P&S Medical Review: December 1995, Vol.3, No.1
Review: Molecular Analysis of Sensory Processing in the Olfactory System

Although in primates vision is the central sensory modality, the sense of smell is crucial in the life cycles of most mammals, lower vertebrates, and insects. Many organisms use the sense of smell to mark their lairs, to avoid predation, to stabilize social organization, to find food at night, and to attract mates over large areas of forest. Even in humans, in whom the sense of smell is not necessary for survival and reproduction, phenomena related to reproductive biology, such as the synchronization of menstrual cycles and mother-infant recognition, have been shown to be dependent upon olfactory cues.1 The olfactory system of mammals provides a rich representation of the external world. Humans are thought to be able to detect and recognize more than ten thousand odors.

What is the molecular basis of odor recognition?

A fuller understanding of molecular recognition and higher sensory processing in the olfactory system was made possible by the cloning of the odorant receptor genes in 1991 by Linda Buck and Richard Axel at the College of Physicians and Surgeons (Figure 1). Using the Polymerase Chain Reaction (PCR), they isolated a novel family of receptor genes encoding seven-transmembrane domain proteins, and estimated that there are one thousand receptor genes, comprising one percent of the active genome.2

The discovery of the size of the repertoire of odorant receptor genes in mammals clarified one of the central questions of sensory processing in olfaction: Are there primary odors?

In the visual system, three pigments are used in combination to detect a large range of hues. The brain compares the inputs from the three classes of cones in the retina (each cone expressing a single pigment), and deduces the color of the subject as a particular blend of red, green, and blue. The cloning of the odorant receptor genes has shown that, in contrast to color vision, in which three receptors interact with all of the visible wavelengths of light, the discrimination of odors requires many receptors, each receptor interacting with one or a few odorants.

Figure 1. The Putative Odorant Receptors are Seven Transmembrane Domain Proteins

The generation of diversity in the olfactory system is thus achieved at the level of the gene. An even greater diversity is achieved in the immune system by the somatic rearrangement of T cell receptor and immunoglobulin genes. However, the extreme capacity for molecular recognition in the immune system does not require precise molecular discrimination by the organism. Although the immune system identifies all ligands as either self or nonself, the structures of the ligands are less relevant to the immune response. In contrast, the olfactory system of mammals is capable of discriminating among thousands of molecular structures. In addition to conscious perception of the odor, each structure may be associated with different behaviors, memories, and states of physiological arousal.

Figure 2. Three Models of Spatial Segregation Each glomerulus contains the terminals of approximately 3000 sensory axons. OE, olfactory epithelium; OB, olfactory bulb; sn, sensory neuron; g, glomerulus; mc, mitral cell; LOT, lateral olfactory tract.

The discrimination of odors requires that the nervous system determine which of the thousand receptors in the nose have been activated. If many different receptor genes are expressed in each sensory neuron, the problem of determining which receptors are activated by an odorant is very complex. However, if only one receptor gene is expressed per cell, it may be reduced to the problem of determining which cells are activated.

How many receptor genes are expressed in each sensory neuron?

Quantitative analyses from in situ hybridization indicate that approximately 0.1% of cells express a given receptor gene.3,4 If the receptor genes are equally represented in the olfactory epithelium and there are one thousand genes, these data imply that one receptor gene is expressed in each cell. Catherine Dulac and Axel have recently shown that amplification of a single-cell cDNA library with degenerate primers to conserved regions of the receptor genes gives a single PCR product, further supporting the notion of one gene per cell.5

Given that one odorant receptor gene is expressed in each sensory neuron, how are sensory neurons organized in the nose?

In other sensory systems, the spatial segregation of neurons or their projections is used to indicate the quality of the stimulus. In the auditory system, for example, the positions of hair cells along the basilar membrane are used to represent different tones. In the somatosensory system, the submodalities of pain and temperature are processed by separate pathways in the spinal cord.

Figure 3. There are Four Zones in the Olfactory Epithelium The four zones of receptor gene expression are numbered 1-4, from dorsal to ventral. The respiratory epithelium does not contain sensory neurons. On the left is a coronal view of the olfactory epithelium, and on the right is a whole mount view of the turbinates, the convoluted structures in the nose which support the epithelium. D, dorsal; V, ventral; A, anterior; P, posterior; R, respiratory epithelium.

Is spatial segregation used to encode odor quality?

Odorants interact with receptors on the cilia of sensory neurons in the olfactory epithelium (Figure 2). Each sensory neuron projects a single axon through the cribriform plate of the ethmoid bone to the olfactory bulb, where it synapses with mitral cells in a spherical structure called a glomerulus. The mitral cells in turn project to the cerebral cortex in the lateral olfactory tract.

The anatomy of the olfactory system suggests three models of segregation. In one model, expression of the receptor genes would be segregated, such that sensory neurons expressing the same receptor gene would be clustered in the olfactory epithelium. This segregation would be preserved in the projections of the neurons to the olfactory bulb (Model 1 in Figure 2). In another model, the expression of receptor genes would be random in the epithelium but the projections of neurons expressing the same odorant receptor gene would be convergent in the olfactory bulb (Model 2 in Figure 2). In the last model, both the expression of receptor genes and the projections of neurons to the bulb would be random, and the discrimination of odors would be accomplished by associative learning (Model 3 in Figure 2). The availability of the receptor genes as molecular probes for distinct sets of sensory neurons has allowed biologists to distinguish among these models.

The first molecular analysis of spatial segregation in the olfactory system was performed by John Ngai and Andrew Chess, both postdoctoral fellows in the Axel laboratory. Ngai and Chess cloned the odorant receptor genes of the catfish by homology to the mammalian genes and analyzed the pattern of gene expression in the olfactory epithelium.6,7 In situ hybridization on olfactory rosettes for receptor messenger RNA revealed a random distribution of sensory neurons in the epithelium.

Is the expression of receptor genes also random in the olfactory epithelium of mammals?

Linda Buck, now at Harvard University, and Robert Vassar, a postdoctoral fellow in Richard Axel's laboratory, have independently shown by in situ hybridization that there are four zones of receptor gene expression in the olfactory epithelium of mammals, each similar in complexity to the entire epithelium of the fish (Figure 3).3,4 Although the expression of each receptor gene is restricted to a single zone, sensory neurons expressing different receptor genes are randomly distributed in a punctate pattern within each zone. Clearly, the limited segregation in the epithelium is inadequate to explain the extreme discriminatory power of the olfactory system.

From these data Model 1 could be excluded from consideration. To distinguish between Model 2 and Model 3, biologists had to analyze the next level of the sensory system.

Previous experiments by Gordon Shepherd, John Kauer, and their colleagues at Yale University using voltage-sensitive dyes and metabolic labeling indicated that exposure to different odorants is accompanied by different patterns of activity in the olfactory bulb.8,9 Electrophysiological experiments in Kensaku Mori's laboratory at the Osaka Bioscience Institute have shown that mitral cells from adjacent glomeruli are sensitive to odorants with different molecular structures.10,11

These experiments showed that the neuropil of the olfactory bulb is segregated into functional units but did not relate its organization to the olfactory epithelium and olfactory nerve. The first anatomical evidence of segregation within the olfactory nerve was provided by Liliane Astic and Diane Saucier at the UniversitŽ Claude-Bernard in Lyon, France. Astic and Saucier visualized the sensory projections to the olfactory bulb by retrograde labeling with horseradish peroxidase. The pattern of staining they observed in the epithelium with limited injections into different parts of the bulb bears a haunting resemblance to the four expression zones of the epithelium described by in situ hybridization (Figure 3). Ventral injections into the bulb stained the ventral epithelium, and dorsal injections stained more dorsal regions.12

Although the Astic and Saucier experiments showed a zone-to-zone projection, the staining they described was also consistent with a more organized projection. Because of the punctate pattern of gene expression in the epithelium, it appeared that a technique for axon tracing with greater spatial resolution than dye injection was necessary to analyze the projections of sensory neurons expressing the same odorant receptor gene. One obvious marker for these axons would be the receptor messenger RNA.

RNA is rarely found in axons.13,14,15 However, Axel and colleagues reasoned that, since sensory neurons express high levels of odorant receptor mRNA, small amounts of RNA entering the axon might be visualized by in situ hybridization, if the projections of neurons expressing the same receptor gene were convergent.

Is there convergence of sensory axons in the olfactory bulb?

In situ hybridization on coronal sections of the rat olfactory bulb shows that sensory neurons expressing the same odorant receptor gene, and therefore sensitive to the same odorants, converge on one or a few glomeruli (Figure 4). Probes detecting 22 genes by Southern hybridization label only 19 of the 3000 glomeruli in the olfactory bulb. Moreover, the positions of labeled glomeruli are bilaterally symmetric and are conserved in all animals.16 A similar hybridization pattern has been described in the mouse by Linda Buck's laboratory at Harvard.17

These data provide physical evidence that the central representation of odors is spatially organized, such that exposure to an odorant is associated with a unique pattern of depolarization in the olfactory bulb. In the visual system, every level of sensory processing must preserve an accurate spatial representation of the world. Neural space is thus constrained to mimic physical space. In contrast, it appears that the olfactory system, having evolved to recognize smells rather than to locate their sources, uses neural space as a matrix in which to arrange the universe of odors.

How are odors analyzed in the brain?

Because most odors are complex mixtures, it is improbable that each odor will be represented by a single locus in the brain, just as it is improbable that there is a single cell in the visual cortex specializing in the detection of tigers. While there may be no tiger cells, it has been shown that there are neurons sensitive to bars of light in a particular orientation, bars moving in different directions, and more complex shapes that combine stripes and curves. The richness of pattern vision is achieved by hierarchical processing. For example, cells with simple center-surround receptive fields may synapse with each other and with higher order cells to create receptive fields in the shape of a bar. The cells with bar-shaped receptive fields may synapse with each other and higher order cells to create a striped receptive field, with alternating bars of light and dark. The olfactory system may use a similar strategy in the processing of odors.

Figure 4. Sensory Neurons Expressing the Same Receptor Gene Converge on One or a Few Glomeruli (A) Synaptic organization of the olfactory bulb. D, dorsal; V, ventral; ONL, outer nerve layer; GL, glomerular layer; EPL, external plexiform layer; MCL, mitral cell layer. (B) Dark field micrograph of a coronal section of the olfactorybulb, showing the hybridization pattern for receptor mRNA. The labeled glomeruli are in the same position in both halves of the olfactory bulb. (Reprinted with permission from Cell.)

In the visual system, as in other sensory systems, the specificity of the sensory process is determined by natural selection. In a world in which all predators are striped, organisms with visual systems specializing in the detection of moving striped forms are more likely to survive and reproduce. By the same logic, organisms that recognize burning smells are more likely to escape forest fires, and organisms that can smell poisons are more likely to avoid drinking or eating them. The discovery of the convergence of sensory axons in the olfactory bulb suggests that a given odor may be represented in the cerebral cortex as the combinatorial activation of different glomeruli. Although a locus in the brain may receive axons from the mitral cells of many different glomeruli, the number of possible combinations of glomeruli exceeds the number of neurons. Selective pressure must determine both the repertoire of odorant receptor genes expressed by an organism and the projections of glomeruli to the cerebral cortex that are preserved. Smells that have adaptive significance will be preserved in the central representation of odors, and smells that are not relevant to survival and reproduction may disappear from the evolving neural map.

Molecular techniques have allowed biologists to begin to elucidate the functional organization of the olfactory system. Many difficult questions remain. For example, how does a sensory neuron choose a single odorant receptor gene to be expressed, and how is the regulation of receptor gene expression linked to axon guidance? Are the receptor proteins used in axon guidance? How is the convergence of sensory axons maintained during the turnover of neurons in the olfactory epithelium? How is the parallel processing of odors accomplished in the brain, and how is the sense of smell related to learning and memory?

In future experiments, odorant receptor genes may be replaced by gene targeting techniques with genes encoding tracer molecules, allowing single axons to be visualized by enzymatic staining. Such mutant animals could be used to study the development of the olfactory system, relating convergence to the onset of receptor gene expression and revealing the contribution of activity-dependent processes to the refinement of the sensory projection. The recent cloning of pheromone receptor genes from the vomeronasal organ of the mouse by Dulac and Axel may permit the extension of gene targeting analysis to sexual behavior and aggression.18 In addition, the strategy of subtractive hybridization used to clone the pheromone receptor genes may be applied to the cloning of genes for axon guidance cues expressed in mitral cells. If the odorant receptor proteins are used in axon guidance, the generation of anti-receptor monoclonal antibodies will show that they are expressed in the growth cones of sensory axons. Finally, in combination with the techniques of molecular biology, the development of advanced imaging techniques, such as functional MRI (magnetic resonance imaging) and hybrid imaging systems integrating MRI and electroencephalography, may provide information about sensory processing events in the limbic system and cerebral cortex.

ACKNOWLEDGMENTS

The author is grateful to Richard Axel and Robert Vassar for critical discussions. This work is supported by a grant from the National Institutes of Health (T32 GM07367).

REFERENCES

1. Agosta, WC. Chemical communication: the language of pheromones. New York: Scientific American Library, 1992.

2. Buck L and Axel R. A novel multigene family may encode odorant receptors: a molecular basis for odor recognition. Cell 1991; 65: 175-187.

3. Ressler KJ, Sullivan SL, Buck LB. A zonal organization of odorant receptor gene expression in the olfactory epithelium. Cell 1993; 73: 597-609.

4. Vassar R, Ngai J, Axel R. Spatial segregation of odorant receptor expression in the mammalian olfactory epithelium. Cell 1993; 74: 309-318.

5. Dulac C and Axel R, unpublished data.

6. Ngai J, Dowling MM, Buck L, Axel R, Chess A. The family of genesencoding odorant receptors in the channel catfish. Cell 1993; 72: 657-666.

7. Ngai J, Chess A, Dowling MM, Necles N, Macagno ER, Axel R. Coding of olfactory information: topography of odorant receptor expression in the catfish olfactory epithelium. Cell 1993; 72: 667-680.

8. Kauer JS, Senseman DM, Cohen, LB. Odor-elicited activity monitored simultaneously from 124 regions of the salamander olfactory bulb using a voltage-sensitive dye. Brain Res. 1987; 418: 255-261.

9. Stewart WB, Kauer JS, Shepherd GM. Functional organization of rat olfactory bulb analyzed by the 2-deoxyglucose method. J. Comp. Neurol. 1979; 185: 715-734.

10. Imamura K, Mataga N, Mori K. Coding of odor molecules by mitral/tufted cells in rabbit olfactory bulb. I. Aliphatic compounds. J. Neurophysiol. 1992; 68: 1986-2002.

11. Katoh K, Koshimoto H, Tani A, Mori K. Coding of odor molecules by mitral/tufted cells in rabbit olfactory bulb. II. Aromatic compounds. Neurophysiol. 1993; 70: 2161-2175.

12. Astic L and Saucier D. Anatomical mapping of the neuroepithelial projection to the olfactory bulb in the rat. Brain Res. Bull. 1986; 16:445-454.

13. Jirkowski GF, Sanna PP, Bloom FE. mRNA coding for oxytocin is present in axons of the hypothalamo-neurohypophysial tract. Proc. Natl. Acad. Sci. USA 1990; 87: 7400-7404.

14. Brunet JF, Shapiro E, Foster SA, Kandel ER, Iino Y. Identification of a peptide specific for Aplysia sensory neurons by PCR-based differential screening. Science 1991; 252: 856-859.

15. Steward O and Banker GA. Getting the message from the gene to the synapse: sorting and intracellular transport of RNA in neurons. Trends Neurosci. 1992; 15: 180-186.

16. Vassar R, Chao SK, Sitcheran R, Nu–ez JM, Vosshall LB, Axel R. Topographic organization of sensory projections to the olfactory bulb. Cell 1994; 79: 981-991.

17. Ressler KJ, Sullivan SL, Buck LB. Information coding in the olfactory system: evidence for a stereotyped and highly organized epitope map in the olfactory bulb. Cell 1994; 79: 1245-1255.

18. Dulac C and Axel R. A novel family of genes encoding putative pheromone receptors in mammals. Cell 1995; 83: 195-206.