An expanded view on the evolution of olfaction

Evolution of olfactory perception is often thought to arise through protein changes in odorant receptors that alter the molecules that they bind. We found that adaptive evolution in olfactory perception also arises through coexpressed receptors that change in copy number and protein sequence.

Differences in sensory perception are fascinating. We catch glimpses of these differences among our friends and family in our varying capacities to detect sensory illusions, in our appreciation for a given perfume, or in our abilities to distinguish the flavor profiles of fine wines. In subtle - and sometimes abrupt – ways, our experience of the external world is unique.

While sensory differences can arise due to environmental factors, my research is focused on the genetic basis of sensory variation. I am interested in understanding how sensory differences emerge between populations and species: What are the genes and neurons underlying sensory differences, and what evolutionary forces drive this diversity? To address these questions, I use Drosophila. This includes the standard workhorse for most fly labs, D. melanogaster, but also ecologically diverse species that are closely related to it. The relatively simple brain of these insects, along with the rapidly expanding genetic toolset that can be applied to non-D. melanogaster species, have been ushering in exciting new ways to understand how sensory perception evolves. A recently published article by myself and colleagues Tom Auer, Raquel Álvarez-Ocaña, Steeve Cruchet, and Richard Benton provides one such illustration of this approach applied to olfactory perception.

The origin of this project can be distilled to two observations. The first has to do with odorant receptor gene families. I suspect anyone who has browsed a paper announcing the release of an animal genome has come across a section describing the number of odorant receptor genes. These gene families are among the most numerous in animal genomes and change rapidly in size between species by gene duplications and deletions.

The second observation concerns the exquisite cellular expression of odorant receptors. In insects, the neurons that express the receptors – olfactory sensory neurons - are found in the peripheral sensory tissues such as antennas. Despite the dozens of odorant receptor genes in Drosophila genomes, each distinct olfactory sensory neuron population will usually “choose” only one receptor to express out of the family. This is sometimes referred to as the one-receptor one-neuron pattern of expression. The olfactory sensory neurons encode the odors that they detect and, with their first synapse, pass their information to the central brain via a region called the antenna lobe.

A question that emerges from these two observations – and one that has long lingered in my mind - is: How are odorant receptor gene duplication events “coordinated” with developmental changes in the olfactory system in order to evolve new populations of olfactory sensory neurons? Influenced by the prevailing one-receptor one-neuron model, the simplified scenario that I (and others) primarily considered had two components. After gene duplication, there would need to be regulatory changes between the duplicated receptor genes that would lead to their exclusive neuron expression patterns, and there would need to be protein changes resulting in unique odor response profiles.

Realizing that advances on these questions are likely to come from studying independent odorant receptor gains and losses, we have been studying an odorant receptor subfamily named Or67a. To our knowledge, this is the most duplicated and deleted subfamily of odorant receptors among the species closely related to D. melanogaster. In breaking new ground with this system, we were met with unexpected results that have expanded my way of thinking about the molecular basis of olfactory perception evolution.  

Figure 1.  Connecting changes in odorant receptor copy number to modifications in the sensory nervous system. a, The four Drosophila species used in this study and the number of Or67a genes in their genomes. b, The initial expectation for how changes in  Or67a gene copy number would relate to the olfactory sensory neuron diversity. c, We found that three Or67a genes that have diverged in function are co-expressed in the same neuron and contribute to differences in species olfactory perception.

The change in my thinking began with our experiments using four species. Two species (D. simulans and D. mauritiana) have three Or67a genes and the other two have one Or67a gene (D. melanogaster and D. sechellia; Figure 1a). Our hypothesis was that the common ancestor had three Or67a genes, each expressed in a unique set of olfactory sensory neurons. Then, after the speciation events leading to the four species, D. melanogaster and D. sechellia lost two receptors and the populations of neurons that had previously expressed these receptors. We expected that this project would lead to new insights into the reduction of sensory neuron populations (Figure 1b). Instead, we found something very different. While we verified that the common ancestor did have three Or67a genes, the surprising finding was that they were not expressed in distinct sensory neurons but co-expressed in the same neuron population, and remain so in D. simulans (and likely D. mauritiana, Figure 1c). Moreover, we were able to demonstrate that positive selection had already diversified their odor response profiles in the common ancestor (with ongoing selection continuing to modify them in each lineage). Together, these results showed that, while the neural circuitry has remained conserved, olfactory perception differences arose due to copy number changes (both duplications and deletions) and positive selection on co-expressed odorant receptors.

Though it currently seems that most insect odorant receptors are uniquely expressed, this perspective is dominated by studies carried out in D. melanogaster. Even with this bias, the Or67a subfamily is unlikely to be unique. There are several other reports of odorant receptor co-expression (in D. melanogaster and other insects), and although analogous evolutionary analyses have not been carried out, similar processes may well be involved. After all, evolution is opportunistic; if the wait time for regulatory mutations that result in neuron-specific expression is longer than the wait time for adaptive mutations in co-expressed receptors, there is not an obvious reason to me that would preclude adaptation from taking this route in other instances.