based upon the assumption that these receptors will 
share considerable homology to the family of rat ol- 
factory receptors. The deduced amino acid se- 
quences of six catfish olfactory cDNAs again define a 
family of proteins that share sequence and structural 
properties with the seven-transmembrane-domain 
superfamily of neurotransmitter and hormone re- 
ceptors. Some sequence motifs are conserved in the 
olfactory receptors identified in both fish and mam- 
mals, whereas other motifs appear specific for fish 
receptors. Overall, the individual fish receptors ex- 
hibit 40-60% amino acid homology to one another 
and 30-40% homology with representative rat 
sequences. 
If the odorant receptor repertoire reflects the 
range of detectable odors, fish might be expected to 
possess a far smaller number of odorant receptors 
than mammals. Extensive PGR (polymerase chain re- 
action) and cloning experiments identify only 20 
different receptor genes in the catfish, whereas simi- 
lar experiments in mammals identify several 
hundred genes. Southern blotting of DNA from a sin- 
gle catfish with three divergent cDNA probes con- 
firms this suggestion. Thus the number of olfactory 
receptor cDNA genes present within the catfish ge- 
nome is significantly smaller than that observed in 
mammalian DNA, as a consequence of the smaller 
size and smaller number of subfamilies. The concor- 
dance between the range of detectable odors in dif- 
ferent vertebrates and the size of the receptor reper- 
toire is consistent with a mechanism of odor 
recognition in which individual odorant receptors 
associate with only one or a few odorants. 
Spatial Segregation and the Processing 
of Olfactory Information 
Although these studies suggest a mechanism by 
which odors are recognized, discrimination among 
the vast array of odors requires that the brain discern 
which of the numerous receptors have been acti- 
vated. By analogy with other sensory systems, it has 
been suggested that this is accomplished by the 
identification of which neurons in the olfactory epi- 
thelium are activated by different odorants. Such 
identification could result from spatial segregation 
of functional classes of neurons in the epithelium 
and from the spatially defined projections of sensory 
neurons to the olfactory bulb. In the simplest form 
of this model, each olfactory neuron would express 
only one type of receptor, and each cell bearing the 
same receptor would project to the same glomeru- 
lus in the olfactory bulb. The numerical simplicity 
of the repertoire of receptors in the fish has allowed 
estimates of the diversity of receptor expression in a 
single neuron. Analysis of in situ hybridization to 
olfactory sensory epithelium with three distinct re- 
ceptor probes provides preliminary evidence that 
individual neurons are likely to express one or a 
small number of receptors. 
If the identity of a given neuron is identified by 
the nature of the receptor it expresses, then the 
identity of a given odorant can be defined by the 
cells it activates. How then does the brain discern 
which neurons have been activated by a specific 
odorant? In one model, neurons expressing a given 
receptor are spatially localized within the epithe- 
lium and project to one or a few spatially segregated 
glomeruli within the olfactory bulb. In a second 
model, neurons expressing a given receptor may ex- 
hibit no spatial order and may be randomly distrib- 
uted throughout the epithelium, while their axons 
project to one or a few glomeruli with spatially de- 
fined loci within the bulb. In either instance, expo- 
sure to a given odorant would result in the stimula- 
tion of a spatially restricted set of glomeruli such 
that individual odorants would be associated with 
specific topographic patterns of activity within the 
bulb. Distinguishing among these other models 
would require an examination of the pattern of re- 
ceptor expression in the neurons of the olfactory 
epithelium and the pattern of projections these neu- 
rons exhibit in the olfactory bulb. 
Dr. Axel and his colleagues have used probes rep- 
resentative of three distinct receptor subfamilies en- 
compassing about 20 different genes and find no 
evidence for the restricted spatial localization of 
neurons expressing specific odorant receptors 
within the olfactory epithelium, with no segrega- 
tion along anterior-posterior, dorsal-ventral, or 
medial-lateral axes. Furthermore, three-dimensional 
reconstruction techniques fail to demonstrate more- 
subtle patterns of receptor expression. Neurons ex- 
pressing specific receptors therefore appear to be 
distributed randomly within the olfactory epithe- 
lium. Thus specific odors are unlikely to elicit de- 
fined spatial patterns of activity within the olfactory 
epithelium. Rather, dispersed olfactory neurons 
with common receptors are likely to converge on 
common glomeruli in the olfactory bulb. 
These data contrast with electrophysiological ex- 
periments in higher vertebrates that suggest the spa- 
tial segregation of functionally discrete classes of 
neurons in olfactory epithelium, which is main- 
tained in the projections to the bulb. The availabil- 
ity of molecular probes for mammalian receptors 
should now permit a direct analysis of the patterns 
of expression of odorant receptors in mammals and 
may provide insight into the strategies employed in 
their more complex olfactory systems. 
NEUROSCIENCE 393 
