Functional expression of a Drosophila odor receptor

Insect olfaction has been a field of intense interest for two reasons. First, insect olfactory systems are well suited for investigating basic principles of olfactory system function and development, which are remarkably well conserved across phylogeny (1). Second, insects cause enormous losses to world agriculture and carry some of the world's most devastating diseases. Because many insects locate their plant and human hosts via olfactory cues, understanding the molecular basis of insect olfaction may lead to new approaches to insect control.

Odorant receptors are central to an understanding of odor sensitivity and discrimination. After many years of effort, a large gene family encoding candidate odor receptors was identified in Drosophila melanogaster (2–4) by using a novel computer search algorithm (5) or other methods. This family, the Or family, fulfills many of the criteria expected of odor receptor genes. It encodes a large family of seven-transmembrane domain proteins; individual members are expressed in small subsets of olfactory receptor neurons (2–4, 6); the number and distribution of neurons expressing a particular Or gene resemble the number and distribution of neurons exhibiting a particular odor response spectrum (7, 8); a mutation that alters the expression of Or genes also alters the odor response profiles of neurons (2, 9); neurons expressing an individual gene converge on common glomeruli in the antennal lobes of the brain (6, 10). What has been missing, however, to establish definitively these genes as odor receptor genes is a direct demonstration of function. This critical advance is now provided in two complementary papers in this issue of PNAS (11, 12).

Stortkuhl and Kettler overexpressed the Or43a gene in the fly antenna and tested for an increase in odor response in vivo (11). Or43a is normally expressed in ≈15 olfactory receptor neurons (ORNs) of the antenna, but the authors were able to drive its expression in a high fraction of the ≈1,200 antennal neurons by using the GAL4/UAS system. They then found a concomitant elevation in antennal response to a subset of odors, as measured by electroantennograms (EAGs), which are extracellular recordings of the receptor potentials of populations of neurons. Stortkuhl and Kettler found that overexpression of the Or43a gene conferred increased response to cyclohexanol, cyclohexanone, benzaldehyde, and benzyl alcohol, each of which contains a six-member carbon ring with a single attached polar group. Responses to several other tested odorants, including some others containing six-member rings, were unaffected. The logic of this experiment follows that of Stuart Firestein and colleagues, who used an adenovirus vector system to overexpress a mammalian odor receptor in the rat olfactory epithelium and measured elevated physiological responses to octanal and some related odorants (13, 14).

These results provide direct functional evidence that an Or gene does, in fact, encode a bona fide odorant receptor.

In Hans Hatt's laboratory, Wetzel et al. (12) found results that nicely complement those of Stortkuhl and Kettler by expressing the Or43a receptor in a heterologous system. Or43a was expressed in Xenopus oocytes, and responses were measured by two-electrode voltage–clamp recordings. Again, cyclohexanone, cyclohexanol, benzaldehyde, and benzyl alcohol elicited responses, with cyclohexanone and cyclohexanol inducing currents at concentrations as low as 500 nM. Six structurally related odorants and two unrelated odorants had no effect.

Taken together, these results provide direct functional evidence that an Or gene does, in fact, encode a bona fide odorant receptor. Moreover, the results are interesting in a number of other respects. The Xenopus oocytes responded to odors in the absence of an insect odorant-binding protein (OBP). OBPs are another large family of divergent proteins that have been identified in the olfactory systems of a variety of insects (15), including Drosophila (16, 17). They are present at high concentrations in the aqueous lymph surrounding the dendrites of ORNs (18, 19), and a mutation of one has been found to affect odor response in Drosophila (20, 21). OBPs are widely believed to play a role in the delivery of hydrophobic odorants through the hydrophilic lymph to odor receptors. However, there is little evidence to support a specific mechanism, and alternative models, including a role in the termination of odor response, have been proposed (22). The results of Wetzel et al. (12) indicate that an odor receptor expressed in Xenopus oocytes is capable of responding to odorants with a substantial degree of sensitivity and specificity in the absence of insect OBPs. Whether the presence of OBPs affects the kinetics or other parameters of the ligand–receptor interaction in vivo remains to be seen.

In a similar vein, one member of the Or family, Or83b, was found to be expressed in all or most ORNs (6), suggesting that it might encode a heterodimerization partner of all other members of the family. However, the results of Wetzel et al. (12) indicate that Or43b can respond to odors in the absence of Or83b expression.

A related implication of the results of Stortkuhl and Kettler (11) is that an odor receptor apparently is able to function in ORNs that normally express a different receptor. Thus Or43a is evidently able to couple with the G protein and other signaling components present in at least some other neurons, as has also been found in Caenorhabditis elegans (23). Unless the expression of Or43a represses expression or function of the endogenous receptor, the results also suggest that at least some ORNs can support the function of two receptors in the same cell. Moreover, the Or43a receptor can function in a sensillum that likely contains OBPs different from those normally in proximity to it.

The degree of increase in odor response observed by Stortkuhl and Kettler (11) in the fly antenna after overexpression of Or43b is interesting. Although the number of neurons expressing Or43a is increased dramatically (by one to two orders of magnitude, apparently), the amplitude of the EAG response is increased only modestly (2-fold at high concentrations and not at all at lower concentrations). How can this discrepancy be explained? The EAG represents a local response and not the response of the entire antenna. The response to a particular odor depends on the location on the antenna from which the recording was taken and is believed to represent the summed receptor potentials of a limited number of neurons in the vicinity of the recording electrode (24). In the wild type, Or43a is expressed only at the distal edge of the antenna, and because the recordings of Stortkuhl and Kettler were evidently made in a different location, it seems likely that their recordings in wild type reflect the activity of receptors other than Or43a. By contrast, the recordings from antennae that overexpress Or43a reflect the summed activity of these other receptors and the ectopically expressed Or43a. That the response is elevated only at high odor concentrations suggests that Or43a is a low-affinity receptor for the tested odorants. ORNs in Drosophila—and presumably the receptors that they express—vary a great deal in their dose–response curves for a particular odorant, with some showing much greater sensitivities than others (7, 8). It seems likely that Or43a has a higher affinity for an odorant other than those tested to date.

The results of these papers provide a foundation for a great deal of future work, including a more detailed examination of the specificity of Or43a and other receptors. The work may also set the stage for a developmental analysis: does the functional expression of a Drosophila odor receptor have any effect on the pattern of axonal projections (25)? Finally, the results invite behavioral investigation. Does increasing the expression of Or43a have an effect on the animal's response to this receptor's ligands?

The demonstration of function for an Or gene is a major advance in the field. It may now seem, in retrospect, that the identity of the Or genes as odor receptors was already clear, given the abundant circumstantial evidence that had been collected previously, and the prior demonstrations of function for odor receptors in other species (13, 26, 27). However, this misapprehension calls to mind the “illusion of retroactive determinism” described by French philosopher Henri Bergson. Only after functional tests, such as those of Stortkuhl, Wetzel, and colleagues, can the identity of receptor genes be clearly established. Moreover, the results of these studies provide an important foundation for further exploration of olfaction in the model insect Drosophila.