Being able to notice an odor does not necessarily mean being able to tell two odors apart. A 2015 mouse study showed that mice lacking some olfactory sensory neurons in the nose had difficulty distinguishing mirror-image molecules even at concentrations where they could detect an odor.
This was not simply a contest to determine how dilute an odor the nose can detect. Through behavioral tests and cellular analyses, the study explored whether input from especially sensitive receptors, among the many olfactory receptors sending signals, is important for telling odors apart.
Telling mirror-image molecules apart by smell
Molecules that are mirror images but cannot be superimposed, like right and left hands, are called enantiomers. Even when they consist of the same atoms, their interactions with olfactory receptors can produce different perceptions.
The researchers used two enantiomer pairs: wine lactone and carvone. They compared ordinary mice with ΔD mice, which had been genetically modified to lack olfactory sensory neurons in the dorsal region of the nasal olfactory epithelium. The latter still retained their ventral olfactory sensory neurons.
Crucially, the researchers separated two tasks. In the detection task, mice distinguished an odor-containing sample from solvent alone. In the discrimination task, they selected the member of an enantiomer pair associated with a reward. Detecting that something smelled and identifying which odor it was were measured separately.
Measuring detection and discrimination in a Y-maze
The mice were trained to choose an odor at a junction in a Y-maze and receive water at the end of the correct arm. The sample and reward positions were switched between left and right, and airflow, sample replenishment, and the time allowed to remove the previous odor were controlled.
Training began with nine male ordinary mice and seven male ΔD mice. After excluding animals that did not adequately learn the tasks during three to four months of training, among other exclusions, six mice in each group were analyzed. The main comparisons involved 18 trials per mouse, or 108 trials for six mice. Samples were diluted in hundredfold steps, and the detection or discrimination threshold was the lowest concentration at which the correct-choice rate was statistically above the 50% chance level.
These results came from small groups of extensively trained, selected mice. The ordinary and ΔD groups also did not have fully matched genetic strain backgrounds. In addition, the extremely low concentrations in the paper are nominal concentrations of the odor-generating solutions, not direct measurements of airborne concentrations reaching the mice’s noses. Dilutions were also checked within the range measurable by an analytical instrument, but its detection limit was about 10 ppb. This should not be confused with directly verifying the amount present in the most extremely dilute samples.
Losing dorsal neurons affected the two tasks differently
Ordinary mice could detect and discriminate both wine lactone and carvone enantiomers at low concentrations. Detection sensitivity was reduced in ΔD mice, but the reduction was not the same for both members of each pair. For both pairs, detection of the plus form was affected more strongly than detection of the minus form.
An even greater change appeared in discrimination. Over a wide concentration range in which ΔD mice could detect one or both molecules, they could not choose correctly between them. Their noses had not become unable to detect odors altogether; rather, the ability to detect an odor’s presence had become separated from the ability to distinguish its identity.
For carvone, for example, the reported discrimination threshold was 10−19 by weight in solution for ordinary mice and 10−9 for ΔD mice. Because ΔD mice could discriminate at higher concentrations, the result does not mean that they were completely unable to distinguish the enantiomers under all conditions.
Discrimination of wine lactone was also markedly impaired in ΔD mice. However, the paper gives inconsistent values for the ordinary mice’s discrimination threshold: 10−17 in Table 1 and Figure 1, but 10−19 in the Results text. Rather than assigning a single precise fold change, the important point is the difference between the two tasks: discrimination was strongly impaired even while detection remained possible.
What cellular responses suggest about important inputs
The study also included experiments measuring calcium responses in olfactory sensory neurons isolated from other mice and examining where receptors were distributed in the nose.
The set of receptors responding to wine lactone was smaller than the set responding to carvone. Some neurons especially sensitive to carvone also showed characteristics of the dorsal region. These findings support the possibility that removing dorsal neurons as a group eliminated highly sensitive, selective inputs important for distinguishing odors.
For a different comparison odorant, nTMT, both groups retained high detection sensitivity. This also shows that losing dorsal neurons did not affect every odor to the same extent.
The authors’ proposed model gives signals different weights
An odor is represented by a combination of responses from multiple olfactory receptors. However, closely related molecules activate strongly overlapping sets of receptors, so simply examining a list of the receptors that respond may provide only limited clues for distinguishing them.
The authors proposed a hierarchical model in which the brain gives greater weight to signals from especially sensitive receptors that respond well to a particular molecule. In this account, strong, early-arriving inputs give priority to certain odor characteristics and suppress similar inputs arriving later, allowing discrimination even when receptor responses overlap substantially.
In ΔD mice, the proposed explanation is that these prioritized inputs are lost, leaving signals shared by the two molecules to dominate. However, the order of signal arrival illustrated in the paper was inferred from measures such as receptor response magnitude. The researchers did not directly measure response timing all the way to the brain or the proposed inhibitory circuits to demonstrate the entire model.
The authors also suggest that the asymmetry in human sensitivity to wine lactone might be explained by the presence or absence of sensitive receptors. This is a hypothesis based on comparing previously reported human data with the present mouse results, not a demonstration that the same deficit exists in human noses.
Summary
Mice lacking dorsal olfactory sensory neurons showed concentration ranges in which they could detect wine lactone or carvone but had difficulty distinguishing their enantiomers. Detecting an odor and telling odors apart can therefore require different conditions, despite seeming similar.
Combining cellular response and distribution analyses, the authors proposed that a mechanism emphasizing inputs from especially sensitive receptors helps support fine odor discrimination. The findings suggest that it matters not only how many receptor types respond, but also which inputs remain and which are lost.
The study received support from Grants-in-Aid for Scientific Research, Japan’s Ministry of Economy, Trade and Industry, and research funding from Takasago International Corporation. One coauthor was affiliated with Takasago at the time of the research. No separate competing-interests statement was found in the original article text examined.
References
Sato T, Kobayakawa R, Kobayakawa K, et al. Supersensitive detection and discrimination of enantiomers by dorsal olfactory receptors: evidence for hierarchical odour coding. Scientific Reports. 2015;5:14073. doi:10.1038/srep14073.
