USE IT OR LOSE IT:
© 2004 AAAS
In A, a 40-day-old mouse in which one nostril was cauterized at birth shows a single X-gal stained M71 glomerulus at the half bulb corresponding with the open nostril (bulbO). In B, two glomeruli appear in the half bulb corresponding to the closed nostril (bulbX) – a seeming sign of immaturity. In the graphs C and D, the critical period for glomerulus maturation appears longer for M71 than for the closely related odorant receptor M72.
A smell can conjure intimate memories and even change behaviors. Scent is primal, so closely related to territoriality, predation, food, and sex in many animals, yet to most humans it seems an accessory, like a dab of perfume. People devote much more conscious attention to vision and taste, which have three and five receptor types, respectively. The human genome, however, has 350 intact odorant receptor genes, and...
A FIRST-ORDER MAP
By happenstance, John Carlson of Yale University learned about a mutant that does not contain the gene for receptor OR22a. In this mutant, the ab3A neuron that usually expresses the OR22a receptor was "empty," expressing no receptor at all.
So, Carlson and graduate student Elissa Hallem used the empty neuron as a decoder.1 They loaded the empty neuron with 31 different ORs in turn and electrophysiologically recorded each ORs response to 55 selected odorants emitted by foods such as oranges, pineapples, and almonds. Comparing the responses to those of normal, wild-type neurons allowed the team to identify which neuron normally houses each OR.
The researchers discovered that ORs also dictate different neuronal responses. "We didn't expect that," Carlson says. They showed that one receptor has two response modes, excitatory and inhibitory. For example, OR59b is excited by the buttery 2,3-butanedione but inhibited by the lemon-scented linalool. Further, ORs respond at varied intensities to diverse odorants, providing an indicator of concentration levels. Finally, an OR displays a temporal response. "Some keep firing long after the odorant is gone," says Carlson. "Some stop firing immediately."
Carlson calls the system a "navigation tool" that guides the fly to a food source based on the signal strength. The temporal response aids memory. Longer response times remind the fly of its final goal, such as a peach, while shorter response times pick up changes along the flight path. "It's an elegant system," Carson comments. "And it's possible mice derive the same types of navigation and memory cues from their OR responses."
"Carlson's work is a large first step in developing a complete view of what's happening, but there are many reasons why it may not be completely applicable to mice and humans," says Reed.
Courtesy of Bob Crimi
A wiring diagram of the axonal projections from olfactory epithelium to olfactory bulb. Four populations of olfactory sensory neurons are shown. They differ by the odorant receptor that they express. Neuronal axons of each population coalesce into a glomerulus, where they form synapses with second-order neurons.
Mammals apparently use more sophisticated mechanisms than flies for relaying the signals from OSNs to the glomeruli. According to three studies last summer,234 neurons express odorant receptor proteins on the axons, where the ORs may influence the development of the olfactory circuitry "after genetics has done its job," says Stuart Firestein of Columbia University.
In mammals, OSN placement does not matter, but their axons must connect to precisely the same glomerulus for the odor code to work. To explain the precision targeting, researchers have assumed that the glomerulus or a preexisting point on the olfactory bulb emits a positioning cue that beckons the axons, says Peter Mombaerts of Rockefeller University, New York.
"We turned the question around," Mombaerts says of research conducted with Paul Feinstein, also at Rockefeller. "We had ... overlooked the fact that glomeruli do not appear until late in fetal development, after axons are extended to the bulb." He and Feinstein propose that axons do not converge on the glomerulus. Rather, axons coalesce into a glomerulus, creating it.
In March, Mombaerts, Reed, and colleagues found that tampering with the OR's genetic instructions in mice changed the axon's target.5 In July, Mombaerts and Feinstein showed that swapping the coding regions of two related odorant receptor genes (
These findings suggest to Mombaerts that odorant receptors in mice have a second function besides detecting odor: indirectly guiding axons to the appropriate glomeruli. ORs mediate a mutual attraction between axons with like receptor proteins, steering the axons to "self-associate and coalesce," he proposes. Then, when OSNs die, as they do frequently, the replacement neurons' axons would migrate towards similar axons and end up at the right place.
REFINING THE SYSTEM
Odorant receptors may have a third function as well: refining the axons' trajectories during postnatal development and in response to sensory input. According to a paper by Firestein, Mombaerts, and Charles Greer of Yale University, the mammalian olfactory system may have a "plastic" developmental timeline, thanks to that suspected role of odorant receptors.7
At birth, the mouse's glomeruli for two odorant receptors, M71 and M72, are "a tangle of fibers," Mombaerts says. The glomeruli develop gradually and at different rates. "Over time, the glomeruli become distinct entities, but surprisingly, they are heterogeneous. As they become larger, they also become purer, more homogenous."
This maturation depends on sensory input, the researchers found. Cauterizing one nostril closed to simulate sensory deprivation at different intervals after birth defined two sensitive periods in development for the two receptors studied: day 15 for the faster maturing M72, and day 25 for the slower M71. Without sensory input before then, the glomeruli remain immature and heterogeneous. After that time, analogous to the "critical period" in vision development, sensory deprivation has no affect.
TOWARDS A CORTICAL MAP
In essence, the olfactory receptors segregate the multiple odorants emitted by, say, a rose into discrete, spatially distinct bits of information on the glomeruli. However, Buck is finding that the olfactory cortex, which lines the sides of the brain, reintegrates that information so that one perceives the complex rose aroma, rather than its component parts. Using molecular tracers that jump synapses in genetically modified mice, Buck and colleagues find evidence that glomeruli separate and connect to different regions of the olfactory cortex. Furthermore, one olfactory cortical region has neurons projecting from different glomeruli, creating a "parallel processing" system. How the olfactory cortex sends that integrated information on to the thinking and feeling parts of the higher brain is one of Buck's continuing interests.
Most likely, ongoing analyses of the human and mice genomes will help scientists put together more pieces of the olfactory puzzle. So far, genomic studies are turning up surprises of their own, according to Barbara Trask, also at the Fred Hutchinson Cancer Research Center. Like the ORs' random scattering in the epithelium, OR genes are splashed around the chromosomes as if in an "evolutionary playground," making Trask wonder how they ever manage to regulate themselves. So for now, the elusive sense of smell continues to play a game of hide and seek with researchers.