FAIR USE NOTICE: This document contains copyrighted material whose use has not been specifically authorized by the copyright owner. This material is available as part of a mission to promote critical thinking about chemistry and chemical issues in a current social and topical context. The use is patterned after CHANCE, a well established program in Statistics.We believe that this constitutes a `fair use' of the copyrighted material as provided for in section 107 of the US Copyright Law. If you wish to use this copyrighted material for purposes of your own that go beyond `fair use', you must obtain permission from the copyright owner.

Using tools of molecular biology, scientists are beginning to figure out how tbe nose smells

Smell is a primitive sense. It helps animals find food and mates and warns them of predators. In humans, scents evoke powerful thoughts and emotions. Until recently, how the nose and brain identify molect~es that pro foundry affect behavior had been little understood. Now, using tools of molecular biology, scientists are beginning to explain odor perception in terms of a map on a part of the brain called the olfactory bulb.

Last month, Richard Axel, a professor of biochemistry and molecular biophysics at Columbia University, described the olfactory sensory map at the annual meeting of the Society for Neuroscience in Washington, D.C. His research suggests that when odor molecules interact with their receptors, they activate sites in the olfactory bulb, as if turning on light bulbs. Like a code, the spatial arrangement of the lighted bulbs is unique, presenting a map that the brain interprets as a particular odor quality.

To explain the discriminatory power of the nose, organisms must have evolved ways to recognize vast arrays of molecular structures. To figure out these mechanisms, Axel's group, using rodents, focused on odor receptors and how the sensory cells project to the brain.

Mammals detect odors initially in the posterior recesses of the nose, a region called the olfactory epithelium, Axel explained. This region is lined with bipolar cells called olfactory sensory neurons. On one end of the cell is a dendrite, which receives signals from odorants. On the other end is an axon, which projects to the olfactory bulb. The cells directly link the brain to the outside world.

The olfactory bulb is the first relay station to the brain. It connects the nose to a part of the brain called the olfactory cortex, which in turn is connected to higher sensory centers in the cerebral cortex. Because the cerebral cortex controls thoughts and behavior, the main olfactory epithelium is in a sense a "cognitive nose," said Axel, leading to the conscious perception of odors.

There is also an "erotic" nose, Axel pointed out. Called the vomeronasal organ, this nose has oll:actory sensory neurons that bypass the main olfactory epithelium and project directly to a separate accessory olfactory bulb and then to the amygdala, the part of the brain that controls instinctive behaviors. It receives information about, for example, the sexual and social status of members of a species, which could lead to mating, nurturing, or aggressive behavior.

Many mammals seem to have vomeronasal organs, but relevant histological evidence for many species is lacking. Humans appear to have cells in the right location to suggest a vomeronasal organ, but whether it is functional is the subject of signif1cant controversy, said Axel.

In lower mammals, however, the main olfactory epithelium and the vomeronasal organ have quite distinct effects on sexual behavior. For example, studies done in the mid-1980s by Robert E. Johnston, a psychologist at Cornell University, showed that male hamsters whose vomeronasal organs have been removed exhibit a clrastic reduction in mating activity, sometimes ceasing to mate altogether. ()n the other hand, a male hamster whose main olfactory nerve is severed continues to mate, but its choice of a partner is affected.

Consider a male hamster with intact main olfactory epithelium that is allowed to mate "to satiety" with a female hamster. Johnston's studies showed that if the male is allowed to mate again and has a choice between the same female or a dffferent female, the male shows a strong preference for the fresh female. But if the main olfactory nerve of the hamster is cut, it will choose either female with equal frequency.

"With these anatomic and physiological distinctions between the two noses," said Axel, "we reasoned that perhaps identifying the receptors_more specifically the genes that encode the receptors_in these two noses, might allow us an understanding of how the outside olfactory world is represented in the brain."

The first step toward this understanding came five years ago, when Linda B. Buck, then a postdoctoral fellow in Axel's group, cloned the genes for odorant receptors. She found that the genes were expressed only in thc main olfactory epithelium. "No matter what sort of molecular chicanery we could bring to the problem," said Axel, "we could not detect the receptors in the vomeronasal organ." The findings suggest_quite surprisingly, he thought_that receptors in the two noses evolved independently.

With the subsequent success of another postdoctoral fellow, Catherine Dulac, in cloning genes for receptors in the vomeronasal organ, the group's efforts yielded two families of genes with properties that make them well suited to encode odor receptors, said Axel.

First, the genes encode proteins belonging to a category of receptors that traverse membranes. Receptors of this type activate signaling proteins called G proteins. When G proteins are activated, they trigger a series of events culminating in an electrical signal being transmitted along the olfactory sensory axon.

Second, one family of genes is active only in the main olfactory epithelium, and the other, only in the vomeronasal organ.

More important, said Axel, is that the genes are expressed only in specific subpopulations of neurons in the main olfactory epithelium or the vomeronasal organ.

And third, there are a lot of receptor genes_at least 1,000. Given that on average the mammalian genome comprises 50,000 to 100,000 genes, then 1 to 2% of the genes is devoted to encoding receptors that recognize odorants. This fact, said Axel, provides the answer to the question of how it is that mammals can recognize a vast array of odors.

The cloning of these families of genes was a breakthrough, but there's a caveat: The genes are likely to encode receptors, "but rather embarrassingly, we have been unable to directly demonstrate that any of the thousand genes makes a protein that can interact with an odor molecule," said Axel. When the genes are introduced into other cells, they are expressed as RNA, but the proteins seem to get stuck and are not expressed in functional form. Axel thinks a transport protein might be required.

Until the receptors are expressed and studied outside the animals, nothing can be known about how odorous molecules actually interact with their receptors. But the ability to manipulate the receptor genes has allowed research to answer one basic question: How does the brain know what the nose is smelling?

That question, said Axel, can be translated into, "How does the brain know which receptors have been activated?" To get at the answer, Axel's group first needed to address the diversity of receptor expression in individual neurons.

Neurons can express the receptor genes in either of two ways. In one model, one neuron makes all receptors. "If this is the case," Axel said, "solutions are hopeless." Every time an odorant binds to its receptor, every neuron will send a signal.

In the other model, one neuron makes only one receptor. "If this is true," he continued, "we can reduce the problem to determining which neuron is activated."

Using the polymerase chain reaction and single-cell cloning techniques, Dulac, now an assistant professor in the department of molecular and cellular biology at Harvard University, found evidence for the latter model: Only one receptor is expressed in each neuron. The problem now becomes one of spatial segregation. Axel's group considered three ways by which neurons can be organized in space.

In the first case, spatial segregation exists in the main olfactory epithelium. Neurons with the same recep tors are separated from neurons expressing other receptors. An odorant then produces localized activity in the oUactory epithelittm that is maintained in the projections to the olfactory bulb.

In the second case, no segregation exists in the epithelium. Neurons expressing identical receptors are randomly dispersed. But their projections converge to specific sites in the olfactory bulb. An odorant then elicits unique patterns of activity in the olfactory bulb.

In the third model, no segregation exists in the epithelium or the olfactory bulb. There is complete randomness.

RobertJ. Vassar, formerly a postdoctoral fellow in Axel's lab and now a research scientist at Amgen, Thousand Oaks, Calif., carried out studies to determine how the receptors are organized spatially. Buck, who had established her independent research and is now an asso ciate professor in the department of neurobiology at Harvard Medical School, performed parallel work addressing the problem. Their experiments showed that the main olfactory epithelium in rodents can be divided into four zones. However,

a given zone is likely to express more than 200 different receptors in a totally random manner. This evidence eliminated the first model: There is no segregation of neurons in the main olfactory epithelium.

To determine which of the latter two models is correct, Vassar performed experiments that Axel initially was extremely skeptical about. Given the abundance of receptor messenger RNA in the neuron cell body, Vassar reasoned that, if, for example, 10,000 axons from cells expressing a given receptor converge to a fixed point in the oUfactory bwlb, there may be enough mRNA at the axon terminus to be detected by in situ hybridization methods. Axel was so doubtful of the approach as to suggest that if Vassar did the experiment that "he should turn the mouse upside down so the RNA will be sure to fall down the axon."

Vassar performed the experiment and found that hybridization of neurons bearing a particular receptor occurs only in two glomeruli in the olfactory bulb. At Harvard, Buck's research independently reached the same conclusion. This was the first evidence that neurons bearing identical receptors project to specific glomeruli in the olfactory bulb, the beginnings of a two-dimensional map.

Although the experiments of Vassar and Buck strongly suggested a map, they did not provide visual evidence of the convergence of the axons. "They did not allow us to demonstrate how complete the convergence was," said Axel. "Nor did the approach allow us to manipulate the sy~ tem and ask whether in fact we can perturb the map in such a way as to learn something about how it works."

Recently, Axel's group completed experiments that yielded direct evidence for convergence [Cell, 87, 675 (1996)]. Using a series of genetic manipulations, Peter Mombaerts, a postdoctoral fellow in Axel's lab and now an assistant professor at Rockefeller University, New York City, generated mice in which neurons expressing a particular receptor can be stained with a blue dye_a mouse with a blue nose. Because the blue color extended the length of the axon, the convergence of axons from neurons bearing the same receptor could be seen.

The blue-nosed mouse showed that the axons of neurons expressing an olfactory receptor called P2 converged to only two of 1,800 fixed glomeruli in the olfactory bulb of the mouse. "The pattern of convergence is absolute; stray axons projecting diffusely are not observed," the team wrote. Furthermore, the positions of the glomeruli are topographically fixed in all individuals of a species. Therefore, the brain knows what the nose is smelling by interpreting the pattern of glomerular activation in the olfactory bulb.

The map poses more questions: How do the axons know where to go? There must be a tight link between the choice of a receptor and the glomerular target to which the receptor-bearing neuron projects its axons. Axel's group believes positional cues in the olfactory bulb guide the axons to the glomeruli. And perhaps the very same receptor, which at the level of the dendrite recognizes an odorant in the environment, might be detecting guidance cues on the axon end of the neuron.

To shed light on this question, Mombaerts constructed a strain of mouse in which the P2 receptor gene was deleted and replaced with a different gene. The consequence of the mutation on axon guidance was profound, said Axel. "We no longer see convergence. Rather we see what appears to be an array of wandering axons looking for their target." Indeed, the receptor is essential for the convergence of axons that establishes the map.

To determine what type of role the receptor plays_permissive or instructive_ Mombaerts designed a different mouse. Instead of deleting the P2 receptor sequences, he replaced them with sequences encoding a different receptor, M12. He chose the M12 receptor gene because neurons expressing this gene project to glomeruli far away from the P2 glomeruli.

If the receptor plays an instructive role, the sequence swap should drive the axons to new convergence sites, the M12 glo meruli. At the other extreme, if the receptor plays a permissive role, the swap should still permit the neurons to converge on the P2 glomeruli.

Mombaerts found that the axons of the receptor-swapped mouse go to new positions close to the P2 glomeruli. Thus, the receptor plays an instructive role in guiding axons, but it is not the only factor in the process.

What is the implication of this map for sensory processing?

There are about 1,000 receptors, individual neurons express a single receptor, and the axons of neurons with a given receptor converge to fixed glomeruli in the olfactory bulb. To reconcile these facts with the ability of organisms to detect far more than 1,000 discrete odors, the map must be part of some combinatorial processing: One receptor must be able to interact with several discrete odorants. Conversely, an odor molecule must be capable of interacting with multiple receptors. By inference, an individual odor will activate multiple glomeruli in the olfactory bulb.

The potential for encoding odors is enormous. Suppose an odor activates 10 receptors and 10 discrete glomeruli. Then the number of possible codes given 1,000 receptors is close to 1023, explained Axel. So in theory, organisms should be able to detect a staggering number of discrete odorants. But in fact, organisms recognize only a very small percentage of the odorants in the environment. Members of a species seem to recognize from the outside olfactory world only those odors that are important for their survival and reproduction, according to Axel.

Despite the advance in the understanding of olfactory processing that these studies have allowed, Axel said many intriguing questions remain: How is the map read? How are the bits of spatially defined electrical information in the olfactory epithelium translated into a meaningful olfactory image? How do the patterns of glomerular activity become translated to thoughts and behavior? The story of the logic of smell has just begun, and many more secrets need to be unlocked to complete it.