This paper reflects the research and thoughts of a student at the time the paper was written for a course at Bryn Mawr College. Like other materials on Serendip, it is not intended to be "authoritative" but rather to help others further develop their own explorations. Web links were active as of the time the paper was posted but are not updated.
Contribute Thoughts | Search Serendip for Other Papers | Serendip Home Page
I began the research for this paper with several questions about the sense of smell. How are smells identified? How and why are they remembered so vividly, and why are they so emotionally charged? What does it mean that cells in the olfactory system are the only neurons to regenerate - what is retained and what is lost in this process? And what does it mean that o lfactory neurons are the only sensory neurons to synapse directly in the brain?
What I know about my own sense of smell is scant and sketchy. It is clear that smell must be functionally very different from sight or sound, because it is bas ed on the recognition of actual objects, of molecules in the gaseous phase. Unlike differences in light or sound, which can both be plotted as a range on the continuum of a single feature (wavelength for color and frequency for pitch), the range of odo rs we perceive cannot be plotted as a wave function. The movement of odorous molecules depends on diffusion, and the gathering and funneling of air that my own nostrils do. Ones nose is involved in a sort of constant sampling of a randomly selected pop ulation of airborne particles, and with every breath it performs a battery of tests for the presence of the molecules in its repertoire. But what is the nature of this testing? And how is the leap made from this molecular interaction to the identifica tion of smells, and from there to response?
At the level of receptors, the perception of color depends on only three types of cells, and from the ratio of activity of these three types of photoreceptors, the brain infers color. Is there a n analogous system at work in the perception of odor? An odiferously active molecule wafts into the nose and into the proximity of the nasal epithelium, the bed of tissue at the top of the nose where all of the chemoreceptive neurons are clustered. Th e molecule, by lovely blind chance, bumps into a receptor protein, with whom it does a little tango. Said tango leaves our protein somewhat bent out of shape, and the permeability of the neuron is altered: it fires. Several questions arise form this e ncounter. The first is, what is the nature of the receptor? Will it promiscuously respond to a wide range of molecules, or is it specific? This question was answered in part by Linda Buck and Richard Axel of the Howard Hughes Medical Institute, using ge netic analysis and statistics. They found that there are between 500 and 1000 genes (in both mice and men) which code for unique receptor proteins and which are only expressed in nasal epithelial cells (4). They also found that "each receptor gene is expressed in only ~0.1% of olfactory sensory neurons, suggesting that each sensory neuron may express only a single type of odorant receptor" (Buck, 4). This is a remarkably different model of perception than that for color: their observations point to a specific, one-to-one relationship between what they call "odorants" and the neurons which respond to them. Linda Bucks lab found that each patch of nasal epithelium is divided into four broad zones, and that each gene is expressed in only one zone, but that expression is random within the zone (4). For our molecule and protein, then, this means that their dance is in fact specific: odorous molecule A will only interact with chemoreceptor protein A, which is embedded in nasal epithelial cell type A, which is randomly distributed across one of four zones in the tissue.
The second question arising from this encounter is, where does the firing neuron go? The axon of every receptor cell in the nasal epithelium synapses directly with c ells of the olfactory bulb (1). The synapse between the axon of the receptor cell and the dendrites of the brain cell take place in glomeruli; in a mouse, each olfactory bulb has about 2000 glomeruli (4). Using mRNA for the receptor protein to identi fy the different receptor neurons at the synapse, Buck found that each type of sensory neuron connected to only 3-5 of the 2000 glomeruli (4). Furthermore, each type of neuron converges to different sets of glomeruli, and the location of any given set i s mirrored in the other olfactory bulb, as well as in other animals of the same species (4). The neurons on the surface are arranged randomly within broad zones; their axons, carrying the message of activation, are sorted out and grouped together. Th ese glomeruli are like the lights on an operators switchboard: each one corresponds to a particular type of neuron, which in turn corresponds to a particular odorous molecule. Theoretically, discrimination between smells at this point is merely a matter of identifying which "switchboard lights" are on. Well, to a certain degree...
Humans are said to be able to discriminate among ~10,000 smells (3). If there are 1000 types of receptor neurons each expressing one type of receptor protein, there is a power of ten missing in the process of odor identification. However, there can be several components to the perception of a single odor. I did a tiny experiment with myself to see if I could identify four different kinds of apples by scent alone, and, as I expected, I could. It is highly unlikely that I have a specific protein receptor and a highly localized region of my olfactory bulb set aside just for Granny Smiths, and another for Macintoshes. Rather, distinguishing between them is p robably a matter of identifying the varying concentrations of all of the component molecules that add up to apple scent. For example, Grannies could be identified by the detection of higher concentrations of certain molecules corresponding to "tart" a nd "sweet", while Macs could be identified with lower concentrations of those molecules along with higher concentrations of some other third, fourth or fifth odorant. In this way, odor identification might still be somewhat like that of color perception , in that it deals in ratios and relationships among receptor activity levels.
A second experiment, culled from experience in the organic chemistry lab, suggests that there may be other paths to odor recognition. In this lab, we looked at t he molecular differences between vanillin, the molecule extracted from vanilla beans which is responsible for their smell, and vanillyl alcohol, which is the molecule used in "artificial" vanilla flavoring. The molecules are nearly identical in struct ure. Both consist of an aromatic ring with three substituents, and the vanillyl alcohol differs from the vanillin only in that one of the substituents has an extra CH2 group (it is an ethyl group instead of a methyl group). (2) Both molecules smelled r emarkably similar, but not the same. This is where the one odorant - one neuron type - one "switchboard light" picture breaks down. If the vanillyl alcohol fits into the same receptor as the vanillin, causing that one type of neuron to fire, my sensa tion should be of vanilla: indistinguishable from the real thing. (The interaction of the protein with the ligand is not qualitative, it is either / or; if the resulting change in permeability is great enough to cause an action potential, that action p otential should have the same result every time.) At the same time, it is hard to believe that there is a separate gene in the genome to code for a different protein receptor to be expressed in a different type of neuron, just for vanillyl alcohol. A nd if such a receptor existed, it does not follow that it would correspond with the sensation of something similar but not equal to vanillin. This apparent conflict is addressed only briefly, by Buck; I found no other website or source of information on this topic. She suggests that the range of identifiable odors may be increased by receptors which interact with certain molecular features (epitopes) rather than the odorous molecule as a whole (4). In the above example, the ability to distinguish between the similar molecules would arise from the interaction of receptors for the two different substituents: perhaps there are different receptors for methyl and ethyl substituents on an aromatic ring. Thus, the ability to discriminate between simi lar but not identical odors would again rely on the interpretation of ratios and relationships across a range of neural responses.
There is very little information available about what happens to the information that is sorted out and spatially organized in the olfactory bulb. What parts of the brain are notified when one of those light bulbs flash? How is that information sorted and organized and coded? How does it eventually sift down through all of the possible options to eventually be identified, responded to, and stored? These questions await further research. The regeneration of these neurons is an active research area now for cellular and molecular biologists: they are the only neural cells of the adult human which undergo full regeneration. What are the conditions which allow / induce the regeneration to happen? Can those conditions be duplicated in the treatment of damaged neurons? The question of why they do it is really just speculation, but it occurs to me that any c ells which are constantly interacting with and taking in molecules from the outside world are susceptible to all kind of damage and contamination. Our skin cells regenerate on approximately the same schedule as that of the olfactory epithelium: every 60 days or so (3). Since the nervous system and the skin arise from the same tissue (ectoderm) developmentally, perhaps they share some programmatic coping mechanisms for dealing with the wear and tear of being exterior.