Harvard University COVID-19 updates

Department News

A Matter of Mixture: How Competition in the Nose Helps the Brain Make ‘Scents’ of Odor Blends [Murthy Lab]

A Matter of Mixture: How Competition in the Nose Helps the Brain Make ‘Scents’ of Odor Blends [Murthy Lab]

Imagine the job of a perfumer. They must be able to identify subtle fragrant notes mixed among intricate blends of dozens of smells, that together, comprise the olfactory profile of a scent. In fact, perfumes may contain dozens of, or even more, unique odorants that all arrive in the nose within a single whiff. Fortunately for the perfumer, and you, the sense of smell (olfaction) is unique due to its large array of receptor subtypes used to sample the world. Humans have around 400 unique types of olfactory sensory neurons (OSNs), while many rodents, such as mice, have over 1,000 unique types (contrast this with the 3-4 types of receptors for sight, 6 for taste and 1 for hearing). Much like perfume, smells in the natural world are complex blends of individual chemicals that each activate distinct, but overlapping, populations of sensory neurons in the nose. However, despite the rich diversity of OSN subtypes, the overlapping nature of odor inputs could result in the entire population of sensory neurons in the nose becoming activated when encountering blends of odors, like those found in a perfume – a fundamental impediment for effective stimulus encoding.

A collaborative effort between the Vergassola Lab at the Ecole Normale Supérieure in Paris and the Murthy Lab at Harvard, recently published in Nature Communications, sought to answer the question: How does the olfactory system contend with a virtually limitless range of potential olfactory stimuli using a limited repertoire of unique receptor subtypes? Or in other words, how does the perfumer maintain their ability to detect subtle fragrances when a complicated perfume might be expected to overwhelm the receptors in their nose.

In an earlier publication (eLife, 7:e34958; 2018), the Murthy and Vergassola groups constructed a theoretical model describing how interactions between odorant molecules in odor blends arise from the biophysical properties of their interactions with, and activation of, olfactory receptors. The model predicted that odorants “compete” for the same receptors, and, in many cases, an odor that is less effective at activating the receptor occupies it, thereby reducing activation by other, more effective odors through a process called antagonism. If antagonism is pervasive, the model predicts that receptors remain unsaturated even for complex mixtures and, consequently, distinct odors are identifiable. In their latest publication, Zak and colleagues tested this prediction by measuring how olfactory sensory neurons respond to odor mixtures in living, freely-breathing mice. To do this, they imaged fluorescent proteins that sensed calcium concentration, which allowed them to visualize the activity of olfactory sensory neurons as they responded to mixtures of odors. A key technical advance of their study was the ability to directly image individual cell bodies of sensory neurons in situ within the olfactory epithelium, along with their axon terminals as readout stimulus information that is conveyed to downstream circuits in the brain.

Zak and colleagues performed two key experiments. First, they measured how olfactory sensory neurons responded to each of two odors across a wide range of concentrations. With these measurements, they could predict how sensory neurons might respond if both odors were presented together as a mixture. They then compared the mixture predictions to experimentally-derived data from real mixtures of the odors at the same concentrations. The data they obtained revealed a widespread prevalence of antagonistic interactions – that is, the mixture stimulus evoked responses that were generally weaker than predicted from the sum of the component responses. This provided a hint that more complex odors, like perfume, are not simply a sum of their parts.
In a second experiment, they mimicked natural odors by constructing mixtures that ranged in complexity, containing up to a dozen unique odors. A key finding from this work was that, when the sensory neuron mixture responses were compared to the summation of responses to the mixture component odors, the mixture-generated responses were far below their linear predictions. These experiments further demonstrated that the prevalence and magnitude of antagonistic interactions between odors is related to the complexity of an odor mixture – more complex mixtures typically led to weaker than predicted responses.

While perfumers have long had the sense that the quality of a fragrance is defined by subtle notes contained within complex blends, these experiments are the first time that such interactions between odor molecules have been measured in live animals within the context of their natural respiration dynamics. These studies also provide new evidence that these antagonistic interactions between odors provide a “normalization’’ mechanism to prevent saturation of input to the olfactory system that is independent of neural processing. Simply put, the competition between odor molecules at their receptors helps you to appreciate complex blends of odors and the intricacies contained within them.

by Joe Zak


Venki Murthy, Murthy lab

Joe Zak (l) and Venki Murthy

Joe Zak (l) and Venki Murthy