidly oxidized, possibly by the receptor (Ziegelberger, 1995). This oxidized odourant-OBP complex is no longer able to stimulate receptor molecules. Finally, the odour is degraded by enzymes (Vogt & Riddiford, 1986; Rybczynski et al., 1990).

The effect of sensitization or facilitation found in our study at low doses of test odour may be the result of the fact that in the presence of background odour the olfactory cell membrane is more depolarized and is thus discharging action potentials at a higher rate than in the non-stimulated situation. The extra stimulus of test odour will increase the depolarization and spike frequency even more. When a high-dose odour puff is presented in high-dose background odour, the response to this puff on top of the response to background odour is lower than in clean air. Competition for odourant binding proteins and receptor sites may occur. The capacity of receptor sites approaches saturation and excess odour molecules might be shunted towards degrading enzymes and away from receptors. Therefore, at high test concentrations, the response may be lower during the presence of background odour than in clean air.

On continuous stimulation with 9.5 × 10−8 m background (±)-1-octen-3-ol odour, some cells responded with a tonic increase in spike frequency, whereas others showed a short phasic increase of spike frequency at the onset of the background odour, quickly returning to the initial spike frequency. These two types of response were also found in other insect species, e.g. moths (Almaas et al., 1991; Heinbockel & Kaissling, 1996) and tsetse flies (Den Otter & Van der Goes van Naters, 1992; Voskamp et al., 1998). The cells showing phasic responses and quickly adapting to odour may be considered to respond to the dynamic increase of odour concentration. The cells that keep a tonic higher spike frequency during the stimulus provide information on both the dynamic and the static phase of odour stimulation. Although we did not find significant differences between the phasic and tonic cells investigated, the sensitivity to new odour puffs may be more affected by background odour in the latter type of cells, as in these cells the continuous response to the background odour may interact with the responses to new odours. It was shown in tsetse flies that the sensitivity and temporal resolution in tonic cells was lower than in phasic cells (Voskamp et al., 1998). The presence of both phasic and tonic responding cells increases the temporal and dynamic range of the olfactory system of the animal.

Our experiments showed that the presence of background odour of (±)-1-octen-3-ol can affect the EAG response to test odour pulses of (±)-1-octen-3-ol, 2-pentanone and, to a lesser extent, R-(+)-limonene. Therefore, olfactory cells are present that are sensitive to (±)-1-octen-3-ol, as well as 2-pentanone and R-(+)-limonene, as was found in most single cell experiments. Two possible models explain this multiple sensitivity: either the three compounds share one unspecific receptor in the olfactory cell membrane, or, alternatively, they may bind to different receptor sites that activate one shared intracellular signal transduction cascade. Both models can explain cross-adaptation and cross-sensitization (Derby et al., 1991).

We tested the odour of chicken manure as a natural background odour. However, the concentration of this complex odour mixture is difficult to estimate, and no information about the concentration in chicken farms was present. If the responses in the natural habitat of the flies are comparable to those found in this study, the presence of a large odour concentration in habitats such as, for example, chicken farms, may not affect the sensitivity of the olfactory system of flies to new odour pulses to a large extent. Therefore, using baits, even containing odour of manure, may be effective in luring animals to control fly populations in smelly, odour-loaded environments.


This work was supported financially by the Technical Sciences Foundation of the Netherlands Organization for Scientific Research (STW-NWO, grant GBI 33.2997).

Accepted 31 January 2002