They say their study, published in the April issue of Nature Neuroscience, should raise firm doubts about the validity of "vibration theory," which states that molecules in each substance generate a specific vibration frequency that the nose can interpret as distinct smells.
The reigning theory of smell, which also is as yet unproven, is that the shape of a chemical determines how it smells - much the same way as taste works.
However, at present there is no way to look at a chemical and predict what it will smell like. This is different from other sensory stimuli that are defined by simple physical properties. Color, for example, is defined by the wavelength of light.
While experiments conducted in this study were not designed to confirm the "shape theory," the results support the theory favored by most scientists, that shape of the odor molecule is the most important determinant of its smell.
"We didn't disprove the vibration theory. We just didn't find anything to support it," says assistant professor Leslie B. Vosshall, Ph.D., head of the Laboratory of Neurogenetics and Behavior. "All of our data are consistent with the shape theory, but don't prove the shape theory."
The findings are important in the sometimes contentious field of smell research because it is the first time vibration theory has actually been put to the rigor of a controlled and double-blind human test, the Rockefeller researchers say.
Andreas Keller, Ph.D., a postdoctoral fellow in Vosshall's lab, conducted a series of experiments that the principle proponent of vibration theory, the biophysicist Luca Turin, Ph.D., said would prove that his theory is correct.
Turin himself proposed the experiments in a theoretical paper but never undertook them, Keller says. Since Turin's theory was based solely on his unverified reports about the smell of certain odorants, the scientific community rejected it as "a universal theory of smell based on one man's olfactory impressions."
Turin's theory has attracted public attention thanks to a BBC documentary about him and last year's publication of the book "The Emperor of Scent." The book's author, Chandler Burr, argues that Turin is a pioneering researcher who is being ignored by the smell research community because of his unconventional ideas.
Because Turin's theory have received so much press attention, Vosshall explains that it was time for science to step in. "Our only goal is to do what Turin said should be done, in a properly controlled fashion," she says.
"I just did the experiments that Luca Turin suggested - but never actually did," says Keller. "He predicted what the outcomes would be, but we couldn't produce them."
Smell is the last of the senses to be explained. Most researchers believe in the "lock and key" shape theory, which says the shape of a chemical (the "key") fits into odorant receptor proteins on the outside of cells ("locks") that are dedicated to the sense of smell. Activated receptors promote neuronal activity in the brain that, by a still mysterious process, leads to the perception of distinct odors. But the problem with the shape theory is that humans have only 347 different odorant receptor proteins dedicated to smell, as researchers working at Senomyx discovered in 2001. A strict lock-and-key mechanism would allow humans to smell only 347 different chemicals, called odorants, when, in fact, thousands are discernable.
So, researchers now believe that only part of the curves and angles that make up odorant chemicals need to fit into the receptor. "It is probably because the lock is a little loose that different keys can fit into the same lock," says Vosshall. Still, loose locks can't explain the phenomenon by which two chemicals, each with a unique shape, can smell the same. "There are cases that are not intuitive for the shape theory, and that is why scientists have been looking for alternative theories for a very long time," Vosshall says.
Turin, who is a physiologist by training and a recognized expert on perfumes, expanded upon a theory first offered in the 1930s that smell depends on intramolecular vibrations of the odor molecule - basically the characteristic "stretching" of its chemical bonds and not the shape of molecules. He hypothesizes that receptors in the lining of the nose function as a biological "spectroscope" to measure vibrations of a chemical odorant. According to Turin, electrons in the receptor protein can lose energy by exciting the vibrational mode of a bound odorant. This only happens for a specific energy of this vibrational mode and, therefore, a receptor is only activated by odorants with a given vibrational energy.
To test Turin's theory empirically, Keller designed a series of three experiments based on experiments that Turin had proposed to prove vibration theory. Keller recruited several dozen human volunteers to the new outpatient unit of The Rockefeller University Hospital to smell different odors presented in vials, which were coded so that he did not know what they contained. The sniffing subjects then answered a series of questions, such as whether the two odors smelled different or the same.
In the first experiment, Turin predicted that if two different chemicals (guaiacol, which smells smoky, and benzaldehyde, which smells like bitter almond) were mixed together, they would smell like vanilla, because their combined molecular vibrations would match those of vanilla. None of Keller's subjects reported that the mixture had a stronger smell of vanilla than did either of the two chemicals by themselves.
In the second experiment, Keller tested whether aldehydes composed of an even number of carbon atoms smell different from those with an odd number. Aldehydes are a family of odorants made famous by being the major components of Chanel No. 5 perfume, and Turin predicted that the vibration of odd versus even aldehydes would not be the same because the aldehyde group of even number aldehydes would have more freedom to rotate, producing different vibrational frequencies. But vials consisting of two odd or two even aldehydes were not perceived by participants as more similar than vials containing an odd and an even number aldehyde, Keller says.
Vosshall adds that, in fact, this experiment supports the shape theory "because the more different in size the aldehydes are from each other, the easier it is for the human subjects to tell them apart."
The final experiment, a test of both the shape and vibration theory, is based on Turin's proposal that two chemicals that have almost identical shapes (acetophenone and deuterated acetophenone) have markedly different molecular vibrations and therefore distinct smells. Deuterated acetophenone is acetophenone that is modified to have all its hydrogen atoms replaced by deuterium atoms. This minor chemical change has only slight effects on shape, but according to Turin has major effects on vibration. In several different tests, none of the subjects could tell the difference between the two.
"They smelled the same to the subjects, which again points to a shape theory," Vosshall says. "Does that mean that no human on Earth is able to tell the difference? No, and we weren't able to test Luca Turin. It is possible that other people can do it, but not our subjects."
Because the study was not designed by the researchers to prove either theory, but rather to put Turin's theoretical approach to the test, "this is a paper of solely negative results," Vosshall says. "It shows us that molecular vibrations alone cannot explain the perceived smell of a chemical."
The study by Keller and Vosshall shows that hypotheses, no matter how intriguing they sound, have to be tested in rigorous experiments. It doesn't tell us how the sense of smell works, yet. But Keller adds that he plans to conduct additional experiments, of his own design, to help tease out the truth behind smell.