(L to R) Erin Song, Sam Kunes, and Benjamin de Bivort

How does a sensory neuron in an animal’s snout or antenna choose one Odorant Receptor (Or) gene to express from a genomic catalog of Or genes sometimes numbering in the thousands? Most genomes have large repertoires of Or genes, but an individual neuron typically expresses just one or a few. And the neurons that express the same Odorant receptor gene are usually scattered across sensory organs, interspersed with neurons that express different Or genes; the resulting olfactory epithelia are ‘crazy quilts’ of chemical sensitivities. Moreover, these patterns are not precise but vary considerably from one individual to another. One model put forward to explain the development of these Odorant receptor patterns is that each neuron pretty much makes an autonomous and stochastic receptor choice, while precluding the activation of any others.

Developmental biology has in recent decades had great success at explaining the development of precise form and pattern. For example, secreted signaling molecules create spatial patterns by forming tissue-wide concentration gradients that trigger distinct cell fates accordingly. We like to think, in fact, that developmental mechanisms are designed to keep multi-cellular chaos to a minimum. Developmental mechanisms that, in contrast, rely on chance are somewhat novel and interesting. The olfactory system has already given us an example of choice by chance; some mouse Or genes are selected via a chance interaction of their promoter regions with distal chromosomal enhancer loci, the P and H elements. We have addressed the question of whether such stochastic mechanisms are necessary to explain the pattern of Or gene expression in Drosophila. The short answer is “probably not”. A more complete explanation can be found in our publication in Developmental Cell.

Neurons that express the same Or gene are scattered across the surface of a fly’s antenna, interspersed with neurons expressing other Or genes. Our insight into the development of this pattern began with the simple observation that soon after the onset of their differentiation, the olfactory sensory neurons, bundled into stereotyped groups called sensilla, become highly motile and move extensively around the early antenna. Clonal analysis revealed that neurons born together don’t stay together. Moreover, we identified two genes, engrailed and dachshund, well known developmental regulators, which function very early and in a precise spatial pattern as part of a combinatorial code to determine olfactory neuron identities. The sensilla, once so defined, migrate through the developing antenna to create the crazy quilt of adult chemical sensitivities. The fly’s Odorant receptor pattern thus arises by two mechanisms we already knew about – spatial patterning and cell migration.

Is this a universal mechanism for establishing the “mosaic, stars-in-the-sky” expression patterns of Odorant receptor genes? Perhaps not. We know that many mammalian Or genes have very simple promoter regions inconsistent with unique transcription factor regulatory codes, and appear to be controlled by an exclusive and chance encounter with distant chromosomal regions containing the P and H loci. For these genes, stochastic choice still makes the most sense. Nonetheless, as our work points out, it would be worth knowing where the neurons were when these decisions were made.

Read more in Developmental Cell

[February 14, 2012]