Odor coding in the Drosophila maxillary palp

Olfactory systems detect and differentiate among many kinds of odor stimuli. Olfactory receptor neurons encode qualitative, quantitative, temporal, and spatial information about odors. Elucidating the mechanisms by which olfactory information is encoded is an intriguing problem in contemporary neurobiology. A major difficulty has been the complexity of most olfactory systems. In the case of mammals, for example, the number of olfactory receptor neurons (ORNs) is very large, as is the number of distinct functional classes into which these neurons may fall. Moreover, in many systems it is difficult to measure systematically the physiological properties of individual ORNs in vivo. Insect ORNs are distributed in sensilla, usually in the form of sensory hairs that protrude from the cuticle, providing accessibility and ease of identification. Odorants pass through tiny pores in the walls of these sensilla and stimulate dendrites bathing in the lymph inside. Drosophila has a relatively simple olfactory system. Olfactory response can be measured in vivo, via either physiological or behavioral means, and a variety of genetic and molecular approaches are available to study its olfactory system (de Bruyne, 1999 and references).

Each of the paired maxillary palps contains only 120 ORNs, housed in 60 sensilla of a single category, sensilla basiconica. The numerical simplicity of the maxillary palp makes it possible to perform an exhaustive study of its neuronal composition and to characterize in detail its functional organization. Extensive extracellular recordings from single sensilla reveal that the neurons fall into six functional classes. Each of the 60 sensilla houses two neurons (A and B), which observe a pairing rule: each sensillum combines neurons of two particular classes A or B in a fixed, not a random, configuration. Behaviorial studies reveal that there are three sensillum types (1, 2 or 3). The three sensillum types are intermingled on the surface of the palp, but their distribution is not random. Since there are two neurons in each sensilla, and three sensillum types, the cellular contents of each of the three sensilla are identified as follows: sensillum type 1 contains neurons 1A and 1B, sensillum type 2 contains neurons 2A and 2B and sensillum type 3 contains neurons 3A and 3B, thus accounting for the six functional classes). Identity of a sensillum is not strictly determined by its position. The neuron pairs in each sensillum exhibit different response characteristics, providing the basis for an olfactory code. A particular odor can excite one neuron and inhibit another; a particular neuron can be excited by one odor and inhibited by another. Some excitatory responses continue beyond the end of odor delivery, but responses to most odors terminate abruptly after the end of odor delivery, with some followed by a period of poststimulus quiescence. Adaptation and cross-adaptation have been documented, and cross-adaptation experiments demonstrate that the two neurons within one type of sensillum can function independently (de Bruyne, 1999).

ORNs from single sensilla on the Drosophila maxillary palp exhibit spontaneous action potentials, clearly distinguishable from background noise. The spikes from an individual sensillum can be resolved into two distinct populations, based on their amplitudes (large and small). This bimodal distribution of spike amplitudes is interpreted as representing the activities of two distinct neurons, an interpretation that has been extensively supported in a wide variety of other insect sensory systems. The large and small spikes represent the output of the two neurons that make up each of the 60 sensilla of the maxillary palp. The two neurons in each sensilla account for the 120 ORNs in each maxillary palp. The two neurons of a sensillum are referred to as the A and B cells (de Bruyne, 1999).

Diversity among the olfactory neurons of the maxillary palp can be observed by analysis of their spontaneous firing rates. The spontaneous action potential frequency of individual cells varies, with most neurons exhibiting frequencies between 3 and 13 spikes/sec. The spontaneous rate of one of these individual cells is relatively constant, however; it stays within a range of +/-3 spikes/sec over the course of recording periods lasting as long as 60 min. In addition to these neurons, some neurons have noticeably higher spontaneous rates of ~30 spikes/sec. Odor stimulation elicits a marked increase in firing frequency in many cases. In most cases one of the two neurons in a sensillum is clearly more excited than was the other by stimulation with a particular odor. The excited neuron can generally be identified by its spike height at the start of the response. In some cases the spike amplitude of the responding neuron gradually changes during the course of stimulation, a phenomenon widely observed in single-sensillum recordings from insects. In such cases, the identities of the neurons can often be confirmed if necessary by examination of the spike shapes, which differ between the two neurons. Excitatory responses start ~80 msec after odor is administered; calculations reveal that 50 msec is required for odor to reach the preparation, indicating a response latency of ~30 msec. The spike frequency rises to a maximum within 50-100 msec after response initiation, depending on the odor and dose, and then declines (de Bruyne, 1999).

Excitatory responses terminate abruptly after the end of the odor stimulation period in most cases. This is consistent with the expected sharp decline in odor levels after the 500 msec delivery period. In these cases a brief period is often observed in which no action potentials are recorded from this neuron. The duration of this poststimulus quiescence appears to be dose-dependent, with higher doses producing longer periods of quiescence, although this relationship has not been examined quantitatively. Poststimulus quiescence is not, however, observed for all odors. In fact, in some cases excitation continues long past the end of odor stimulation. Thus the variations in response at the end of odor stimulation show that different odors may have different effects on the kinetics of excitation, thereby providing one putative mechanism for odor discrimination. Inhibitory responses were also observed in some sensilla. Interestingly, in these sensilla some odors inhibit one cell but excite the other; the cell with the large spikes is inhibited, and the cell with the small spikes is excited. Thus an individual odor can have opposite effects on the firing frequency of different neurons, and an individual neuron can respond oppositely (excitation vs. inhibition) to different odors (de Bruyne, 1999).

To define the basic elements of the olfactory code, attempts were made to determine whether ORNs fall into discrete functional classes and, if so, to determine the number and odor specificity of such classes. This analysis was initiated with a chemically diverse group of 16 odorants selected on the basis of three criteria. Some odorants (ethyl acetate, 3-octanol, and benzaldehyde) were selected because they have been used extensively in previous research on Drosophila olfaction and are known to induce strong behavioral responses. Others were selected because they play important roles in the ecology of related dipteran insects. For example, 4-methylphenol is present in cattle urine and attracts tsetse flies to their hosts; E2-hexenal (leaf aldehyde) is a plant odor that attracts many insect species. Finally, odors were selected to represent certain chemical groups (e.g., ketones, aldehydes, alcohols, esters, and aromatics) (de Bruyne, 1999 and references).

Initial recordings from individual sensilla clearly indicated that sensilla can be divided into three functional types, termed palpal basiconic 1 (pb1), pb2, and pb3. Although of the same morphological category, the three types of sensilla contain neurons with response spectra that are distinct from one another and from neurons of the other sensillum types. The two neurons of the pb1 sensillum are denote pb1A and pb1B. Of these two neurons, pb1B cell responds strongly to only one of the tested odorants, 4-methylphenol, to which it shows an increase in spike frequency of 178 +/- 47 spikes/sec (+/- SD; n = 13). The only other response from pb1B that is significantly different from that of the paraffin oil control is the response to 4-methylcyclohexanol, an odor molecule structurally similar to 4-methylphenol. The pb1A cell, by contrast, shows a broader range of responses to the tested stimuli: it responds most strongly to ethyl acetate, showing an increase of 138 +/- 32 spikes/sec (+/- SD; n = 13), but also responds to several other stimuli (de Bruyne, 1999).

The different neurons of the maxillary palp could be distinguished using a diagnostic subset of 7 of the 16 odors. These 7 odors were therefore used to extend the analysis to a larger number of sensilla. The results confirm the presence of exactly three sensillum types, each containing two neurons, yielding a total of six types of neurons with distinguishable response properties: pb1A, pb1B, pb2A, pb2B, pb3A, and pb3B. The pb1 sensilla contain an A cell that responds strongly to ethyl acetate and a B cell that responds strongly to 4-methylphenol. In the pb2 sensilla, the A cell responds strongly to benzaldehyde. The other cell, pb2B, is excited by 4-methylphenol, although not as strongly as is pb1B. In addition, pb2B is strongly inhibited by 3-octanol and several other odors. Specifically, the spontaneous firing frequency of pb2B is 32 +/- 7 spikes/sec and is reduced 80-100% by 3-octanol. The pb3 sensillum contains two neurons that are both excited by 3-octanol and isoamyl acetate, but pb3B is more strongly stimulated by isoamyl acetate than is pb3A. In summary, this analysis reveals that the maxillary palp contains six distinguishable neuronal types. The responses of some neuronal types are overlapping (e.g., 4-methylphenol excites both pb1B and pb2B neurons), but comparisons of responses to multiple odors clearly shows the profiles to be distinct. In pb1 and pb3 sensilla, A neurons consistently have larger spike amplitudes than B neurons. pb2 sensilla differ from pb1 and pb3 sensilla, not only because the B neuron consistently has higher spontaneous activity than the other five neuronal types but also because pb2A has smaller spikes than pb2B in some sensilla. However, spikes of pb2A are consistently different in shape, having larger negative phases relative to the positive phase (de Bruyne, 1999).

The two neurons in one sensillum observe a pairing rule; for example, the cell that is excited by benzaldehyde (pb2A) is always paired with a cell that is inhibited by 3-octanol (pb2B). Among 232 sensilla examined, the activity of two cells were recorded in 225 cases (in the 7 exceptional sensilla, one cell was absent or unresponsive, perhaps because of damage caused by the recording electrode in at least some cases); in 222 of the 225 cases, one of the three characteristic combinations of cells was found (de Bruyne, 1999).

Although the existence of exactly six neuronal classes seemed clear from this physiological analysis, the classification scheme was tested more rigorously. A cluster analysis of the responses of 54 neurons in 27 sensilla to five of the odors was performed. This analysis yielded five clusters based on the response profiles. Cluster 1 groups cells with a strong response to ethyl acetate (corresponding to pb1A), and cluster 2 unites cells responding strongly to benzaldehyde (pb2A). Clusters 5 and 4 contain cells that respond to 4-methylphenol but that either are or are not inhibited by 3-octanol (pb2B and pb1B, respectively). Cluster 3 includes cells that are excited by both isoamyl acetate and 3-octanol but that show no strong response to the other odors; this cluster includes both pb3A and pb3B cells. This analysis confirms the initial observation that ORNs on the palp can reliably be assigned to cell classes that have clearly distinct response characteristics. This cluster analysis, however, did not resolve pb3A and pb3B. Therefore further evidence was sought that pb3A and pb3B are distinguishable by testing their responses to an additional set of odors. Significant responses from pb3 cells were only observed to 3 of the initial 16 odorants. A homologous series of esters was chosen because the previous analysis had revealed a difference in the responses of pb3A and pb3B to isoamyl acetate. Aliphatic esters with varying chain lengths of both the alcohol and acid moieties were tested. The responses of pb3A and pb3B are clearly distinguishable, in two respects. (1) For most of the esters there is a significant difference in the absolute responses of the two neurons; for example, for pentyl acetate, the response of pb3B is 99 +/- 12 spikes/sec, whereas the response of pb3A is only 9 +/- 2 spikes/sec. (2) The relative responses of the two neurons to different odors vary; for example, the pb3A neuron responds dramatically better to ethyl butyrate than to hexyl acetate, whereas for pb3B the converse is true. This analysis confirms that the response spectra of pb3A and pb3B are clearly different and that the pb3 sensillum, like pb1 and pb2, contains two distinguishable neurons. It is noted with interest that for pb3A, the length of the odorant molecule correlates with potency; odors that elicit the strongest mean responses are those with a total of 6 carbons, with the next most potent being those with 4, 5, or 7 carbons, followed by those with 8 or 10 carbons. However, none of the tested odors elicits more than ~30 spikes/sec from this neuron, leading to a suspicion that there are other molecules in odor space that are more effective stimuli for this neuron (de Bruyne, 1999).

All three sensillum types are found on both male and female palps, and no evidence was found of sexual dimorphism of any kind. A quantitative analysis of spike frequency shows no differences between the sexes in the response of pb1 sensilla to any of seven odors tested (ethyl acetate, isoamyl acetate, 4-methylphenol, benzaldehyde, 3-octanol, E2-hexenal, and cyclohexanone); limited data reveal no differences between the sexes for pb2 or pb3 sensilla. These results are consistent with the lack of sexual dimorphism in the glomeruli of the antennal lobes and form a striking contrast with the dimorphism observed in the olfactory sensilla and glomeruli of moths (de Bruyne, 1999).

The number of functional types of neurons on the maxillary palp (six) is of the same order as the estimated number of five glomeruli that receive afferent fibers from it. This approximate numerical equivalence, in which the number of neuronal classes is determined by direct physiological analysis, is consistent with the approximate equivalence of the number of glomeruli with the number of neuronal types in the mammalian main olfactory epithelium. In moths, there is also a well documented pattern of projection from particular functional classes of neurons, pheromone-sensitive neurons, to a cluster of specialized glomeruli, the macroglomerular complex. However, it remains to be seen whether the two neurons in each sensillum type project to one glomerulus, giving three palpal glomeruli, or to two glomeruli, which would give six palpal glomeruli. The total number of olfactory glomeruli reported for the antennal lobe of Drosophila is ~40 (Laissue, 1999). Hence the total peripheral input to the olfactory-processing centers in the CNS may consist of ~40 basic types of input elements (de Bruyne, 1999 and references).

A major way in which the insect olfactory system differs from that of vertebrates is that the neurons of the sensory field are compartmentalized in sensilla. In this respect, the insect olfactory system can be considered (by analogy to the compound eye) as a 'compound nose'. Neurons of the palp are ordered with respect to this level of organization; the six types of neurons are distributed within sensilla in stereotyped pairs, with a neuron of one particular response spectrum cohabiting in a sensillum with a neuron of another particular response spectrum. Despite the intimate cohabitation of neurons within a sensillum, cross-adaptation experiments show that they are able to function primarily as independent units of perception. This independence and their distinct response spectra are observed despite the fact that both neurons in a sensillum share a common pool of binding proteins (Hekmat-Scafe, 1997; see Drosophila Lush for more information on odorant binding proteins or OBPs) and a common electrical circuit. In this sense, olfactory tissues in insects may be similar to those in mammals (de Bruyne, 1999 and references).

The largely overlapping distribution of the three sensillum types across the sensory field shows that in the maxillary palp there is not an obvious odotopic layout at the primary neuron level. However, the presence of a pb2 'exclusion' zone provides some heterogeneity and likely points to zones that are developmentally distinct. This analysis of sensillum function and organization raises interesting questions about how such a sensory field develops. Olfactory sensilla develop from founder cells (Ray, 1995) that build different sensillum categories. The mixed distribution of functional types on the palp and the variability in the positions of individual sensilla suggest that sensilla are not committed to one of the three alternative fates strictly according to their position in the field. It is not known how the expression of genes encoding the choice between pb1, pb2, and pb3 is regulated. Moreover, further studies will be necessary to reveal the developmental logic by which the stereotyped pairing of neurons is produced. What is the mechanism that regulates expression of class-specific elements such as receptors and coordinates it between the two neurons of one sensillum? (de Bruyne, 1999).

The olfactory system of adult Drosophila contains two organs, the antenna and the maxillary palp. Each antenna is covered with ~500 sensilla of three morphological types, whereas the maxillary palp is covered with 60 sensilla of a single morphological type. It is tempting to compare this arrangement with the presence of two olfactory organs in mammals, the main olfactory epithelium (MOE) and the vomeronasal organ (VNO). Like the VNO, the maxillary palp has been associated with pheromone response and the modulation of sexual behavior. The VNO neurons project to an accessory olfactory bulb that does not overlap with the main olfactory bulb; likewise, projections from the palp have been reported to map to a subset of glomeruli distinct from those that receive input from the antenna. However, VNO neurons differ from MOE neurons in that they are microvillous rather than ciliate, whereas both palpal and antennal ORNs are ciliate. In addition, the VNO neurons probably perceive a different type of odorants (less volatile and larger molecules) from those perceived by the MOE, whereas the present study shows that the maxillary palp is sensitive to many small, volatile molecules that also stimulate the antenna. No evidence was found for a special role of the palp in pheromone perception. This is based on two findings: (1) cis-vaccenyl acetate, a compound reported previously as an inhibitory sex pheromone induces a response from pb1A. However, this neuron responds more strongly to many other odors, and cis-vaccenyl acetate has also been found to stimulate an antennal neuron. (2) No differences were found between male and female ORN populations, making it unlikely that sex-specific pheromonal receptor cells occur on the palps (de Bruyne, 1999 and references).

Odor coding in the Drosophila antenna

Odor coding in the Drosophila antenna is examined by a functional analysis of individual olfactory receptor neurons (ORNs) in vivo. Sixteen distinct classes of ORNs, each with a unique response spectrum to a panel of 47 diverse odors, were identified by extracellular recordings. ORNs exhibit multiple modes of response dynamics: an individual neuron can show either excitatory or inhibitory responses, and can exhibit different modes of termination kinetics, when stimulated with different odors. The 16 ORN classes are combined in stereotyped configurations within seven functional types of basiconic sensilla. One sensillum type contains four ORNs and the others contain two neurons, combined according to a strict pairing rule. A functional map of ORNs is provided, showing that each ORN class is restricted to a particular spatial domain on the antennal surface (de Bruyne, 2001).

Electrophysiological recordings from basiconic sensilla on the antenna consistently show spontaneous action potentials. Spikes from most recordings could be resolved into two populations based on their amplitudes. The two populations of spikes in these sensilla indicate the presence of two neurons, as found in basiconic sensilla on the maxillary palp (de Bruyne, 1999). The neuron with the larger spike amplitude is referred to as the A neuron and the one with the smaller spikes as the B neuron (de Bruyne, 2001).

Neurons show excitatory responses to odors, with different neurons excited by different odors. For instance, one A neuron increases its firing frequency in response to ethyl acetate, while the B neuron spike frequency is unaffected. By contrast, the B neuron in the same sensillum is excited by hexanol. In another sensillum, the A neuron is excited by 1-octen-3-ol, whereas the B neuron is unaffected. In this sensillum, the B neuron is excited by ethyl butyrate; the A neuron also shows a modest increase in firing frequency (de Bruyne, 2001).

Recordings of spontaneous activity from some large basiconic sensilla reveal the presence of four neurons, whose spike amplitudes fell into four classes. These physiological results confirm ultrastructural observations that while most basiconic sensilla house two neurons, many large basiconic sensilla contain four neurons. Strikingly, one of these neurons, the C neuron, responds strongly to CO2 (de Bruyne, 2001).

Analysis reveals that sixteen ORN classes are housed in stereotyped combinations in seven functional types of basiconic sensilla. A set of 47 odorants was chosen to characterize the responses of neurons in antennal basiconic sensilla. The odorants were chosen from a variety of chemical classes and include compounds that contain branched chains, double bonds, two functional groups, and other structural features. Most odorants were chosen because of evidence that they play a role in the chemical ecology of Drosophila or other flies. In addition to testing a large number of neurons with the entire set of 47 odorants, a subset of 12 diagnostic odorants was identified that is particularly useful in identifying and distinguishing among neuronal classes: these 12 odorants were used in most subsequent experiments (de Bruyne, 2001).

Extensive recordings from large basiconic sensilla reveal that they fall into three distinct types based on the odor response spectra of their ORNs. These three types are referred to as antennal basiconic (ab) types ab1, ab2, and ab3, as distinct from pb sensilla (basiconic sensilla on the maxillary palp [de Bruyne, 1999]). Each ab type contains a stereotyped combination of neurons (de Bruyne, 2001).

The ab1 sensillum is unique in that it houses four neurons. The A neuron within this sensillum type, referred to as ab1A, responds most strongly to ethyl acetate. This neuron also responds to several other odorants, notably ethyl butyrate, but it is unusual in responding strongly to the diluent control stimulus, paraffin oil. It is likely that for this particular neuron, much of the response observed to other odors represents a response to an element of the delivery system. The ab1B, C, and D neurons respond with a high degree of specificity to 2,3-butanedione, CO2, and methyl salicylate, respectively; they do not respond to the diluent control stimulus. The other two types of large basiconic sensilla, ab2 and ab3, contain only two neurons. ab2 sensilla contain an A neuron that responds strongly to ethyl acetate, with a somewhat weaker response to 2,3-butanedione, and a B neuron that responds moderately to ethyl butyrate and hexanol. ab3 contains an A neuron that is excited by ethyl butyrate and pentyl acetate and a B neuron that responds to heptanone and hexanol (de Bruyne, 2001).

The functional classification of eight neuronal types in large basiconic sensilla was confirmed through a cluster analysis. The eight ORNs are distinct from the six of the maxillary palp. Furthermore, of the 145 large basiconic sensilla from which recordings were taken, all could be classified as either ab1, ab2, or ab3; no evidence was found for an additional type (de Bruyne, 2001).

In addition to recording from large basiconic sensilla, recordings were taken from 128 small basiconic sensilla on the anterior and posterior surfaces of the third antennal segment. An additional four sensillum types, called ab4, ab5, ab6, and ab7, were identified based on odor response profiles. Each of the four sensillum types contains two neurons distinct from each other and from all other previously defined neuronal types, including those on the maxillary palp. Thus, an additional eight ORN classes housed within the small basiconic sensilla have been defined. Moreover, the neurons in the small basiconic sensilla, like those in the large basiconic sensilla, are housed in characteristic combinations; they observe a strict pairing rule (de Bruyne, 2001).

The ab4 sensillum houses a neuron, ab4A, that responds strongly to E2-hexenal, a 'green leaf volatile'; other insects have ORNs that appear to be narrowly tuned to this molecule. Not only does this odor elicit a high frequency of firing from the ab4A neuron, but the neuron continues to fire above spontaneous levels for more than a minute following a 0.55 s pulse of E2-hexenal. ab5A responds strongly to geranyl acetate, a monoterpene ester, whereas ab5B responds strongly to pentyl acetate and heptanone. ab6A shows a strong response to a number of the selected subset of 12 odors, but 1-octen-3-ol is distinguished by evoking a long-lasting excitation such as that for ab4A. The ab7A neuron shows moderate responses to several odors. None of the 12 odors strongly excited ab4B, ab6B, or ab7B. However, ab6B is excited by 4-methylphenol and inhibited by ethyl acetate and pentyl acetate, and ab7B is mildly excited by ethyl butyrate (de Bruyne, 2001).

Although all of the large basiconic sensilla from which recordings were taken could be clearly identified as ab1, 2, or 3, some of the small basiconic sensilla (23 of 128) did not fall into a well-defined category. The neurons in these sensilla exhibited spontaneous action potentials, and most yielded weak excitatory responses to at least some odors, but none responded strongly to any of the 47 odors in the full panel (de Bruyne, 2001).

16 classes of ORNs have been found on the antenna. How are these classes distributed spatially? As useful coordinates in describing the distributions of these functional units, regions drawn on the basis of sensillar morphology are referred to. Specifically, the anterior and posterior surfaces of the third antennal segment are covered with olfactory sensilla, but the sensory field is heterogeneous in terms of sensillar morphology and can be divided accordingly into five regions. Regions I and II contain exclusively large basiconic sensilla. Region III is largely devoid of sensilla, except for a number of coeloconic sensilla. Region IV contains both small basiconic and coeloconic sensilla. Finally, region V contains a high density of trichoid sensilla, but also contains some small basiconic and coeloconic sensilla (de Bruyne, 2001).

Each functional type of sensillum -- and consequently each functional class of neuron -- is restricted to a particular spatial domain of the antennal surface. Although certain types are intermingled, each is compartmentalized by particular spatial boundaries. Ab1, ab2, and ab3 are restricted to regions I and II, whereas the four remaining sensillar types are localized exclusively in regions IV and V (this segregation follows from the fact that ab1, ab2, and ab3 are the only large basiconic sensillar types, and from the definition of the regions). However, ab1, ab2, and ab3 do not show identical distributions. ab3 shows a more restricted distribution; it is limited to region I and, moreover, is restricted to the dorso-medial portion of region I. The distribution of ab1 overlaps closely with that of ab2. ab4 and ab6 are found in region IV but not region V, and their distributions appear similar. ab5 and ab7 are found in both regions IV and V. Thus, four functional domains can be distinguished: those for (1) ab1 and ab2; (2) ab3; (3) ab4 and ab6 and (4) ab5 and ab7. Within a functional domain the distributions of functional types may not be identical, but a more detailed analysis will be required to establish these distinctions conclusively (de Bruyne, 2001).

How many neurons of each functional class are located on the antenna? Among the large basiconic sensilla, 50% were identified as ab1, with 30% as ab2 and 20% as ab3. The results suggest that there are on the order of 45 ab1A, B, C, and D neurons, 27 of the ab2 neurons, and 18 of the ab3 neurons. This estimate predicts that 50% of large basiconic sensilla would contain four neurons, which is in excellent agreement with the estimate of 52% from ultrastructural studies of males. Similar calculations suggest that there are 13-28 of the neuronal types in small basiconic sensilla, again assuming that the sampling bias is not large (de Bruyne, 2001).

What relationship is there between the number of ORN classes and the number of glomeruli? The most informative data are available for the ORNs within large basiconic sensilla. Eight classes of these ORNs have been defined and no evidence has been found for any additional classes. Approximately seven glomeruli are indicated as receiving input from the region of the antenna containing large basiconic sensilla. While there are some uncertainties in comparing the two studies, it is nonetheless striking that the number of ORN classes is similar to the number of identified glomeruli. This numerical similarity is consistent with the finding that neurons expressing a particular odorant receptor project to one or two glomeruli, both in vertebrates and in Drosophila. Furthermore, the total number of ORN classes identified in this study, 16, when added to the 6 classes characterized in an extensive analysis of maxillary palp sensilla (de Bruyne, 1999) and the 7 classes for which evidence is found in a preliminary characterization of trichoid and coeloconic sensilla on the antenna, yields a total of 29, a number which approaches 41, the number of olfactory glomeruli that have been identified in Drosophila (de Bruyne, 2001).

An approximate numerical correspondence is also observed between the number of neurons per ORN class and the number of ORNs expressing an individual odor receptor gene. Specifically, it is estimated that the number of neurons in a functional class ranges from 13 to 45. In situ hybridization experiments have shown that among four individual odor receptor genes tested quantitatively, each was expressed in 16 to 40 cells per antenna, a figure that is in reasonable agreement with data for a number of other receptor genes. This numerical correspondence is consistent with the concept that neurons of a particular functional class, i.e., neurons with a particular odor response spectrum, express the same odor receptor gene(s) (de Bruyne, 2001).

Three variables describe an olfactory stimulus: odor identity (analogous to hue), odor concentration (intensity), and time. Each of these variables is represented in the action potential frequency of different receptor neurons in the Drosophila antenna. Different ORN classes show diverse odor response spectra, sensitivities, and dynamics. Correspondingly, a particular odor signal produces a different response among different ORN classes (de Bruyne, 2001).

Although individual ORNs may have evolved to respond to particular ligands of ecological importance, 'cross-stimulation' of ORNs by other ligands is likely to be critical in the coding of odor identity. Of the 16 ORN classes defined in this study, 11 responded to more than one of the test stimuli. Conversely, most (73%) of the test stimuli that elicited a response from any ORN elicited a response from multiple ORNs, and 23% elicited a response from five or more ORN classes. The response of individual ORNs to multiple odors, and the response of multiple ORNs to individual odors, is consistent with a number of other studies of vertebrate ORNs and some other insect ORNs. An example of how odor identity could be encoded by integrating the responses of multiple ORNs is illustrated by the responses of ab1A and ab3A to ethyl acetate and ethyl propionate. The ORN that is most sensitive to ethyl acetate, ab1A, is also highly sensitive to ethyl propionate. How then can the fly discriminate between a high dose of ethyl acetate and ethyl propionate? One possible answer is by integrating the response of ab1A with that of ab3A, which responds to ethyl propionate but not ethyl acetate (de Bruyne, 2001).

Is the structure of the code arbitrary? In principle, if each ORN class represented a binary coding unit, the 16 classes described here could encode 216, or some 65,000, different odors. However, in interpreting the response matrix, it is useful to consider that the ORN repertoire of Drosophila is not designed for maximal computational efficiency across the entirety of odor space, but rather has evolved to fit the organism's need for survival and reproduction. Esters and alcohols, which are commonly present in fermenting fruits, elicit responses from a relatively large number of neurons, which may thereby provide resolving power to aid in discrimination among these odors (de Bruyne, 2001).

Coding of odor concentration is also likely to depend upon the multiplicity of responding ORNs. For example, ab1A is very sensitive to ethyl acetate, saturating at a 10^-4 dilution. ab2A is less sensitive, showing a linear relationship that initiates at 10^-4 and remains linear until 10^-2; thus the presence of two ethyl acetate neurons of differing sensitivities effectively expands the dynamic range of the response. Low doses of ethyl acetate are likely to be encoded by ab1A; at higher doses, not only is a stronger signal sent by ab1A to a glomerulus in the antennal lobe, but an additional signal is transmitted by ab2A, presumably to an additional glomerulus. It has been noted that among the dose-response relationships characterized for ORNs of both the antenna and the maxillary palp (de Bruyne, 1999), the slopes are remarkably constant. Moreover, the linear phase of the curve in each case extends over at least a 100-fold range of odor dilutions (de Bruyne, 2001).

Both excitation and inhibition of ORNs by odors have been documented. The ability of a neuron to exhibit two modes of response adds a degree of freedom that may be used in the coding of odor identity. Inhibition may provide a mechanism for enhancing signal recognition; for example, an odor may inhibit some ORNs while exciting others, which may enhance contrast in the glomeruli. In fact, linalool can inhibit firing of the ab2A neuron, but can excite the ab4A, ab6A, and ab7A neurons. Linalool is a terpene common among plants and excites ORNs from a variety of insect species; it has previously been shown to inhibit pheromone-sensitive ORNs in a moth (de Bruyne, 2001).

In addition to odor quality and quantity, the temporal profile of an odor stimulus also must be encoded by spike trains. The temporal profile of an odor is critical to flying insects that seek odor sources. Such insects encounter intermittent 'odor pockets' while traversing an odor plume, and the frequency of such encounters is an essential parameter in navigation (de Bruyne, 2001).

How is temporal information encoded? In addition to whether an individual ORN fires, and the frequency of firing, a third degree of freedom is the distribution of spikes in time. The spike frequency is not a simple reflection of the instantaneous odor concentration. For example, in some cases the spike train terminates abruptly after stimulus termination, and in other cases the spike train continues long after the odor stimulus has ceased; these radically different patterns can be observed for two neurons in the same sensillum, stimulated with the same odor. The underlying molecular basis of these kinetic differences is not known. One possibility is that they relate to differences in ligand-receptor binding affinities; another possibility is that they represent differences in the kinetics of receptor deactivation, with post-stimulus inhibition perhaps representing an extreme form of receptor deactivation. In any case, this third degree of freedom could provide a means of expanding the coding capacity of the system. Fast off-responses may convey information about rapid changes in odor concentration, such as those encountered by a flying insect. Slow off-responses may play another role in the fly's orientation behavior, perhaps by providing a 'memory' of a recently encountered odor stimulus. Thus, the two different response modes may simultaneously present a phasic and tonic representation of selected features of the odor environment (de Bruyne, 2001).

The spatial organization of the antennal domains, as determined by physiological recording, are of interest in a developmental context. The boundaries of these domains are roughly perpendicular to the axis that extends from the dorso-medial portion of the antenna to the ventro-lateral portion. Thus, ab3 sensilla are the most dorso-medial, followed by ab1 and ab2, with ab4 and ab6 occupying an intermediate domain, and ab5 and ab7 located more ventro-laterally. This axis is parallel to the direction of many mitotic clones observed in developmental studies of the antenna and to the orientation of individual sensilla. It is perpendicular to the boundaries between a number of sensillar types as defined anatomically. Not only do large basiconic sensilla lie dorso-medial to the small basiconic sensilla, but at the dorso-medial extreme lies an ultrastructural subclass of large basiconic sensilla called LBI-1, which corresponds well in distribution and number to the ab3 sensilla. Another ultrastructural subclass of basiconic sensilla, TB-2, shows a distribution pattern similar to that of the ab4 and ab6 sensilla. It is noted also that the distribution of most of the sensilla that could not be classified is similar to that of 'sensilla intermedia,' a type of sensillum which, by electron microscopy, exhibits features intermediate between basiconic and trichoid sensilla. It is tempting to speculate that the different sensillar classes that have been defined functionally can also be distinguished anatomically, and that their distinct characteristics arise during development by differential expression of many genes, including receptor genes, along a morphogenetic axis (de Bruyne, 2001).

Chemosensory coding by neurons in the coeloconic sensilla of the Drosophila antenna

Odor coding is based on the diverse sensitivities and response properties of olfactory receptor neurons (ORNs). In the Drosophila antenna, ORNs are housed in three major morphological types of sensilla. Although investigation of the Drosophila olfactory system has been expanding rapidly, the ORNs in one of these types, the coeloconic sensilla, have been essentially unexplored. Four functional types of coeloconic sensilla were defined through extracellular physiological recordings. Each type contains at least two neurons, with a total of at least seven distinct ORN classes that vary remarkably in their breadth of tuning. Analysis of 315 odorant-ORN combinations reveals how these neurons sample odor space via both excitation and inhibition. A class of neurons was defined that is narrowly tuned to small amines, and humidity detectors were found that define a cellular basis for hygroreception in Drosophila. The temporal dynamics of responses vary widely, enhancing the potential for complexity in the odor code. Molecular and genetic analysis shows that a broadly tuned ORN, antennal coeloconic 3B (ac3B), requires the odor receptor gene Or35a for its response in vivo. The activity of ac3B is not required for the response of the other ORN within that sensillum, ac3A. The functional analysis presented here, revealing a combination of highly specialized neurons and a broadly tuned ORN, along with the ancient origin of coeloconic sensilla, suggests that the specificities of these ORNs may reflect basic needs of an ancestral insect (Yao, 2006; Full text of article).

The first step in the coding of an olfactory stimulus is the activation of ORNs in olfactory organs. This study examined in detail the coding of olfactory information by ORNs in a major class of olfactory sensilla, the coeloconic sensilla, whose function in Drosophila has until now been essentially unexplored (Yao, 2006).

Physiological recordings identified four distinct functional types of coeloconic sensilla on the antennal surface. Each type contains at least two ORNs, and the data show that there are at least seven distinguishable classes of ORNs; it was not possible to distinguish with confidence the activities of the different ORNs in ac4 sensilla, but it is clear from the response to phenylacetaldehyde, for example, that at least one of the ORNs in ac4 is distinct from those in ac1, ac2, and ac3. It is noted that an anatomical study of coeloconic sensilla revealed two subtypes: one containing two neurons and the other containing three neurons. It thus seems likely that at least one functional type, such as ac4, contains three ORNs (Yao, 2006).

Neurons were identified in the Drosophila antenna that respond strongly to humidity. Hygroreception is critical in the insect world. Many insects, such as mosquitoes, lay eggs in water, and dry environments can lead to desiccation and death. Drosophila exhibits preferences when given a choice between environments of different humidities; however, the molecular and genetic basis of hygroreception is not understood, and even its cellular basis has until now been unknown in Drosophila (Yao, 2006).

In defining a cellular basis of hygroreception in Drosophila, a focus is provided for a molecular and genetic analysis of the mechanism of transduction. It is noted that ablation experiments had provided evidence in a previous study that hygroreception localizes to the arista, a feathery projection extending from the third antennal segment. However, the current study also found that spineless^aristapedia, a homeotic mutant in which aristae are transformed into leg-like structures, had normal hygroreception. These studies do not exclude the possibility of hygroreceptive cells associated with the arista or elsewhere; however, identification of hygroreceptors in coeloconic sensilla is consistent with studies of other insects in which hygroreceptive cells were identified in morphologically similar sensilla (Yao, 2006).

The temporal dynamics of humidity responses appear largely tonic. The responses to odors vary in their dynamics. The diverse temporal dynamics shown by coeloconic ORNs may reflect another degree of freedom used by the antenna in odor coding: the temporal structure of olfactory information has been shown to be critical in odor coding in a variety of systems (Yao, 2006).

It is noted with interest that, although the transduction mechanisms underlying hygroreception and olfaction may differ, they are likely to be housed in a common cell (i.e., in a multimodal neuron: isoamylamine inhibits all neuronal activity in ac2 sensilla), including that of humidity-sensitive cells, suggesting that hygroreceptive cells contain a receptor for isoamylamine (Yao, 2006).

Amine detectors were also detected. One class of ORN, ac1A, is highly sensitive to ammonia, and another, ac2A, responds strongly to 1,4-diaminobutane. A particularly important example of amine detection in insects is the attraction of mosquitoes to ammonia, a component of human sweat. Ammonia-sensitive cells have been detected in double-walled, grooved-peg sensilla on the antenna of Anopheles mosquitoes and Triatoma infestans, a vector of trypanosomiasis, or Chagas disease. Identification of these cells in Drosophila lays a foundation for examining the molecular genetics of amine detection, which, in turn, could be useful in designing new means of pest control (Yao, 2006).

The response spectra of the coeloconic ORNs vary remarkably in their apparent breadth of tuning. At one extreme is ac3B, which reveals exceptionally broad tuning. ac3B is excited by 36 of 45 odors at the test concentrations; there is evidence that some coeloconic sensilla in the moth Bombyx mori yield responses to a variety of acids and alcohols. ac1B, in contrast, is excited by none of these 45 odors; it may be the cell in ac1 that responds to humidity. ac1A responds strongly to ammonia and weakly to two other amines, but shows no responses to the other 42 tested odors. The profile of ac2A suggests that it may have evolved to signal the presence of 1,4-diaminobutane. No strong excitatory responses were detected among the neurons of ac4, other than a modest response to phenylacetaldehyde, and it seems likely that they detect the presence of biologically significant stimuli that are not included in the odor set. Conclusions about the breadth of ORN tuning are limited by the number of odors tested. Odor space is vast and discontinuous, and the sampling of odors in this study is necessarily limited. Nonetheless, the pattern that emerges from this analysis is one of a single broadly tuned ORN, ac3B, and others that are excited by only one or a small number of stimuli (Yao, 2006).

In light of the widely varying tuning breadths of these ORNs, it seems plausible that the coeloconic sensilla have evolved in large part to signal the presence of a small number of specific chemosensory stimuli, such as water vapor, ammonia, and 1,4-diaminobutane, also known as putrescine. The exceptionally broad spectrum of ac3B could have evolved as a more general sensor to signal the simultaneous presence of food sources. The coeloconic sensilla have an ancient origin, and perhaps their specificities reflect the most basic needs of an ancestral insect (Yao, 2006).

The specificities of these neurons are also of interest in light of the recent construction of a receptor-to-neuron map for the Drosophila antenna. Or genes were individually expressed in a mutant in which a particular ORN, ab3A, loses its odorant response because of a deletion that removes two odorant receptor genes, Or22a and Or22b. When particular Or genes were expressed in this mutant ORN, they conferred in many cases an odor response spectrum that matched that of a defined ORN. In this way, individual Or genes could be assigned to particular ORNs. Testing of virtually the entire repertoire of antennal Or genes mapped many receptor genes to basiconic ORNs, but only one, Or35a, to a coeloconic ORN. Because a number of receptors in that study conferred response profiles that did not match those of any defined ORNs, these unmapped receptors were further tested, using a panel of odorants from that the present study has identified as diagnostic for particular coeloconic ORN classes. However, it was not possible to map any of these receptors to coeloconic ORNs (Yao, 2006).

Why has only a single class of coeloconic ORN been found to derive its odor response profile from an Or gene? One possible interpretation is that some coeloconic ORNs, for example those sensitive to humidity or ammonia, do not rely on receptors of the Or family to detect the molecules that activate them. It is noted that no Or gene was mapped to ab1C, a basiconic ORN that is a sensor of CO₂. A Gr (Gustatory receptor) gene has been mapped to this neuron, although at present there is no evidence that it is a CO₂ receptor. In the present study, it was shown that Or35a is necessary for the response of ac3B. An important goal now is to determine which receptors are necessary for the responses of the other coeloconic neurons. The analysis of coeloconic ORNs presented in this study provides a foundation for investigating the molecular and genetic basis of the mechanisms by which coeloconic ORNs transduce chemosensory signals. It also provides a basis for investigating the roles of these neurons in olfactory-driven behaviors (Yao, 2006).

Molecular, anatomical, and functional organization of the Drosophila olfactory system

Olfactory receptor neurons (ORNs) convey chemical information into the brain, producing internal representations of odors detected in peripheral receptors. A comprehensive understanding of the molecular and neural mechanisms of odor detection and processing requires complete maps of odorant receptor (Or) expression and ORN connectivity, preferably at single-cell resolution. Near-complete maps have been constructed of Or expression and ORN targeting in the Drosophila olfactory system. These maps confirm the general validity of the 'one neuron-one receptor' and 'one glomerulus-one receptor' principles and reveal several additional features of olfactory organization. ORNs in distinct sensilla types project to distinct regions of the antennal lobe, but neighbor relations are not preserved. ORNs grouped in the same sensilla do not express similar receptors, but similar receptors tend to map to closely appositioned glomeruli in the antennal lobe. This organization may serve to ensure that odor representations are dispersed in the periphery but clustered centrally. Integrated with electrophysiological data, these maps also predict glomerular representations of specific odorants. Representations of aliphatic and aromatic compounds are spatially segregated, with those of aliphatic compounds arranged topographically according to carbon chain length. These Or expression and ORN connectivity maps provide further insight into the molecular, anatomical, and functional organization of the Drosophila olfactory system. These maps also provide an essential resource for investigating how internal odor representations are generated and how they are further processed and transmitted to higher brain centers (Couto, 2005).

The 60 Or genes of Drosophila are predicted to encode a total of 62 odorant receptors, from transcripts originating from 62 distinct Or promoters. A set of mCD8-GFP reporter lines was constructed for all 62 promoters, as well as 59 GAL4 reporter lines. In parallel, a set of in situ hybridization experiments was performed to detect mRNA for 54 distinct Or genes in sections of the adult olfactory organs, the antennae and maxillary palps. For all Or genes for which expression was detected by both of these methods, sections were also probed simultaneously for the Or mRNA and the corresponding Or transgenic reporter. In all, 42 Or genes were detected for which the reporter line and in situ hybridization labeled identical subsets of ORNs. None of these Or genes is sex specific. In addition, the Or83b reporter is broadly expressed in ORNs, as is Or83b itself. No validated reporter was detected for five of the Or genes known to be expressed in the antenna or maxillary palp (Or1a, Or33a, Or71a, Or85b, and Or98b); a validated reporter for Or71a has already been obtained in another study. Thus, the expression of 44 Or genes can be mapped in the antenna and maxillary palp (Couto, 2005).

A related family of 60 Gr genes encodes 68 predicted gustatory receptors, some of which may actually function as odorant receptors. A systematic survey of Gr expression in the olfactory system was performed, preparing mCD8-GFP reporter lines for 67 of 68 Gr promoters. Only Gr21a could be mapped to specific ORNs by both reporter expression and in situ hybridization. Several other Gr reporters are also expressed in the antenna, but their expression is generally either weak or broad or could not be confirmed by in situ hybridization. It is therefore concluded that few if any of the other Gr genes are likely to encode odorant receptors (Couto, 2005).

The larval olfactory system comprises just 21 ORNs, which also expresses members of the Or gene family. Since 14 of the Or reporters could not be detected in the adult, it is suspected that these Or genes may in fact encode larval odorant receptors. Indeed, 11 of these reporters are expressed specifically in the larval olfactory system. This search was expanded to cover all Or reporters and another seven Or genes were identified that are expressed in the larval as well as the adult olfactory system. This analysis confirmed 8 of the 10 larval Or genes previously identified by both RT-PCR and reporter experiments, but only 3 of the additional 13 receptors with RT-PCR evidence alone. Thus, a total of 20 larval Or genes have been identified (in addition to Or83b). This is a close match to the 21 ORNs, consistent with the notion that, in the larva, each ORN expresses a single and distinct odorant receptor. This analysis also confirms that the larval and adult odorant receptors are encoded by phylogenetically dispersed members of the same Or family, some of which are stage specific whereas others are common to both stages (Couto, 2005).

In the adult, ORNs are housed in sensory sensilla of four distinct morphological types: basiconic, trichoid, coeloconic, and intermediate sensilla (in order of decreasing abundance). All four sensilla types are found on the antenna. The maxillary palp contains only basiconic sensilla. Each adult Or gene was mapped to a specific sensillum type simply by noting the morphology of the sensilla innervated by GFP-positive dendrites in Or-mCD8-GFP or Or-GAL4, UAS-mCD8-GFP flies (Couto, 2005).

Most sensilla contain 2-4 ORNs, so attempts were made to identify pairs of Or genes expressed in distinct ORNs of the same sensillum. For this, double labelings were performed for most possible pairings of Or genes within both the basiconic and trichoid classes. This generally involved immunofluorescence detection of one Or transgenic reporter and in situ hybridization to detect the second Or, although in some cases either double in situ hybridization or double immunofluoresence were performed (using the mCD8-GFP reporter for one Or and the GAL4 reporter and UAS-τlacZ for the other). These methods allowed the mapping of 44 Or genes and Gr21a to 38 ORN classes that innervate 19 distinct and highly stereotyped sensilla types (Couto, 2005).

The basiconic sensilla of the antenna comprise three large (ab1-3), three thin (ab4-6), and four small sensilla (ab7-10), following the nomenclature used for their morphological and physiological classification . The sensilla classes define as ab9 and ab10 were not identified in the physiological studies, possibly because they are located more laterally on the antenna. The expression was confirmed of Or genes in 11 antennal basiconic ORN classes, most of which had previously only been inferred from functional data, and Or genes were identified for another 11 ORN classes. The map of the maxillary palp comprises three thin basiconic sensilla (pb1–3), with Or genes assigned to all six ORN classes. In summary, 31 receptor genes expressed in the basiconic ORNs were identified, defining a total of 28 distinct ORN classes and 13 sensilla classes. With the exception of one neuron in the ab6 sensillum, the receptor-to-neuron map of the basiconic sensilla of the antenna and palp is now most likely complete (Couto, 2005).

The trichoid sensilla are innervated by one, two, or three neurons (T1, T2, or T3 sensilla, respectively). There are only very limited physiological data available for the trichoid sensilla, and no receptor-to-neuron assignments have previously been made. Twelve Or genes expressed in 9 ORN classes in 4 distinct classes of trichoid sensilla were identified: a single T1 sensillum (which is referred to as at1), a single T2 sensillum (at2), and two distinct classes of T3 sensillum (at3 and at4). The compositions of at2 and at3 seem to be strictly stereotyped, but at4 may be slightly variable. The total numbers of trichoid sensilla and ORNs that were identified closely match the numbers reported in the morphological survey, and since at least one Or was assigned to each ORN class, it is anticipated that this receptor-to-neuron map of the trichoid sensilla is also likely to be complete (Couto, 2005).

Coeloconic sensilla house either two or three ORNs, and based on the number of glomeruli innervated by coeloconic ORNs, it is anticipated that there are eight distinct classes of coeloconic ORNs. However, only a single Or gene, Or35a, could be assigned to the coeloconic sensilla. Or or Gr genes might be expressed only at very low levels in the coeloconic ORNs, making them difficult to detect by these methods. Alternatively, these neurons may express some other type of chemoreceptor (Couto, 2005).

The intermediate sensilla number only about 20-30, and also contain either 2 or 3 ORNs. Or13a was tentatively assigned to the intermediate sensilla, although it is also possible that one or another of the sensilla classes identified as basiconic or trichoid by light microscopy may in fact correspond to sensilla described as intermediate by electron microscopy (Couto, 2005).

What logic, if any, guides the selection and pairing of Or genes in individual sensilla? With regard to sensillum type, one might predict that the different sensilla types would express different subfamilies of Or genes, as defined either by the sequences of the receptors they encode or their chromosomal locations. However, no obvious pattern emerged when either a phylogenetic tree or a genomic map of Or genes was annotated with the corresponding sensillum type. With regard to the specific combinations of receptors expressed in ORNs of the same sensilla, two extreme possibilities are that paired receptors might be closely related (because the odorants they detect must pass through a shared extracellular enviroment, and so may be chemically related) or maximally divergent (in order to minimize passive interference between ORNs. To test these hypotheses, the sequence distance was determined between two receptors for each of the 990 possible pairs of the 45 odorant receptors on the map, and each pair was binned into one of four categories: pairs expressed in the same ORN class (e.g., both in ab3A), in distinct ORNs of the same sensillum (e.g., ab3A and ab3B), in different sensilla classes of the same type (e.g., ab3A and ab4A), or in different sensilla types (e.g., ab3A and at1A). It was found that those receptors expressed in different neurons of the same sensillum are, on average, no more and no less closely related to each other than those expressed in different sensilla, different types of sensilla, or indeed any receptor pair chosen at random. Thus, although highly stereotyped, the selection and pairing of Or genes into distinct sensilla types and classes does not seem to follow any particular logic with regard to either the sequence of the receptor or the location of its gene (Couto, 2005).

Anatomical studies using general synaptic markers have defined some 40-50 glomeruli in the Drosophila antennal lobe, with some minor discrepancies between different studies. The Or transgenic reporter lines provided a set of molecular markers for individual glomeruli; these reporters can be used individually or in combination to refine existing maps and establish an atlas of 49 glomeruli. To assign receptors to individual glomeruli, ORN projections were examined for each of the promoter-mCD8-GFP fusions. Previously, this approach had been used to map receptors to 13 glomeruli. The set of 44 verified adult Or reporters facilitated correction of 4 of these earlier assignments and to extend the coverage to a total of 37 glomerul. For each Or, the projections were identical in each animal examined, with no differences between the sexes. Each Or reporter also labels just a single glomerulus (although a few reporters are weakly or ectopically expressed in additional ORN classes that target other glomeruli). Through a series of unilateral deafferentation experiments, it was determined that only the V glomerulus is innervated unilaterally; all other glomeruli on the map receive bilateral innervation (Couto, 2005).

It was anticipated that many of the glomeruli that remained unassigned were likely to be innervated by coeloconic ORNs. To identify these glomeruli, an ato-GAL4 reporter was used: this reporter is expressed in all coeloconic ORNs of the antenna and the basiconic ORNs of the maxillary palp. A total of 14 glomeruli are labeled in ato-GAL4, UAS-mCD8-GFP flies: the six glomeruli already assigned to the six ORN classes of the maxillary palp, the one glomerulus assigned to an antennal coeloconic ORN (VC3), and seven still unassigned glomeruli (DC4, DL2, VL1, VL2a, VM1, VM4, and VM6). It is therefore inferred that VC3 and these seven additional glomeruli are the targets of the antennal coeloconic ORNs (Couto, 2005).

The ORN connectivity map reveals a spatial organization in the antennal lobe that was not apparent from the few ORN classes previously examined. Specifically, afferents from ORNs in distinct sensilla types project to distinct regions of the antennal lobe: ORNs in antennal trichoid sensilla project to the lateral anterior region, antennal basiconic sensilla to the medial region, palp basiconic sensilla to the central-medial region, and antennal coeloconic sensilla to the posterior (Couto, 2005).

This segregation of sensory input according to sensillum type does not extend down to the level of individual sensilla classes. Specifically, ORNs that are neighbors in the same sensillum do not always project to neighboring glomeruli. Nevertheless, ORNs in the same sensilla class might still innervate glomeruli that are generally close to each other within the antennal lobe. To test this possibility, the distances between all 666 pairs of the 37 assigned glomeruli was determined and then it was asked whether glomeruli innervated by ORNs in the same sensilla are generally closer to each other than those innervated by ORNs in different sensilla. Distances were calculated in two different ways for each pair. An average physical distance between the geometric centers of the two glomeruli was determined from 3D reconstructions of four male antennal lobes. In addition, the 'degrees of separation' between each pair was determined: 1 for neighboring glomeruli, 2 for glomeruli that are not themselves neighbors but have a common neighbor, and so on. It was found that, by either distance measure, pairs of glomeruli innervated by ORNs in the same sensilla are, on average, no closer together or further apart than those innervated by ORNs from different sensilla of the same type. Thus, sensory inputs from the different sensilla types are segregated into distinct regions of the antennal lobe, but within each of these regions the arrangement of glomeruli bears no obvious relationship to the location or pairing of ORNs in the periphery (Couto, 2005).

It was next asked whether receptors that are more closely related by sequence tend to map to glomeruli that are physically closer within the antennal lobe. To test this, all 938 possible pairs of the 44 odorant receptors on the map were examined (excluding the eight pairs that are coexpressed in the same neuron). For each pair, the sequence divergence of the two receptors and the separation of the corresponding glomeruli in the antennal lobe were determined. There is a strong positive correlation between the two, using either the actual distances between pairs of glomeruli or their degrees of separation. This correlation can be attributed entirely to the antennal basiconic sensilla, as it is not observed at all for the other two sensilla types . Thus, among the antennal basiconic sensilla, pairs of ORNs that express more closely related receptors tend to map to more closely positioned glomeruli (Couto, 2005).

By integrating the molecular and anatomical maps with existing electrophysiological data, odor-evoked activity patterns could be predicted for a total of 29 glomeruli. Each of these 29 glomeruli were classified according to whether the test odorants that elicited a strong response (above a threshold of 50 spikes/second) were linear aliphatic compounds or aromatic compounds containing a benzene ring. This analysis suggested a spatial separation of aliphatic and aromatic odor representations in the antennal lobe. Glomeruli that respond primarily to aromatic odorants are clustered in a ventral-central region of the antennal lobe, whereas those that respond preferentially to aliphatic odorants are clustered in the medial region. This clustering does not bear any relationship to the clustering of inputs from the different sensilla types (Couto, 2005).

The test odorants used in these physiological studies had been selected primarily in order to maximize their chemical diversity, rather than to systematically sample 'odor space'. Esters are, however, particularly well represented in these data sets and range in size from 4 to 12 carbons. Collectively, the odorant receptors or ORNs that are activated above a threshold of 50 spikes/second by these esters map to 16 of the 20 'aliphatic' glomeruli on the map. For each of these 16 glomeruli, the carbon number of the ester that gave the maximum response was noted. This revealed a broad ordering of glomeruli along the posterior to anterior axis, with more anterior glomeruli generally preferring larger esters. A similar trend was also observed for alcohols and ketones, but for these compounds the data are too sparse to draw any strong conclusions (Couto, 2005).

In both mammals and insects, individual ORNs are thought to express only a single functional odorant receptor, although exceptions have been documented in both rats and Drosophila. The critical test of this hypothesis is to map, at single-cell resolution, the expression of the entire family of odorant receptor genes. This task has now been almost completed for the adult olfactory system of Drosophila melanogaster; this study mapped 45 odorant receptors to 38 distinct ORN classes (Couto, 2005).

Only six ORN classes express more than one receptor (excluding the widely expressed Or83b, which heterodimerizes with other odorant receptors but is not functional by itself; as well as the low levels of some additional Or or Gr genes in some neurons). In four of these six cases, the coexpressed Or genes are closely linked and highly conserved, suggesting that they arose through a relatively recent gene duplication. These pairs of coexpressed receptors are likely to detect the same odorants, and so do not represent a meaningful exception to the one neuron-one receptor principle. The two cases of coexpressed but unrelated and unlinked Or genes are Or33c and Or85e in pb2A and Or49a and Or85f in an ab10 neuron. For the Or33c/Or85e pair, both receptors are functional when ectopically expressed, but the response profile of the pb2A neuron in which they are endogenously coexpressed can be attributed to Or85e alone. Similar comparisons are not yet possible for the Or49a/Or85f pair, since there are no electrophysiological data available for Or49a or the ab10 sensillum (Couto, 2005).

In mammals, a negative feedback mechanism ensures that only one functional receptor is expressed. The choice of a specific receptor is largely stochastic, although each ORN is somehow restricted to selecting from a large subset of 'available' receptor genes according to its position in the olfactory epithelium. In contrast, receptor choice in Drosophila appears to be entirely deterministic, as indicated by the highly stereotyped patterns of Or expression in olfactory sensilla. There is also no evidence for any negative-feedback mechanism in Drosophila; the loss of an endogenous receptor does not lead to the expression of an alternative receptor, nor does ectopic expression of a second receptor block the expression of an endogenous receptor (Couto, 2005).

The 45 receptors that were mapped can be paired in nearly 1000 different ways, yet less than 20 distinct combinations are actually deployed in olfactory sensilla. Why have these specific combinations been selected? One possibility is that ORNs compartmentalized into the same sensillum might express closely related receptors, as the odorants they detect are transported and processed by the same set of molecules in their common sensillar lymph, and so may be chemically related. However, this study found that pairs of Or genes expressed in the same sensillum are no more closely related to each other than any randomly selected pair. Similarly, electrophysiological surveys of a more limited set of basiconic sensilla have shown that ORNs housed in the same sensilla tend to have distinct rather than similar response spectra. These observations are more readily explained by a model in which ORNs housed in the same sensilla instead express divergent receptors, so as to minimize their functional overlap and afford each ORN a greater dynamic range (Couto, 2005).

Nevertheless, there are still many different ways in which pairs of divergent odorant receptors could be combined, so this consideration alone cannot completely explain the specific combinations deployed. An additional factor may be that two odorants could be discriminated with a higher spatial and temporal resolution if the ORNs that detect them are placed in the same rather than distinct sensilla, possibly even allowing the insect to discern whether two odorants are present in the same or different filaments of an odor plume. This might be particularly relevant for odorants such as pheromones, for which it may be critically important to distinguish whether the individual components of a blend originate from a single source (a potential mate) or from two closely spaced sources (Couto, 2005).

Whatever the logic behind these pairings, it will be of great interest to determine how they are programmed developmentally. At present, little is known of these mechanisms. ORNs in the same sensillum are likely to be related by lineage, and so selection of a specific Or gene might be part of the instrinsic mechanisms that generate diverse cell fates within each lineage. Alternatively, by analogy to the signaling mechanisms that coordinate rhodopsin gene selection between R7 and R8 cells in the same ommatidium in the eye, one ORN in each sensillum might choose its Or first and then instruct the Or choice of its neighbor(s). The promoter regions that were defined will be a valuable guide in computational and experimental approaches aimed at defining the cis-acting determinants of Or choice, while the transgenic reporters should facilitate genetic screens to identify the trans-acting factors (Couto, 2005).

In Drosophila, axons of ORNs that express the same odorant receptor are thought to converge upon a single glomerulus, with each glomerulus receiving input from just a single class of ORN. The Or axonal reporters generated prior to this study all label a single glomerulus per antennal lobe, as does each of the 45 verified reporters in this study. The few cases in which Or reporters have been observed targeting multiple glomeruli can most likely be explained by low levels of 'ectopic' Or expression or by reporters that do not faithfully mimic the endogenous Or expression. Innervation of a single glomerulus thus appears to be a strict rule (Couto, 2005).

More difficult to verify, in any species, has been the postulate that each glomerulus receives input only from a single class of ORNs (the one glomerulus-one receptor hypothesis). No exceptions have yet been reported in the main olfactory systems of mice or Drosophila. However, with only a small fraction of odorant receptors examined in each case, the chances of detecting a glomerulus with multiple inputs had, until now, been vanishingly small. With the map of ORN connectivity now almost complete for Drosophila, it can now be confirmed that most, and probably all, glomeruli do indeed receive input from just a single class of ORN. Specifically, each of 38 glomeruli could be assigned to a single and distinct ORN class and most of the remaining glomeruli to a nonoverlapping set of unidentified coeloconic ORNs (Couto, 2005).

The connectivity map also reveals an unanticipated topographic organization of the antennal lobe, with ORNs in distinct sensilla types projecting into distinct regions of the antennal lobe. A similar topographic organization may apply in the vertebrate main olfactory system. Within these regions, however, neighbor relationships are not preserved -- ORNs that are neighbors in the same sensillum do not target neighboring glomeruli, or indeed even closely positioned glomeruli (Couto, 2005).

For the antennal basiconic sensilla, the distance between glomeruli does, however, correlate with the sequence distance between the corresponding odorant receptors. Thus, whereas ORNs are grouped into sensilla in ways that appear to favor combinations of divergent receptors, their target glomeruli may be arranged in part in ways that tend to juxtapose ORNs that express similar receptors. This redistribution of ORNs between the periphery and the antennal lobe may contribute to the formation of chemotopic maps in the brain (Couto, 2005).

The peripheral and central mechanisms of Drosophila olfaction have been well described. What has been missing until now is the causal link between the two. The expression and connectivity maps of this study provide this link. The internal representations of specific odorants can now be explained and predicted from the knowledge of the receptors they activate, the ORNs that express these receptors, and the glomeruli that these ORNs target (Couto, 2005).

This study has predicted odor respresentations covering 29 glomeruli for the diverse set of odorants used in the physiological studies. These odor maps reveal a functional organization of the antennal lobe that is not apparent from imaging studies with Drosophila but is consistent with imaging and electrophysiological data from other insects. Specifically, aromatic and aliphatic compounds are predicted to activate spatially distinct regions of the antennal lobe. A similar segregation of aromatic and aliphatic representations has also been suggested for the larval olfactory system. It was also found that, within the 'aliphatic cluster' of the adult antennal lobe, compounds of increasing carbon chain length are predicted to successively shift the activity pattern in an anterior direction (Couto, 2005).

These features are also not unique to insect olfactory systems. Accumulating evidence points to a similar functional organization of the mammalian olfactory bulb, with distinct chemical classes activating distinct glomerular clusters and carbon chain length represented topographically within each cluster. Thus, the chemotopic organization of the antennal lobe that emerges from the map of the Drosophila olfactory system appears to be a common feature of both insect and mammalian olfactory systems (Couto, 2005).

In the antennal lobe, ORN axons synapse with second order projection neurons (PNs), which extend axons to the protocerebrum. Since high-resolution anatomical maps are beginning to emerge for PN axons, it may soon be possible to predict odor representations at higher brain levels as well. It will be fascinating to learn to what extent this chemotopic map is retained or transformed at higher levels. Before doing so, however, it will be essential to understand the transformations that take place within the antennal lobe itself. Imaging studies have indicated a high degree of correlation between the ORN and PN responses for individual glomeruli, suggesting that the antennal lobe is primarily a relay station with little transformation of olfactory information. In contrast, electrophysiological data suggest that PNs are more broadly tuned and dynamic in their responses than the corresponding ORNs. A caveat to this result, however, was that ORN and PN responses could be compared only for a single glomerulus, DM2 (Couto, 2005).

The connectivity map will now allow more systematic comparisons of ORN input and PN output in the antennal lobe. The electrophysiological data for PNs are still too limited to significantly extend this analysis. Nevertheless, ORN versus PN comparisons can be made for two additional glomeruli: VA7l and DM1. The VA7l PNs appear to be more broadly tuned than their presynaptic pb2B ORNs, whereas the DM1 PNs seem to be a much closer match to the corresponding ORNs, most likely of the ab1B class. Thus, some odors may undergo complex transformations in the antennal lobe, whereas others may be transmitted to higher brain centers with little further processing (Couto, 2005).

In conclusion, these detailed maps of Or expression and ORN connectivity have not only confirmed and extended understanding of the basic molecular and anatomical principles of the olfactory system, they also provide a framework for understanding its functional organization. With these maps, it is now possible to explain and predict how the peripheral activation of odorant receptors produces a chemotopic represention in the antennal lobe. In future, these maps can also be used to determine precisely how olfactory information is further processed in the antennal lobe and transmitted to higher brain centers (Couto, 2005).

Precise and fuzzy coding by olfactory sensory neurons

The exact nature of the olfactory signals that arrive in the brain from the periphery, and their reproducibility, remain essentially unknown. In most organisms, the sheer number of olfactory sensory neurons (OSNs) makes it impossible to measure the individual responses of the entire population. The individual in situ electrophysiological activity of OSNs in Drosophila larvae were measured in response to stimulation with 10 aliphatic odors (alcohols and esters). Control larvae (a total of 296 OSNs) and larvae with a single functional OSN were studied, created using the Gal4-upstream activator sequence system. Most OSNs showed consistent, precise responses (either excitation or inhibition) in response to a given odor. Some OSNs also showed qualitatively variable responses ('fuzzy coding'). This robust variability was an intrinsic property of these neurons: it was not attributable to odor type, concentration, stimulus duration, genotype, or interindividual differences, and was seen in control larvae and in larvae with one and two functional OSNs. It is concluded that in Drosophila larvae the peripheral code combines precise coding with fuzzy, stochastic responses in which neurons show qualitative variability in their responses to a given odor. It is hypothesized that fuzzy coding occurs in other organisms, is translated into differing degrees of activation of the glomeruli, and forms a key component of response variability in the first stages of olfactory processing (Hoare, 2008).

Against expectations, it was found that in whole Drosophila larvae, responses to a narrow range of ecologically significant odors involved a mixture of (1) precise, reproducible neuronal coding in which specific neurons were consistently excited, inhibited, or never responded, and (2) variable, fuzzy coding, in which some neurons responded inconsistently to certain odors. The result is a noisy peripheral signal, in which patterns of OSN activity are rarely repeated, even when an identical stimulus is presented. The first assumption was that these responses must reflect some kind of artifact. However, control experiments showed that these phenomena were not a function of odor, of concentration, of stimulus duration, or of stimulus delivery, nor were they related to differences in responsiveness shown by OSNs or by individual larvae. Strong evidence for the existence of fuzzy coding is provided by the existence of such profiles in [OrX-Gal4/UAS-Or83b; Or83b^-/-] larvae in which only one OSN is functional and in which the activity of the given OSN is readily identifiable (this also shows that the procedure for detecting the activity of single OSNs in a multiunit recording was not the cause of the observation of qualitative response variability). Following the prediction, the typical response profiles seen in single-functional OSN larvae of two different OSN classes were also found in larvae in which these two OSNs were expressed together. Gal4 lines can vary in penetrance across a population or between lines, thus the existence of fuzzy responses could be produced by low or absent OR83b levels in particular individuals or lines. It found that given single-functional OSN lines consistently showed both precise and fuzzy responses and that all individuals of that genotype showed both types of response, strongly suggesting that the results are not caused by differential penetrance of Or83b across or within lines. Finally, it should be recalled that stochastic activity in response to certain odors was also observed in the full complement of larval OSNs in two unmanipulated control strains: this was not a phenomenon that was limited to genetically manipulated strains (Hoare, 2008).

Faced with this accumulation of evidence it was conclude that, as expected, OSNs can show precise, consistent responses to odors, but also that they can show stochastic, fuzzy responses to other odors. Responses of mouse OSNs expressing MOR71 shows consistent calcium increases when stimulated with acetophenone, but only 33% showed such a response to benzaldehyde. Another study found that response thresholds to lyral in MOR23 mouse OSNs varied over three log units of odor concentration. A third study found an intraclass correlation of only 0.65 in the responses of adult Drosophila OSN types and also described heterogeneity in the response profile of OSNs projecting to the same glomerulus. Finally, in a clear parallel to the current findings, it has been reported that Anopheles gambiae TE1A OSNs responded >80% of the time to 4-ethylphenol but <20% of the time to pentatonic acid. Overall, these studies support the current findings and are given coherence by the interpretation that some OSNs, when presented with some odors, can show stochastic or fuzzy responses (Hoare, 2008).

The mechanism underlying fuzzy coding is unknown. It is speculated that fuzzy conding may reflect less effective receptor-ligand binding at the 'edge' of the molecular receptive range of a receptor, or that lateral peripheral interactions between OSNs may shape the response. No evidence was found for such peripheral interactions in [Or42a-Gal4/UAS-Or83b; Or83b^-/-] larvae (stimulation with an odor led to no significant change in the activity of the nonresponsive Or83b^-/- OSNs), but this may merely indicate that such peripheral interactions require the presence of a functioning receptor in the OSN membrane. In this case, it is possible that fuzzy coding in wild-type 21-OSN larvae reflects the existence of lateral interactions between OSNs, whereas fuzzy coding in single OSN larvae would be attributable to the lack of such interactions because the nonfunctional OSNs present a constant signal. Further electrophysiological, genetic, and pharmacological investigations will be required to test this possibility (Hoare, 2008).

The function of fuzzy coding is also unclear, although behavioral data suggest it may form part of the peripheral code in this organism. Recent studies of central processing have suggested that the introduction of noise via lateral inhibition and excitation may be involved in gain control of faint signals. The existence of noise in a subset of OSNs that respond to a particular odor may increase this effect. In organisms in which there is more than one OSN for each Or (that is, for the vast majority of organisms but not the Drosophila larva), the consequence of fuzzy coding would presumably be an intermediate level of excitation in the appropriate glomerulus, since only a proportion of the OSNs would respond to a given presentation of an odor. In this way, fuzzy coding may represent a means by which greater signal variability is introduced into central processing by peripheral events. This is reinforced by the growing appreciation that variability (noise) may be an essential component of neuronal function, in particular for synchronizing the activity of groups of neurons and by data from models of the activity of sensory neuronal networks, which suggest that fuzzy coding may be involved in processing incomplete or variable data (Hoare, 2008).

In Drosophila larvae the combinatorial code involves not only lock and key-style precise coding but also patterns of stochastic, fuzzy activity. Previous data showing heterogeneous responses in mouse OSNs, in Drosophila and in Anopheles gambiae mosquitoes, can be explained by the current findings, suggesting that this may be a general feature of the peripheral olfactory code. In organisms in which a particular OSN class consists of more than one cell, it is speculated that fuzzy coding would lead to only a subset of that class responding to a given odor; as a result the glomerulus to which they projected might show lower levels of excitation than in response to precise coding. In Drosophila adults, the antennal lobe carries out nonlinear transformation of OSN inputs, amplifying weak but not strong signals, introducing noise by excitatory lateral neurons, carrying out gain control by lateral presynaptic inhibition, and detecting synchrony in patterns of projection neuron output. These processes may be necessary at least partly because there is variability in the responses of classes of OSN; in the larva, in which the olfactory system is reduced to a single pair of each type of OSN, it is more difficult to see how animal copes with the unreliability of aspects of the peripheral signal. More extensive research on a wider range of organisms will be necessary to subject this radical hypothesis to rigorous testing (Hoare, 2008).

Chemotaxis behavior mediated by single larval olfactory neurons in Drosophila

Odorant receptors (ORs) are thought to act in a combinatorial fashion, in which odor identity is encoded by the activation of a subset of ORs and the olfactory sensory neurons (OSNs) that express them. The extent to which a single OR contributes to chemotaxis behavior is not known. This question was investigated in Drosophila larvae, which represent a powerful genetic system to analyze the contribution of individual OSNs to odor coding. Twenty-five larval OR genes expressed in 21 OSNs were identified and genetic tools were generate that allow engineering of larvae missing a single OSN or having only a single or a pair of functional OSNs. Ablation of single OSNs disrupts chemotaxis behavior to a small subset of the odors tested. Larvae with only a single functional OSN are able to chemotax robustly, demonstrating that chemotaxis is possible in the absence of the remaining elements of the combinatorial code. Behavioral evidence is provided that an OSN not sufficient to support chemotaxis behavior alone can act in a combinatorial fashion to enhance chemotaxis along with a second OSN. It is concluded that there is extensive functional redundancy in the olfactory system, such that a given OSN is necessary and sufficient for the perception of only a subset of odors. This study is the first behavioral demonstration that formation of olfactory percepts involves the combinatorial integration of information transmitted by multiple ORs (Fishilevich, 2005).

The 'nose' of the Drosophila larva resides in a pair of dorsal organs at the anterior tip of the animal, each containing 21 OSNs. Previous studies showed that up to 23 of the 61 Drosophila ORs are expressed in larvae by PCR and transgenic analysis. RNA in situ hybridization was performed to provide direct evidence that OR genes are expressed in larval OSNs. Or83b, which is necessary for the proper localization and function of conventional ORs, is broadly expressed throughout the dorsal-organ ganglion. Twenty-four of the 30 ORs tested in this study are expressed in a single larval neuron in the dorsal organ. The expression of Or10a, Or43b, or Or49a mRNA or OR43b protein was not detected, although RT-PCR analysis detects these transcripts in larvae. Or92a and Or98b are also not detected by RNA in situ hybridization. Most larval OSNs express a single OR along with Or83b, but two OSNs coexpress a pair of ORs along with Or83b: Or33b/Or47a and Or94a/Or94b. Such OR coexpression has also been documented in the adult olfactory system (Fishilevich, 2005).

In parallel with the RNA in situ hybridization analysis, a collection of 42 different OR-Gal4 transgenes were examined that drive the expression of Gal4 under the control of OR promoter elements. To visualize gene expression in the dorsal organ, individual OR-Gal4 lines were crossed to UAS-GFP, encoding cytoplasmic green fluorescent protein (GFP) and the olfactory-neuron marker Or83b-Myc. Or83b-Gal4 labels all 21 larval OSNs. Per dorsal organ, 19 of the remaining 41 OR-Gal4 transgenes label a single larval OSN that is also positive for Or83b-Myc. Although Or49a mRNA was not detected in larvae, Or49a-Gal4 labels one dorsal-organ OSN along with a single terminal-organ gustatory neuron. Gustatory receptor (GR) genes are expressed in both olfactory and gustatory organs of the adult fly. GR-Gal4 transgenes are expressed only in the gustatory terminal organ or in nonolfactory dorsal-organ neurons that do not express Or83b-Myc. A total of 25 Drosophila ORs expressed in the larval dorsal organ were identified and direct evidence is provided that 24 of these OR mRNAs are expressed in situ. Of these, 14 are expressed only at the larval stage, whereas 11 are utilized by both larval and adult olfactory systems (Fishilevich, 2005).

Larval OSNs project axons to the larval antennal lobe of the brain. Patterns of axonal projections to the larval antennal lobe were examined in larvae carrying each of 20 larval OR-Gal4 transgenes along with UAS-GFP or UAS-CD8-GFP. Each OR-Gal4 line reveals a single labeled axonal arbor that terminates in an antennal-lobe glomerulus whose position is conserved between animals (Fishilevich, 2005).

The availability of genetic tools that uniquely label 19 of the 21 larval OSNs allows manipulation of the odor code by deconstructing the peripheral olfactory input and examining effects on behavioral output. Toward this end, a chemotaxis assay was establised of sufficient sensitivity to quantify differences in odor-evoked behavior. Chemotaxis of wild-type larvae was measured in response to 53 synthetic monomolecular odorants and three natural Drosophila attractants. The assay involves single-animal analysis in which the position of individual chemotaxing larvae is tracked over the course of a 5 min experiment (Fishilevich, 2005).

This assay was used to screen larval chemotaxis to a panel of 53 synthetic odors and quantified the median distance to odor for Or83b^−/− and Or83b^+/+ larvae. Forty of the 53 odors are naturally present in fruit, and of these 40, 13 are known to elicit behavioral and electrophysiological responses in Drosophila. Anosmic Or83b^−/− larvae do not respond to any odors, but wild-type (yw) larvae respond to many odors with strong chemotaxis (Fishilevich, 2005).

It was next asked how sensitive larvae are to odors by performing chemotaxis experiments at various odor concentrations. The responses to 1-hexanol are weak and not statistically different from anosmic controls for low dilutions, whereas responses increase steeply between 0.02 μl and 0.2 μl doses and appear to reach a plateau for higher concentrations. No evidence was found that higher concentrations elicit repulsion. Response thresholds to heptanal and isoamyl acetate are one and two log orders, respectively, below that of 1-hexanol (Fishilevich, 2005).

To test whether the weak responses observed for some odors at 2 μl could be explained by high detection thresholds, seven of these odors were further tested with 20 μl. Under these conditions, 1-butanol and 2,3-butanediol elicit chemotaxis, whereas the remaining five odors do not. Thus the 2 μl stimulus dose elicits robust chemotaxis across a large group of different odors, in accord with previous behavioral studies (Fishilevich, 2005).

Upon loading of an odorant stimulus in the closed-dish assay, the spatial distribution and average airborne concentration of this odor in the dish will be partly determined by the odor's vapor pressure. Vapor pressure is thus likely to affect the behavioral response observed for a particular odorant stimulus. In addition to this factor, it is anticipated that the olfactory system of the larva may be differentially tuned to different stimuli. In the initial phases of this study, no clear correlation was found between the vapor pressure of a given odor and its corresponding behavioral efficacy. It was therefore decided to avoid any normalization of stimulus concentration and used the same quantity of odor (2 μl) for all 53 stimuli tested (Fishilevich, 2005).

Whether chemotaxis elicited by single odors is comparable to that obtained with natural stimuli was examined. Chemotaxis was measured in the same assay to banana mush, balsamic vinegar, and yeast paste at different concentrations. It was found that attraction elicited by single synthetic odors is qualitatively similar to that obtained with natural odor blends and that the same steep threshold and stable plateau properties are seen for both stimulus types (Fishilevich, 2005).

The relative contribution of any given OSN to the formation of an odor percept was examined. Diphtheria toxin (DTI), an attenuated version of the cell-autonomous protein-translation inhibitor diphtheria toxin, was used to ablate identified OSNs selectively. Most but not all larval OSNs are ablated by the expression of DTI along with GFP under control of the Or83b-Gal4 driver in all 21 larval OSNs. In Or83b-ablated animals, GFP expression is not detected, and sensory dendrites are severely atrophied but not completely absent. In Or49a-ablated animals, the Or49a-GFP marker is not visible, and expression of other ORs is not perturbed (Fishilevich, 2005).

Chemotaxis of animals with single neurons ablated (Or1a, Or42a, or Or49a) was measured with a panel of 20 odors and compared to results obtained with the Or83b-ablation. Or83b-ablated larvae fail to respond to 17/20 odors. If a single false discovery (FD) is allowed for, Or83b-ablated animals fail to respond to 19/20 odors. Or1a-ablated and Or49a-ablated animals each show reduced chemotaxis to a single different odor, (E)-2-hexenal and 1-hexanol, respectively, but show normal chemotaxis to the other 19 odors. In contrast, ablation of the Or42a OSN causes decreases in chemotaxis to four of 20 odors. If FD = 1 is allowed, Or1a-ablated animals are impaired in responses to three of 20 and Or42-ablated animals to five of 20 odors (Fishilevich, 2005).

It was next asked which OSNs are sufficient to produce chemotaxis to a given odor by constructing animals with only one or combinatorials of two functional OSNs. This was achieved by exploiting the Or83b mutation, which prevents OR trafficking to the sensory dendrite. Or83b function was restored in individual OSNs by crossing animals with specific OrX-Gal4 drivers to UAS-Or83b animals, allowing assessment of the contribution of single neurons to odor-evoked behavior in the OrX-functional progeny (Fishilevich, 2005).

Only a single OR83b-expressing neuron is seen in Or42a-functional, Or49a-functional, and Or1a-functional animals, whereas two OR83b-positive neurons are visible in Or1a-/Or42a-functional and Or1a-/Or49a-functional animals. The remaining OSNs are present but unlabeled in these animals because the Or83b mutation eliminates OR83b protein expression. No evidence was found that the glomerular map is distorted by the activation of a single OSN in a background of nonfunctional neurons as evidenced by the normal position and volume of the Or1a glomerulus in Or1a-functional and Or83b mutant larvae (Fishilevich, 2005).

These animals along with genetically matched control animals were screened for chemotaxis to 53 odors by using the same behavioral assay and nonparametric statistical analysis employed for the ablation experiments. Consistent with the strong Or42a-ablated phenotypes, Or42a-functional animals respond to 22 odors compared to 36 odors in Or83b^+/+ controls possessing 21 functional OSNs. Or42a-functional animals respond to three of four odors to which Or42a-ablated animals are anosmic. The broad behavioral response profile observed for Or42a-functional larvae is in agreement with the broad ligand specificity of this OR as defined by electrophysiological experiments (Fishilevich, 2005).

In contrast to the broad odor response profile of Or42a-functional larvae, Or1a- and Or49a-functional animals do not show significant chemotaxis to any of the 53 odors tested, consistent with the weak phenotype of ablating either the Or49a-expressing or Or1a-expressing neuron. These behavioral results are in accord with the ligand profiling of Or49a, which does not show strong electrophysiological responses to any of 27 odors tested (Fishilevich, 2005).

Although Or1a- and Or49-functional larvae do not chemotax to any odors tested, it was asked whether these neurons contribute to chemotaxis in concert with the Or42a neuron. Chemotaxis performance of larvae with two functional neurons was compared to data from animals with only a single functional neuron. Larvae with two functional neurons respond to a somewhat different subset of odors than animals having either single functional neuron alone (Fishilevich, 2005).

To examine the existence of interactions between these neurons and identify cases of combinatorial enhancement, a linear regression model was developed to compare chemotaxis data across genetically matched controls for larvae with one or two functional OSNs. The model was designed to identify potential cases where single-neuron chemotaxis behavior differs from two-neuron behavior. The linear model suggests six cases of potential positive cooperativity between Or1a and Or42a chemotaxis that merited further experimental investigation. Additional chemotaxis experiments were carried out with four odors (1-pentanol, 2-pentanol, 2-hexanol, and 3-octanone) at three concentrations. 1-pentanol shows significantly stronger chemotaxis in Or1a/Or42a-functional animals than Or42a-functional or Or1a-functional animals at all three concentrations. A qualitative view of this behavioral enhancement is seen in the sector-plot distributions comparing the anosmia of Or83b mutants to the progressive increase in chemotaxis to 1-pentanol of Or1a-functional or Or42a-functional compared to Or1a/Or42-functional. The Or1a/Or42-functional animals spend comparatively more time in the sector containing the odor than animals having either single functional neuron alone. For the other three odors, this cooperative effect is significant at a single odor concentration (Fishilevich, 2005).

This study has used behavioral analysis to measure the contribution of individual neurons to the odor code and provide a missing link between the understanding of the molecular biology of ORs, the neurophysiological properties of the olfactory network, and complex odor-evoked behaviors. The goal was to approach the question of how the combinatorial activation of ORs encodes odor stimuli and elicits olfactory behavior. The results suggest that there is a high level of redundancy in the larval olfactory system, such that ablating a single neuron has minimal effects on odor detection. Among these olfactory inputs, the Or42a neuron plays a more important role in odor detection than the Or1a or Or49a neuron. Animals engineered to have the Or42a neuron functional are able to chemotax to multiple odors. The addition of a second OSN to such animals results in enhanced chemotaxis for several odors. Whereas Or1a-functional animals show no significant responses to any odor tested, it was observed that responses of Or1a/Or42a-functional animals to four odors are enhanced relative to Or42a-functional animals. This suggests that although olfactory input contributed by the Or1a-expressing OSN is not sufficient alone to elicit robust chemotaxis, it enhances the perception of odors in conjunction with the information transmitted by the Or42a-expressing OSN. (Fishilevich, 2005).

Behavior is the ultimate output of a sensory system that integrates all aspects of external-information processing. These experiments demonstrate the feasibility and value of integrating behavioral analysis into the study of odor coding. It is proposed that the simple olfactory system of Drosophila larvae will be an invaluable model in any attempt to correlate the cellular basis of the odor code with its behaviorally relevant output (Fishilevich, 2005).

Drosophila is a holometabolous insect that undergoes dramatic changes in lifestyle from the larval to adult stage. In a sense, these animals can be considered to occupy completely separate ecological niches. Larvae maintain constant contact with food until pupation, whereas adults are flying insects that use their sense of smell to identify suitable food sources and appropriate sites for egg-laying. In essence, larvae are specialized for feeding and growth, whereas adults are devoted to breeding and dispersal. To what extent have these two life stages of the same species evolved a different chemosensory system? This study shows 14 of 25 larval OR genes are stage specific and not used again by the adult animal. All larval OSNs are histolyzed in metamorphosis and replaced in the adult by newly differentiated antennal and maxillary palp OSNs. Perhaps this developmental changeover has led to largely separate OR genes with transcriptional regulatory regions specific for either larval or adult olfactory organs. Alternatively, the segregation of larval- and adult-expressed ORs could be functional and relate to the different ecological niches that these life stages occupy: larvae may cope with much higher odor concentrations because of their direct contact with food (Fishilevich, 2005).

Odor processing occurs at various levels in the nervous system, from peripheral sensory neurons to primary processing centers, such as the olfactory bulb in vertebrates and the antennal lobe in insects, and further to higher brain centers of the olfactory cortex in vertebrates and mushroom body and lateral horn in insects. How the combinatorial code established by the ORs at the periphery is transmitted through this olfactory circuitry to produce the perception of an odor in any species is unknown. The data support the notion that peripheral sensory neurons constitute information channels that are not independent but subject to interactions in the olfactory circuit. Otherwise, one would expect that the behavioral response profile observed for the Or1a/Or42a-functional genotype be given by the union of the best performances of the single Or1a- and Or42a-functional genotypes. Where and how the information is processed remains unclear, but part of this transformation may occur in the antennal lobe (Fishilevich, 2005).

A number of conclusions about odor coding in the Drosophila larva can be drawn from this work. There appears to be no clear structural relationship between the odors that elicit chemotaxis mediated by a given OSN, as has been previously shown in an analysis of the ligand response properties of ORs in the adult fly. The Or42a-expressing neuron differs from other neurons studied here in the large number of odors that attract animals having only this neuron active. Interestingly, the behavioral response profile of the Or42a-functional genotype indicates that an OR may not need to be strongly activated by a given odor to allow for chemotaxis toward the odor source. This point is best illustrated by 3-octanol and anisole, which both elicit strong chemotaxis in Or42a-functional animals whereas they seem to induce relatively weak electrophysiological activity (Fishilevich, 2005).

Finally, the behavioral receptive field of animals having combinatorials of functional neurons cannot be predicted from a simple model where the responses of animals having either single OSN functional are added. The chemotaxis results reported in this study highlight the existence of strong nonlinearities in the processing of olfactory information in such a way that in the arithmetic of sensory coding, the whole is greater than what the parts can produce independently. Such a scheme would be consistent with the extraordinary needs of the olfactory system to detect numbers of odors that greatly exceed the number of OR genes in any given animal. The functional redundancy observed here could buffer the olfactory system against mutations and allow animals to adapt to changing or new odor environments (Fishilevich, 2005).

The genetic tools presented in this study should permit a systematic analysis of the peripheral and central components that generate an odor response in the Drosophila larva. A number of key unanswered questions remain for future studies. Electrophysiological or optical imaging tools must be used to analyze the neuronal correlates of the observed behavior. Greater understanding of the second- and third-order neurons that communicate information from the antennal lobe to eventual motor output is needed. This study has been restricted to simple chemotaxis assays, and no attempt has been made to query larvae for their powers of odor discrimination. Animals missing a single OSN may chemotax normally but experience olfactory-perception not uncovered in these chemotaxis assays. By coupling associative learning of odors in intact animals followed by generalization tests in the same animals that conditionally lack a single OSN, it should be possible to determine whether odor salience is altered in larvae missing a single OSN. Finally, it will be important to determine whether the phenomena reported in this study can be considered general olfactory-coding principles that also apply to more complex animals (Fishilevich, 2005).

REFERENCES

Couto, A., Alenius, M., and Dickson, B. J. (2005). Molecular, anatomical, and functional organization of the Drosophila olfactory system. Curr. Biol. 15: 1535-1547. 16139208

de Bruyne, M., Clyne, P. J. and Carlson, J. R. (1999). Odor coding in a model olfactory organ: the Drosophila maxillary palp. J. Neurosci. 19(11): 4520-32. PubMed Citation: 10341252

de Bruyne, M., Foster, K. and Carlson, J. R. (2001). Odor coding in the Drosophila antenna. Neuron 30: 537-552. 11395013

Fishilevich, E., Domingos, A. I., Asahina, K., Naef, F., Vosshall, L. B. and Louis, M. (2005). Chemotaxis behavior mediated by single larval olfactory neurons in Drosophila. Curr. Biol. 15(23): 2086-96. 16332533

Hekmat-Scafe, D. S., et al. (1997). Coexpression of two odorant-binding protein homologs in Drosophila: implications for olfactory coding. J. Neurosci. 17(5): 1616-24. PubMed Citation: 9030621

Hoare, D. J., McCrohan, C. R. and Cobb, M. (2008). Precise and fuzzy coding by olfactory sensory neurons. J. Neurosci. 28(39): 9710-9722. PubMed Citation: 18815257

Laissue, P. P., et al. (1999). Three-dimensional reconstruction of the antennal lobe in Drosophila melanogaster. J. Comp. Neurol. 405: 543-552. PubMed Citation: 10098944

Ray, K. and Rodrigues, V. (1995). Cellular events during development of olfactory sense organs in Drosophila melanogaster. Dev. Biol. 167: 426-438. PubMed Citation: 7875369

Yao, C. A., Ignell, R. and Carlson, J. R. (2005). Chemosensory coding by neurons in the coeloconic sensilla of the Drosophila antenna. J. Neurosci. 25(37): 8359-67. PubMed Citation: 16162917

date revised: 20 December 2009

Zygotically transcribed genes

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.