Why the Same Molecule Can Smell Sweaty or Vanilla-Like

The same molecule can smell unpleasantly sweaty or urinous to one person, sweet to another, and barely noticeable to someone else. Androstenone is known for just this kind of striking variation between people.

A 2007 study linked part of that variation to genetic forms of OR7D4, a protein that detects odor molecules in the nose. Its central contribution was to connect two kinds of experiments: testing receptor function in cultured cells and asking people to smell the same compounds.

First, finding a responsive receptor in cells

The researchers expressed 335 putative human odorant receptors in cultured cells and tested their responses to androstenone. OR7D4 produced the strongest response.

They then examined its response range using 66 odorants and two solvents. The common form of OR7D4 responded selectively to androstenone and the structurally related androstadienone. No response to the other 64 odorants or the solvents was observed under these test conditions.

A receptor is the entry point that converts stimulation by odor molecules into cellular signals. However, some receptors may not function well in cultured cells. A lack of response in this screen therefore does not prove that a receptor has no role in the human nose. The authors also noted that contributions from other receptors remained possible.

A form with two different amino acids responds less strongly

Sequencing OR7D4 in 391 people revealed several variants that change amino acids in the protein. Two common changes were inherited together. The researchers called the common form RT and the form with changes at both positions WM.

In the cell experiments, the WM form responded much less strongly than RT. The researchers also made receptors carrying each change separately; these produced intermediate responses. Comparisons using deliberately altered sequences therefore confirmed that both positions contribute to receptor function.

Other, relatively rare forms included variants with lower responses and one with a higher response than RT. The high-response form was more abundant at the cell surface, prompting the authors to discuss possible links to receptor trafficking or stability. In contrast, the differences involving the main low-response forms could not be explained simply by a lower total amount of protein produced in the cells.

Intensity and pleasantness in 391 people, detection thresholds in 121

The human experiments analyzed 391 participants after exclusions for factors such as generally reduced olfactory function. They were aged 19–75 and included 210 women and 181 men. Participants rated the intensity and pleasantness of 66 odorants at two concentrations on seven-point scales. Sample identities were concealed, and measurements were repeated twice, at least four days apart. For between-group analyses, ratings were converted into ranks within each participant to reduce differences in how people used the scales.

In 121 of these participants, androstenone and androstadienone were also progressively diluted to measure detection thresholds, an indication of the concentration needed to detect them. Odor quality was examined in a further task in which participants selected descriptors from a list of 146 terms.

People inherit one OR7D4 form from each parent. The researchers therefore compared groups such as RT/RT, with two RT copies, and RT/WM, with one copy of each form.

The low-response group perceived weaker, less unpleasant odors

Compared with the RT/RT group, the RT/WM group rated androstenone and androstadienone as weaker and less unpleasant. The small group with two WM copies also gave both compounds lower intensity ratings than the RT/RT group.

This was not a general tendency to perceive everything as weaker. In the main comparison between RT/RT and RT/WM, there were no similarly statistically clear associations with intensity or pleasantness for the remaining 64 odorants and the solvents. The compounds that activated the receptor in cells matched those for which perception differed between people.

The threshold comparison likewise showed that the RT/WM group needed higher concentrations than the RT/RT group: the reported group thresholds were 11 times higher for androstenone and 16 times higher for androstadienone. The same reduced sensitivity was not found for three control odors. These are comparisons between the groups under the study’s measurement conditions, not fixed multipliers for every individual’s sensitivity.

Cell responses for the rare forms also showed trends in the same direction as human ratings, but participant numbers were small and the group differences in intensity or pleasantness were not significant. Even when functional differences are clear in cells, sufficient participant numbers are needed to establish perceptual differences.

The likelihood of choosing “vanilla” also differed

When selecting words to describe androstenone, the RT/WM group was less likely than the RT/RT group to choose nauseating descriptors and more likely to choose “vanilla.” The low-response form was thus associated not only with lower intensity and unpleasantness but also with how participants described the odor.

This does not classify everyone carrying WM as perceiving vanilla. It is a difference in how often the supplied descriptors were selected, with variation still present within each group.

Interestingly, a secondary finding was that the RT/WM group was also more likely to select “honey,” “sweet,” and “vanilla” when smelling vanillin itself. However, there was no similarly clear group difference in vanillin intensity or pleasantness, nor did the study show that OR7D4 responds to vanillin. Verbal descriptions of odor quality must be distinguished from direct receptor responses.

Genetic forms explain part of the variation

In this population, OR7D4 genotype was estimated to account for 19% of the variation in pleasantness and 39% of the variation in intensity for the steroid odors studied. These are statistical measures of differences in ratings across the population, not claims that 19% or 39% of an individual’s sensation is made by that gene.

Other receptors, experience, and neural processing may contribute to the remaining variation. The authors also stated that the sequence of a single receptor does not fully explain why some people cannot smell these compounds.

The experiments concerned perception while smelling molecules. They did not test sexual attraction, romantic compatibility, or pheromonal effects that reliably steer human behavior. The achievement was to establish a specific connection between receptor function and differences in the perception of particular odors.

Summary

Compared with RT/RT participants, groups carrying the low-response OR7D4 form perceived androstenone and androstadienone as weaker and less unpleasant, and required higher concentrations to detect them. The frequency of describing androstenone as “vanilla” also differed.

The major finding was the correspondence between receptor function demonstrated in cultured cells and differences in human perception. Even when people smell the same molecule, differences in receptors at the entrance to the olfactory system contribute to sensations that are not entirely shared.

The research was supported by the NIH, the Irma T. Hirschl Trust, the Marco S. Stoffel Fellowship, and other sources. Human Pheromone Sciences and International Flavors & Fragrances supplied some samples. The authors declared competing financial interests. The original paper refers to separate online details, but those details could not be verified for this article.

Reference

Keller A, Zhuang H, Chi Q, Vosshall LB, Matsunami H. Genetic variation in a human odorant receptor alters odour perception. Nature. 2007;449:468–472. doi:10.1038/nature06162