Making Sense of Scents

Science News, April 10, 1999 by John Travis

Scientists begin to decipher the alphabet of odors

Linda Buck can't smell musk. This comes as a great disappointment to her because Buck loves perfumes. She recalls dabbing them--especially Chanel No. 5, Marilyn Monroe's favorite--under her nose for fun when she was a little girl. Of course, remarks Buck, placing perfume that close to the nose can overwhelm the olfactory system to the point where the aroma is no longer sensed.

She should know. A Howard Hughes Medical Institute (HHMI) investigator at Harvard Medical School in Boston, Buck is a leader in the effort to tease out how the nose works with the brain to make sense of scents.

In 1991, she and HHMI researcher Richard Axel of Columbia University thrilled the olfaction field with the long-awaited discovery of cell-surface proteins in the mammalian nose that detect odor molecules, or odorants. Mammals appear to have around 1,000 genes for these odorant receptors, the largest gene family ever found.

In two recent papers, Buck's group and one led by HHMI investigator Randall Reed of the John Hopkins Medical Institutions in Baltimore have started to tally the odorants recognized by various receptors. This endeavor is expected to eventually reveal how one molecule can carry a pleasant scent of flowers while an almost identically shaped molecule has the stench of rotting garbage.

The preliminary answer: Distinct odorants bind to different arrays of receptors, a strategy that allows people to discriminate more than 10,000 odors even though there are only 1,000 or so odorant receptors. Buck compares this to employing the alphabet's 26 letters to form an entire dictionary of words.

"By using the letters in different combinations, you can describe an almost unlimited number of things with words. That's what the olfactory system is doing," she says

The basics of how the nose detects smells have been known for some time. Odorants waft up through the two nasal cavities until they strike a region that contains approximately 50 million olfactory neurons, the cells that bear odorant receptors. These sensory cells extend long fibers, known as axons, from the nose to the olfactory bulb, the brain region that first processes olfactory information and then sends signals to other areas of the brain (SN: 8/15/98, p. 106).

The 1991 discovery of odorant receptors spurred many new investigations into the mammalian sense of smell. In the past few years, for example, Buck and her colleagues have shown that each nasal cavity has four so-called expression zones. Any receptor type appears in only one of them.

Within each zone, however, the sensory cells bearing the same receptor are strewn about randomly, perhaps to prevent a loss of smell if a small area becomes damaged. Axons from sensory cells bearing the same receptor all converge on identical targets in the olfactory bulb.

Despite such findings, this emerging picture of the sense of smell had a major hole. Scientists hadn't been able to match odorants with their receptors.

"We knew a lot about how the wiring was set up from the primary olfactory neurons to the olfactory bulb," says Reed. "But you'll never understand how we perceive, or code, for an odor, unless you know something about the selectivity of [odorant] receptors on cells."

Last year, Stuart Firestein of Columbia University and his colleagues finally linked a rat receptor to a specific odorant, octanal, which has a meaty smell (SN: 1/10/98, p. 23). Yet the technique used, which employed a genetically engineered virus that forced sensory cells in rat nasal cavities to overproduce the receptor, is too labor-intensive to apply to all 1,000 or so receptors.

"Our real goal is to associate lots of [odorants] with lots of receptors," says Reed.

He and his colleagues therefore tried another tactic, which they described in the Dec. 23, 1998 CELL. Other scientists adding unaltered receptor genes to cells had found that the large proteins rarely make it to the cell's surface and thus seldom function. Reed and his team, however, edited the DNA sequences of various genes encoding odor receptors. The new genes resulted in surface proteins that are smaller than normal odor receptors. They retained most, but not all, of the potential odorant-binding regions.

Reed and his colleagues then joined each edited receptor gene to a gene encoding rhodopsin, a protein that they hoped would help guide the altered receptors to the cell surface. Reed's group created 80 such chimeric receptor genes, inserted them into kidney cells growing in a laboratory culture, and exposed the cells to 26 different odorants.

As indicated by changes in intracellular chemistry, each of three odorants--carvone (spearmint or carroway seeds), citronellal (citrus), and limonene (lime)--activated a different chimeric receptor. The researchers also confirmed that the rat receptor studied by Firestein's group responds to the meaty smelling octanal, but they found that the mouse version of the same receptor recognized heptanal, an herbal odor, instead. The change apparently stems from a single amino acid difference between the two receptor proteins.