Olfactory Physiology in the Drosophila Antenna and Maxillary Palp: acj6 Distinguishes Two Classes of Odorant Pathways Richard K. Ayer, Jr. and John Carlson*
Department of Biology, Yale University, New Haven, Connecticut 0651 1
SUMMARY This article provides characterization of the electrical response to odorants in the Drosophila antenna and provides physiological evidence that a second organ, the maxillary palp, also has olfactory function in Drosophila. The acj6 mutation, previously isolated by virtue of defective olfactory behavior, affects olfactory physiology in the maxillary palp as well as in the antenna. Interestingly, abnormal chemosensory jump 6 ( a c j 6 ) reduces response in the maxillary palp to all odorants tested except benzaldehyde (odor of almond), as if response to benzal-
dehyde is mediated through a different type of odorant pathway from the other odorants. In other experiments, different parts of the antenna are shown to differ with respect to odorant sensitivity. Evidence is also provided that antennal response to odorants varies with age, and that odorants differ in their age dependence. ic3 1992 John
INTRODUCTION
in processing olfactory information distributed among different odorant pathways? Evidence for patterns of organization along both of these lines has been found in other organisms. For example, in the moth Manducu sexta, units responding to female attraction pheromone are housed in a particular class of sensory hairs on the male antenna, and they project to a particular region in the antennal lobe (Schneiderman and Hildebrand, 1985). In vertebrates, there is evidence that response to certain odorants is mediated through a CAMPpathway, whereas response to certain other odorants is mediated through an IP, pathway ( Boekhoff, Tareilus, Strotmann, and Breer, 1990; Sklar, Anholt, and Snyder, 1986). The antenna of the fruit fly Drosophila melanogaster is morphologically complex. The third antennal segment, which has olfactory function, is covered with three classes of sensilla, each arranged in a stereotypic pattern. Patterns of gene expression in the antenna are also complex, as evidenced by a study of enhancer trap lines designed to reveal patterns of gene expression within the Drosophila
Fruit flies are attracted to a remarkable number of odorants. More than 1000 different odorants were shown to attract one species, Dams dorsalis (Beroza and Green, 1963). Fruit flies are also capable of odor discrimination, as has been shown in another species, Drosophila melunogaster ( Siddiqi, 1983). How is the fruit fly olfactory system organized to process such an enormous diversity of chemical information? The question of how the system is functionally organized can be considered in spatial terms: for example, how are the units that mediate response to a particular odorant spatially distributed on the sensory surface? The question can also be considered in mechanistic terms: how are the individual molecular components that act Received May 19, 1992; accepted July 14, 1992. Journal of Neurobiology, Vol. 23. No. 8, pp. 965-982 (1992) 0 1992 John Wiley & Sons, Inc. CCC 0022-3034/92/080965- 18 * To whom correspondence should be addressed.
Wiley & Sons, Inc.
Keywords: olfactory physiology, antenna, maxillary palps, olfaction.
965
966
10 -2 dilutions), responses
showed different kinetics from those observed using lower concentrations (Figure 3). Responses to undiluted ethyl acetate, benzaldehyde, acetone, and butanol showed a slow return to baseline. The response to undiluted propionic acid, after an initial small negative polarization. quickly returned and overshot the baseline by up to 10 mV. Similar EAG responses to short-chain acids, including propionic acid, were reported for the Mediterranean fruit fly, Cerutitis cuppitutu. by Light et al. ( 1988). Additionally, after the positive-going EAG response to propionic acid, Vslanddeclined to near 0 mV. During the next several minutes, EAG responses to smaller doses of propionic acid and other odorants were diminished or eliminated entirely, after which a gradual
24
-
20
.
16
-
65 -
65 -
m V 12
.
4-
mV 4 .
8 -
32-
32.
A
B
4 -
0-
A
C
87-
1
1-
2.5 x 104
A
0-
OIL lo5 10" lo9 10' WENZALDEHYDEI
H ~ O 1 0 ~ 1 0 " 10" 10-l loo [ETHYL ACETATE1
81 -
1-
0-
loo [PROPIONIC ACID]
28 . 24 . 20 .
mV 1 6 . 12 1
10-3
.
8 . 4. 0-
- ~ loo H ~ O 10" lo4 1 0 . ~ 1 010-I [ACETONE]
H ~ O 1 0 ' ~ 10" [BUTANOL]
10-l
loo
Figure 2 EAG amplitude dose-response curves. Arrows along the concentration axis indicate intermediate odorant dose (see Materials and Methods and Table 1 ). Each curve represents 16-20 flies of mixed sex. Error bars, too small to be seen in some cases, indicate S.E.M. The diluent was water in all cases except for benzaldehyde, where parafin oil was used. Although data for each curve were generally gathered within several consecutive days, the five curves were obtained at different times over the course of 3 years. This, in part. may explain why the response to water vapor differs in the different panels (see Materials and Methods). Recordings were made in the vicinity of the arrow in Fig. I (B).
and partial recovery occurred. This effect was specific to propionic acid: undiluted concentrations of benzaldehyde and ethyl acetate did not markedly reduce or eliminate subsequent responses.
Controlling Experimental Variation EAG amplitudes showed relatively small variation among flies on a given day, but demonstrated substantial day-to-day variation when days were separated by pe-
Table L
Intermediate Odorant Dilutions
Odorant
Intermediate Dilution
Acetone Propionic acid Benzaldeh yde Butanol Ethyl acetate
1.0 x 10-3 1.5 x 10-3 2.5 x 10-4 2.0 x 10-3 4.0 x 10-4
riods of time on the order of months. Variables that might be relatively well-controlled among individual flies o n a particular day, or over consecutive days, but uncontrolled over longer time periods, include alterations in environmental factors such as humidity, or subtle changes in technique. We have attempted t o minimize variation by measuring experimental and control groups in parallel, and by performing experiments over a short period of time, that is, on one day, or on a few consecutive days (except as indicated in the legend to Fig. 2 ) . Anesthetics were not used in this study. Limited data suggested that some animals exposed to CO, or ether showed reduced EAG amplitudes when tested shortly after treatment: when tested I day later, animals treated with CO, appeared to have recovered, bct some animals treated with ether still showed reduced EAG amplitude. Effects of ether narcosis on the EAG of moths have been documented previously by Schneider ( 1957; reviewed in Kaissling, 197 I ), who used ether-induced elimination of the EAG as evidence that the signal was a function ofthe antenna1 tissue, and did not represent a chemical alteration of electrode potential caused by odorant application.
970
ilycr und Curlson
Fly Stocks and Cultures. The wild-type stock, CS-5, was isogenic for a single X chromosome, derived from the wild-type strain Canton S. The derivation of CS-5 and the mutant ucj6 have been described in detail (McKenna et al., 1989). CS-5 was the parental wild-type from which ucjh was derived. Methodology used for Drosopphila culture was as described in Monte et al. ( 1989).
RESULTS Figure I ( A ) shows the head ofDrosophilu rnelanogaster and serves to illustrate two pairs of its sensory organs, the antennae and the maxillary palps. Higher-magnification views of the antenna [Fig. 1 ( B ) ] and the maxillary palp [Fig. 1 ( C ) ] show the sensilla that cover them. Morphological studies have demonstrated that the third antennal segment carries three types of sensilla (s.): s. basiconica, s. trichodea, and s. coeloconica (Venkatesh and Singh, 1984; Stocker and Gendre, 1988). The maxillary palp is populated by s. basiconica and s. trichodea only (Singh and Nayak, 1985; Harris, 1972). All three types of sensilla are innervated and are believed to have sensory function. Morphological studies agree that basiconica and coeloconica contain pores or channels and are thus likely to be olfactory, but the studies disagree as to whether s. trichodea have pores in Drosophila antennae (Venkatesh and Singh, 1984; Link, 1983, cited in Venard and Stocker, 199 1 ) . In many moths, s. trichodea are present in much higher numbers in males than females and are the site of pheromone reception (Schneider and Steinbrecht, 1968).
Electrical Response of the Drosophila Antenna to Odorants With recording and reference electrodes positioned as described in Materials and Methods, odorant sdmulation produces EAG records like those seen in Figure 3. From top to bottom, the figure shows EAG responses of increasing amplitude, produced by stimulation with vapor of increasing concentrations of ethyl acetate. The top trace is a control response to the water diluent. For this study we used four other test odorants in addition to ethyl acetate: benzaldehyde, acetone, butanol, and propionic acid. Each odorant represents a different chemical class: acetate esters, aldehydes, ketones, alcohols, and organic acids. Responses to all tested odorants appear qualitatively similar at dilutions of lo-' or below. At the
highest concentrations tested (> all of the tested odorants produce EAG responses that are more complex [Fig. 3 (bottom trace): see Materials and Methods]. Using the same recording method, we were unable to record electrical responses from the eye with any of three odorants tested: ethyl acetate, benzaldehyde, or propionic acid, even at undiluted concentrations (data not shown). These observations suggest that the odorant-generated potentials discussed in this article represent signals specific to olfactory sensory epithelium. Figure 2 shows that the amplitudes of EAG responses are dose dependent. The absolute amplitude of the EAG response, especially at low odorant doses, is highly dependent on the amplitude of the EAG response to the diluent (Fig. 2: and Materials and Methods). As can be seen in the figure, benzaldehyde and propionic acid show saturation at high concentrations. We do not know whether ethyl acetate, acetone, and butanol would show saturation if more data were available for dilutions between lop2and 10' (solubility of these odorants in water is limited). If not, then it is possible that saturation would be reached at higher doses: dosage is measured in terms of the dilution of the odorant solution from which the vapor derives, and not in terms of the number of molecules delivered to the antennal surface, which varies according to the volatility of the test odorant. It is also possible that a large nonphysiological response occurs at the higher doses of ethyl acetate, acetone, and butanol (but not propionic acid or benzaldehyde) and occludes a saturated physiological response. In the remainder of this study, only dilutions of odorants are used.
The Drosophila Maxillary Palp Generates an Odorant-Dependent Signal The maxillary palp has long been suggested to have olfactory function in Drosophila, a proposal that has received support from anatomical evidence (Harris, 1972; Singh and Nayak, 1985; Stocker, Lienhard, Borst, and Fischbach, 1990) and, recently, behavioral experiments (Stocker and Gendre, 1989; Venard and Stocker, 199 1 ) . Electrophysiological experiments have provided evidence for olfactory function in the maxillary palps of blowflies (Dethier, 1952; Van der Starre and Templaar, 1976). To our knowledge, however, there has been no direct physiological demonstration of olfactory function in Drosophila. In order to determine whether the maxillary palp responds electri-
L Figure 3 Typical EAG responses to a range of dilutions of ethyl acetate ( E A ) , diluted in water. Relative to the unstimulated, positive, voltage baseline, stimuli produce negative voltage deflections that quickly return to baseline (upper five traces). In contrast. the response to undiluted EA (lower trace) shows a slow recovery and can be accompanied by an initial small polarization, prior to stimulus delivery, probably due to diffusion of odorant from the saturated filter disk (see Materials and Methods). High odorant dilutions (> I 0 - 2 )were not used in this study except where indicated otherwise. EAG amplitude is measured from the prestimulation baseline to the maximal odorant-induced polarization, Traces are offset by -5 m V so that EAG responses d o not overlap. Scale bar = 5 mV, 1 s.
cally to olfactory stimuli in Drosophila, palps were stimulated with vapors of the five test odorants. Figure 4 demonstrates that the maxillary palp gives a robust electrical response to odorants. Using a set of intermediate concentrations (chosen according to criteria defined in Materials and Methods for the antenna and listed in Table 1 ), responses varied from -2 mV in the case of benzaldehyde to -7 mV in the case of ethyl acetate (Figs. 4 and 5 ). Figure 5 ( A ) also shows results of a control experiment, indicating that the signal recorded from the maxillary palp is not generated by the antenna: the response measured from the maxillary palp is not affected by surgical removal of the antennae. We will henceforth refer to a recording from the maxillary palps as an electropalpogram, or EPG. Interestingly, two distinct differences in sensitivity were seen between the maxillary palp and the antenna in a separate experiment that directly compared the amplitudes of EAGs and EPGs [Fig. 5 ( B ) ] . First, the palp shows virtually no measurable response to water vapor. Second, the palp
shows a relatively small response to benzaldehyde. Whereas the amplitudes of the EPGs range between -40% and -65% of the corresponding EAGs for the other four odorants, for benzaldehyde the EPG amplitude is only -20% that of the EAG. Dose-response curves for maxillary palps and antenna, performed in parallel, would be required in order to determine whether the reduced response of the palps relative to the antenna represents a threshold shift, or perhaps a uniformly reduced sensitivity at all concentrations. The Olfactory Mutant acj6 Exhibits an Odorant-Specific Defect in Maxillary Palp Physiology
Given that maxillary palps have olfactory function, with similar but distinguishable characteristics from those of the antenna, how similar are the molecular genetic underpinnings of olfaction in the two organs? As a first step in addressing this question, we have used the olfactory mutant acj6, originally isolated due to its abnormal olfactory be-
972
A y r and Curlson
Figure 4 Typical EPG responses. The top trace shows the response to water, the middle trace shows response to benzaldehyde, and the bottom trace shows response to ethyl acetate. Relative to the unstimulated, positive voltage baseline, stimuli produce negative voltage deflections that quickly return to baseline. Electropalpogrdm (EPG) amplitude is measured from the prestimulation baseline to the maximal odorant-induced polarization. Scale bar = 2 mV. 500 ms. (Note the scale is different from that in Fig. 3 . )
havior ( McKenna et a]., 1989). ucJ6 was subsequently found to have an EAG defect, but a normal electroretinogram ( Ayer and Carlson, 199 I ). It is defective in a test of larval olfactory behavior, but normal in a test of larval visual behavior. ucj6 has been shown to have reduced EAG amplitude in response to the odorants ethyl acetate and benzaldehyde as well as to water vapor ( Ayer and Carlson, I99 1; see also below, Figure 10). The mutant also has reduced EAG responses to the odorants acetone, propionic acid, and butanol ( R . K. Ayer and J. Carlson, unpublished results); however, due to the mutant’s reduction in EAG response to water vapor, a reduction in the EAG response to an odorant could be, a priovi, due to a loss of either odor responsiveness, or response to the water which was used as a diluent. Ayer and Carlson ( I99 1 ) were able to show that a c j 6 had a defective response to ethyl acetate vapor per w, but did not find clear evidence of decreased response to benzaldehyde vapor per se. The EPG, which contains a negligible response to water vapor, presents a useful means with which to investigate the nature of the acj6 defect. Figure 6 ( A ) demonstrates that EPG amplitudes in the mutant are severely reduced, by approximately 4-6 mV, for the odorants acetone, propionic acid, butanol, and ethyl acetate. This reduction is almost certainly not due to the loss of water vapor response, as the magnitude of the response to water vapor alone is :a Drosophila mutant lacking antennal basiconic sensilla (Diptera: Drosophilidae). J . Imect Behuv. 4:683-705. VENARD,R. and PICHON. Y. ( 198 1 ). Etude tlectro-antennographique de la rCponse pCriphCrique de I'antenne de Drosophila mclunogistc~r2 des stimulations odorantes. C R. Acad. Sci. Paris 2932339-842. VENARD,R. and PICHON, Y. ( 1984). Electrophysiological analysis of the peripheral response to odours in wild type and smell-deficient o//Cmutant of Drosophila melanogaster. J. Insect. Physiol. 30: 1-5. VENKATESH,S. and SINGH,R. (1984). Sensilla on the third antennal segment of Drosopkila melanogaster Meigtw. Int. J . Insecl Morphol. Emhryol. 135 1-63. WHITE, P. R. ( 1991 ). The electroantennogram response: effects of varying sensillum numbers and recording electrode position in a clubbed antenna. J. Insect P h ~ ~ . ~ 37: i o l 145-1 . 32. WOODARD,C., ALCORTA,E., and CARLSON,J. ( 1992). The rdgB gene of Drosophila: a link between vision and olfaction. 1.Neirrogtwetic.s 8: 17-32.
Department of Biology, Yale University, New Haven, Connecticut 0651 1
SUMMARY This article provides characterization of the electrical response to odorants in the Drosophila antenna and provides physiological evidence that a second organ, the maxillary palp, also has olfactory function in Drosophila. The acj6 mutation, previously isolated by virtue of defective olfactory behavior, affects olfactory physiology in the maxillary palp as well as in the antenna. Interestingly, abnormal chemosensory jump 6 ( a c j 6 ) reduces response in the maxillary palp to all odorants tested except benzaldehyde (odor of almond), as if response to benzal-
dehyde is mediated through a different type of odorant pathway from the other odorants. In other experiments, different parts of the antenna are shown to differ with respect to odorant sensitivity. Evidence is also provided that antennal response to odorants varies with age, and that odorants differ in their age dependence. ic3 1992 John
INTRODUCTION
in processing olfactory information distributed among different odorant pathways? Evidence for patterns of organization along both of these lines has been found in other organisms. For example, in the moth Manducu sexta, units responding to female attraction pheromone are housed in a particular class of sensory hairs on the male antenna, and they project to a particular region in the antennal lobe (Schneiderman and Hildebrand, 1985). In vertebrates, there is evidence that response to certain odorants is mediated through a CAMPpathway, whereas response to certain other odorants is mediated through an IP, pathway ( Boekhoff, Tareilus, Strotmann, and Breer, 1990; Sklar, Anholt, and Snyder, 1986). The antenna of the fruit fly Drosophila melanogaster is morphologically complex. The third antennal segment, which has olfactory function, is covered with three classes of sensilla, each arranged in a stereotypic pattern. Patterns of gene expression in the antenna are also complex, as evidenced by a study of enhancer trap lines designed to reveal patterns of gene expression within the Drosophila
Fruit flies are attracted to a remarkable number of odorants. More than 1000 different odorants were shown to attract one species, Dams dorsalis (Beroza and Green, 1963). Fruit flies are also capable of odor discrimination, as has been shown in another species, Drosophila melunogaster ( Siddiqi, 1983). How is the fruit fly olfactory system organized to process such an enormous diversity of chemical information? The question of how the system is functionally organized can be considered in spatial terms: for example, how are the units that mediate response to a particular odorant spatially distributed on the sensory surface? The question can also be considered in mechanistic terms: how are the individual molecular components that act Received May 19, 1992; accepted July 14, 1992. Journal of Neurobiology, Vol. 23. No. 8, pp. 965-982 (1992) 0 1992 John Wiley & Sons, Inc. CCC 0022-3034/92/080965- 18 * To whom correspondence should be addressed.
Wiley & Sons, Inc.
Keywords: olfactory physiology, antenna, maxillary palps, olfaction.
965
966
10 -2 dilutions), responses
showed different kinetics from those observed using lower concentrations (Figure 3). Responses to undiluted ethyl acetate, benzaldehyde, acetone, and butanol showed a slow return to baseline. The response to undiluted propionic acid, after an initial small negative polarization. quickly returned and overshot the baseline by up to 10 mV. Similar EAG responses to short-chain acids, including propionic acid, were reported for the Mediterranean fruit fly, Cerutitis cuppitutu. by Light et al. ( 1988). Additionally, after the positive-going EAG response to propionic acid, Vslanddeclined to near 0 mV. During the next several minutes, EAG responses to smaller doses of propionic acid and other odorants were diminished or eliminated entirely, after which a gradual
24
-
20
.
16
-
65 -
65 -
m V 12
.
4-
mV 4 .
8 -
32-
32.
A
B
4 -
0-
A
C
87-
1
1-
2.5 x 104
A
0-
OIL lo5 10" lo9 10' WENZALDEHYDEI
H ~ O 1 0 ~ 1 0 " 10" 10-l loo [ETHYL ACETATE1
81 -
1-
0-
loo [PROPIONIC ACID]
28 . 24 . 20 .
mV 1 6 . 12 1
10-3
.
8 . 4. 0-
- ~ loo H ~ O 10" lo4 1 0 . ~ 1 010-I [ACETONE]
H ~ O 1 0 ' ~ 10" [BUTANOL]
10-l
loo
Figure 2 EAG amplitude dose-response curves. Arrows along the concentration axis indicate intermediate odorant dose (see Materials and Methods and Table 1 ). Each curve represents 16-20 flies of mixed sex. Error bars, too small to be seen in some cases, indicate S.E.M. The diluent was water in all cases except for benzaldehyde, where parafin oil was used. Although data for each curve were generally gathered within several consecutive days, the five curves were obtained at different times over the course of 3 years. This, in part. may explain why the response to water vapor differs in the different panels (see Materials and Methods). Recordings were made in the vicinity of the arrow in Fig. I (B).
and partial recovery occurred. This effect was specific to propionic acid: undiluted concentrations of benzaldehyde and ethyl acetate did not markedly reduce or eliminate subsequent responses.
Controlling Experimental Variation EAG amplitudes showed relatively small variation among flies on a given day, but demonstrated substantial day-to-day variation when days were separated by pe-
Table L
Intermediate Odorant Dilutions
Odorant
Intermediate Dilution
Acetone Propionic acid Benzaldeh yde Butanol Ethyl acetate
1.0 x 10-3 1.5 x 10-3 2.5 x 10-4 2.0 x 10-3 4.0 x 10-4
riods of time on the order of months. Variables that might be relatively well-controlled among individual flies o n a particular day, or over consecutive days, but uncontrolled over longer time periods, include alterations in environmental factors such as humidity, or subtle changes in technique. We have attempted t o minimize variation by measuring experimental and control groups in parallel, and by performing experiments over a short period of time, that is, on one day, or on a few consecutive days (except as indicated in the legend to Fig. 2 ) . Anesthetics were not used in this study. Limited data suggested that some animals exposed to CO, or ether showed reduced EAG amplitudes when tested shortly after treatment: when tested I day later, animals treated with CO, appeared to have recovered, bct some animals treated with ether still showed reduced EAG amplitude. Effects of ether narcosis on the EAG of moths have been documented previously by Schneider ( 1957; reviewed in Kaissling, 197 I ), who used ether-induced elimination of the EAG as evidence that the signal was a function ofthe antenna1 tissue, and did not represent a chemical alteration of electrode potential caused by odorant application.
970
ilycr und Curlson
Fly Stocks and Cultures. The wild-type stock, CS-5, was isogenic for a single X chromosome, derived from the wild-type strain Canton S. The derivation of CS-5 and the mutant ucj6 have been described in detail (McKenna et al., 1989). CS-5 was the parental wild-type from which ucjh was derived. Methodology used for Drosopphila culture was as described in Monte et al. ( 1989).
RESULTS Figure I ( A ) shows the head ofDrosophilu rnelanogaster and serves to illustrate two pairs of its sensory organs, the antennae and the maxillary palps. Higher-magnification views of the antenna [Fig. 1 ( B ) ] and the maxillary palp [Fig. 1 ( C ) ] show the sensilla that cover them. Morphological studies have demonstrated that the third antennal segment carries three types of sensilla (s.): s. basiconica, s. trichodea, and s. coeloconica (Venkatesh and Singh, 1984; Stocker and Gendre, 1988). The maxillary palp is populated by s. basiconica and s. trichodea only (Singh and Nayak, 1985; Harris, 1972). All three types of sensilla are innervated and are believed to have sensory function. Morphological studies agree that basiconica and coeloconica contain pores or channels and are thus likely to be olfactory, but the studies disagree as to whether s. trichodea have pores in Drosophila antennae (Venkatesh and Singh, 1984; Link, 1983, cited in Venard and Stocker, 199 1 ) . In many moths, s. trichodea are present in much higher numbers in males than females and are the site of pheromone reception (Schneider and Steinbrecht, 1968).
Electrical Response of the Drosophila Antenna to Odorants With recording and reference electrodes positioned as described in Materials and Methods, odorant sdmulation produces EAG records like those seen in Figure 3. From top to bottom, the figure shows EAG responses of increasing amplitude, produced by stimulation with vapor of increasing concentrations of ethyl acetate. The top trace is a control response to the water diluent. For this study we used four other test odorants in addition to ethyl acetate: benzaldehyde, acetone, butanol, and propionic acid. Each odorant represents a different chemical class: acetate esters, aldehydes, ketones, alcohols, and organic acids. Responses to all tested odorants appear qualitatively similar at dilutions of lo-' or below. At the
highest concentrations tested (> all of the tested odorants produce EAG responses that are more complex [Fig. 3 (bottom trace): see Materials and Methods]. Using the same recording method, we were unable to record electrical responses from the eye with any of three odorants tested: ethyl acetate, benzaldehyde, or propionic acid, even at undiluted concentrations (data not shown). These observations suggest that the odorant-generated potentials discussed in this article represent signals specific to olfactory sensory epithelium. Figure 2 shows that the amplitudes of EAG responses are dose dependent. The absolute amplitude of the EAG response, especially at low odorant doses, is highly dependent on the amplitude of the EAG response to the diluent (Fig. 2: and Materials and Methods). As can be seen in the figure, benzaldehyde and propionic acid show saturation at high concentrations. We do not know whether ethyl acetate, acetone, and butanol would show saturation if more data were available for dilutions between lop2and 10' (solubility of these odorants in water is limited). If not, then it is possible that saturation would be reached at higher doses: dosage is measured in terms of the dilution of the odorant solution from which the vapor derives, and not in terms of the number of molecules delivered to the antennal surface, which varies according to the volatility of the test odorant. It is also possible that a large nonphysiological response occurs at the higher doses of ethyl acetate, acetone, and butanol (but not propionic acid or benzaldehyde) and occludes a saturated physiological response. In the remainder of this study, only dilutions of odorants are used.
The Drosophila Maxillary Palp Generates an Odorant-Dependent Signal The maxillary palp has long been suggested to have olfactory function in Drosophila, a proposal that has received support from anatomical evidence (Harris, 1972; Singh and Nayak, 1985; Stocker, Lienhard, Borst, and Fischbach, 1990) and, recently, behavioral experiments (Stocker and Gendre, 1989; Venard and Stocker, 199 1 ) . Electrophysiological experiments have provided evidence for olfactory function in the maxillary palps of blowflies (Dethier, 1952; Van der Starre and Templaar, 1976). To our knowledge, however, there has been no direct physiological demonstration of olfactory function in Drosophila. In order to determine whether the maxillary palp responds electri-
L Figure 3 Typical EAG responses to a range of dilutions of ethyl acetate ( E A ) , diluted in water. Relative to the unstimulated, positive, voltage baseline, stimuli produce negative voltage deflections that quickly return to baseline (upper five traces). In contrast. the response to undiluted EA (lower trace) shows a slow recovery and can be accompanied by an initial small polarization, prior to stimulus delivery, probably due to diffusion of odorant from the saturated filter disk (see Materials and Methods). High odorant dilutions (> I 0 - 2 )were not used in this study except where indicated otherwise. EAG amplitude is measured from the prestimulation baseline to the maximal odorant-induced polarization, Traces are offset by -5 m V so that EAG responses d o not overlap. Scale bar = 5 mV, 1 s.
cally to olfactory stimuli in Drosophila, palps were stimulated with vapors of the five test odorants. Figure 4 demonstrates that the maxillary palp gives a robust electrical response to odorants. Using a set of intermediate concentrations (chosen according to criteria defined in Materials and Methods for the antenna and listed in Table 1 ), responses varied from -2 mV in the case of benzaldehyde to -7 mV in the case of ethyl acetate (Figs. 4 and 5 ). Figure 5 ( A ) also shows results of a control experiment, indicating that the signal recorded from the maxillary palp is not generated by the antenna: the response measured from the maxillary palp is not affected by surgical removal of the antennae. We will henceforth refer to a recording from the maxillary palps as an electropalpogram, or EPG. Interestingly, two distinct differences in sensitivity were seen between the maxillary palp and the antenna in a separate experiment that directly compared the amplitudes of EAGs and EPGs [Fig. 5 ( B ) ] . First, the palp shows virtually no measurable response to water vapor. Second, the palp
shows a relatively small response to benzaldehyde. Whereas the amplitudes of the EPGs range between -40% and -65% of the corresponding EAGs for the other four odorants, for benzaldehyde the EPG amplitude is only -20% that of the EAG. Dose-response curves for maxillary palps and antenna, performed in parallel, would be required in order to determine whether the reduced response of the palps relative to the antenna represents a threshold shift, or perhaps a uniformly reduced sensitivity at all concentrations. The Olfactory Mutant acj6 Exhibits an Odorant-Specific Defect in Maxillary Palp Physiology
Given that maxillary palps have olfactory function, with similar but distinguishable characteristics from those of the antenna, how similar are the molecular genetic underpinnings of olfaction in the two organs? As a first step in addressing this question, we have used the olfactory mutant acj6, originally isolated due to its abnormal olfactory be-
972
A y r and Curlson
Figure 4 Typical EPG responses. The top trace shows the response to water, the middle trace shows response to benzaldehyde, and the bottom trace shows response to ethyl acetate. Relative to the unstimulated, positive voltage baseline, stimuli produce negative voltage deflections that quickly return to baseline. Electropalpogrdm (EPG) amplitude is measured from the prestimulation baseline to the maximal odorant-induced polarization. Scale bar = 2 mV. 500 ms. (Note the scale is different from that in Fig. 3 . )
havior ( McKenna et a]., 1989). ucJ6 was subsequently found to have an EAG defect, but a normal electroretinogram ( Ayer and Carlson, 199 I ). It is defective in a test of larval olfactory behavior, but normal in a test of larval visual behavior. ucj6 has been shown to have reduced EAG amplitude in response to the odorants ethyl acetate and benzaldehyde as well as to water vapor ( Ayer and Carlson, I99 1; see also below, Figure 10). The mutant also has reduced EAG responses to the odorants acetone, propionic acid, and butanol ( R . K. Ayer and J. Carlson, unpublished results); however, due to the mutant’s reduction in EAG response to water vapor, a reduction in the EAG response to an odorant could be, a priovi, due to a loss of either odor responsiveness, or response to the water which was used as a diluent. Ayer and Carlson ( I99 1 ) were able to show that a c j 6 had a defective response to ethyl acetate vapor per w, but did not find clear evidence of decreased response to benzaldehyde vapor per se. The EPG, which contains a negligible response to water vapor, presents a useful means with which to investigate the nature of the acj6 defect. Figure 6 ( A ) demonstrates that EPG amplitudes in the mutant are severely reduced, by approximately 4-6 mV, for the odorants acetone, propionic acid, butanol, and ethyl acetate. This reduction is almost certainly not due to the loss of water vapor response, as the magnitude of the response to water vapor alone is :a Drosophila mutant lacking antennal basiconic sensilla (Diptera: Drosophilidae). J . Imect Behuv. 4:683-705. VENARD,R. and PICHON. Y. ( 198 1 ). Etude tlectro-antennographique de la rCponse pCriphCrique de I'antenne de Drosophila mclunogistc~r2 des stimulations odorantes. C R. Acad. Sci. Paris 2932339-842. VENARD,R. and PICHON, Y. ( 1984). Electrophysiological analysis of the peripheral response to odours in wild type and smell-deficient o//Cmutant of Drosophila melanogaster. J. Insect. Physiol. 30: 1-5. VENKATESH,S. and SINGH,R. (1984). Sensilla on the third antennal segment of Drosopkila melanogaster Meigtw. Int. J . Insecl Morphol. Emhryol. 135 1-63. WHITE, P. R. ( 1991 ). The electroantennogram response: effects of varying sensillum numbers and recording electrode position in a clubbed antenna. J. Insect P h ~ ~ . ~ 37: i o l 145-1 . 32. WOODARD,C., ALCORTA,E., and CARLSON,J. ( 1992). The rdgB gene of Drosophila: a link between vision and olfaction. 1.Neirrogtwetic.s 8: 17-32.
