How the nose knows: Instinctive organization

Buck, who first discovered TAAR genes in 2006, has spent 25 years delving into the science of smell, launched by her discovery, together with Dr. Richard Axel of Columbia University, of the genes responsible for the majority of scent detection. Those genes comprise mammals’ largest gene family and code for proteins known as odorant receptors. The researchers found that mice have about 1,000 odorant receptors, humans between 350 and 400.

Buck and Axel found these odorant receptors inside olfactory neurons, nerve cells embedded in a small area lining the upper part of the nose. Buck and her colleagues later found evidence that each olfactory neuron produces a protein from only one odorant receptor gene — all the other hundreds of odorant receptor genes remain dormant in that cell.

When airborne scent molecules waft into the nose, those tiny chemicals are captured by copies of the single odorant receptor protein sitting on the surface of nose neurons. Each odorant slots into only specific odorant receptors, like a key fitting in a lock. When that key and lock find each other, the odorant receptor triggers signals in its neuron that travel to the brain.

But even that groundbreaking discovery didn’t explain the full panoply of the sense of smell. Buck and her colleagues later found that odorant receptors are combinatorial — each scent triggers the activation of a small and unique subset of odorant receptor proteins and the neurons that house them, making an “odorant code” in the brain. And that detailed and precise organization finally illuminated how humans and other animals — even fish — can detect so many different odors. 

“Even if each odorant is only recognized by three odorant receptors, this combinatorial strategy could generate millions of different odorant codes,” Buck said. “That was a key breakthrough.”

The large number of odorant receptors and their combinatorial use in odor detection also explains how our noses can easily distinguish between even closely related chemicals, Buck said. Odorants with scents of oranges and sweaty socks, for example, are nearly identical molecules, but we perceive them very differently.

A pungent mystery

So Buck’s discovery of the TAAR genes and proteins was a big surprise, she said. The proteins are similar to odorant receptors, although a much smaller family — mice have 14 TAARs, humans six. And a small class of olfactory neurons produce TAAR proteins instead of odorant receptors (as with odorant receptors, neurons that express TAARs seem to choose only one TAAR gene to activate).

“This was really puzzling,” Buck said. “Why would there be TAARs in addition to odorant receptors?”

Then she and her team found that some of the scent molecules that TAARs recognize come from other animals. The researchers discovered that several mouse TAARs recognize compounds in mouse urine, and that was their first hint about the proteins’ special role.

“Urine is a very rich source of social cues in the mouse,” Buck said. “It can stimulate all kinds of things, like maternal behavior, mating behavior or aggression. So we thought, ‘Aha! Maybe the TAARs are involved in sensing social cues.’”

Most responses to smells (for mice or humans) are learned behaviors, said Dr. Zhonghua Lu, a postdoctoral fellow in Buck's group and one of the study’s authors. But some are innate — like attraction to something pleasing or repulsion from something potentially dangerous. Buck and other research teams later showed that TAARs recognized chemicals that trigger innate responses, either attraction or aversion. Buck’s team also found that a human TAAR recognizes spoiled fish, a smell that’s almost universally repugnant to humans, suggesting that these proteins can drive our instinctive reactions, too.

The science of attraction and repulsion

Those behaviors and their link to TAARs are intriguing, but it’s not clear how the neurons that produce individual TAARs drive innate responses.

“If that’s the case, it must be that some neurons are preprogrammed to express TAARs and may be sending signals to different pathways in the brain,” Buck said. “There must be something special about those neurons.”

To try to get at that question, Buck’s team asked how these proteins and their associated neurons are organized in the nose — and how neurons choose among the more than 1,000 different odorant receptors or TAARs available to them in the genome.

Taking advantage of the observations that each olfactory neuron always houses one, and only one, functional scent-sensing protein (either an odorant receptor or a TAAR), the researchers looked at mice with mutations in one of two different TAAR genes.

The Buck group and others found that neurons  that chose a mutant TAAR tended to replace that nonfunctioning protein with another TAAR instead of an odorant receptor, even though there are far more odorant receptor genes than TAAR genes. And Buck's group further found two classes of TAARs — a neuron missing a functional TAAR tended to choose another gene from the same subset of TAAR genes.

These two subsets of TAAR neurons are spatially separate in the nose, Buck said. These findings suggest that there’s something unique about neurons that contain TAARs, which could be a clue to help map the cells responsible for TAAR-driven instinctive behaviors.

How the neurons choose their fate

To understand how these special classes of neurons arise, the researchers then looked at TAAR gene organization inside the nucleus, the part of the cell that contains all its genes. Cells often bundle genes together spatially in the nucleus, said Dr. Tobias Ragoczy, one of the study’s authors and a staff scientist in the fundamental research laboratory of Dr. Mark Groudine, Fred Hutch executive vice president and deputy director.

“If you keep them all together, it’s easier to regulate them as a group,” Ragoczy said.

The nucleus contains “dead zones” and “hot zones” for gene activity — physical spots in the cellular compartment where genes are more or less likely to spur production of a protein. Clustering odorant receptor and TAAR genes in such dead zones within neurons makes sense, Ragoczy said, when you consider that only one out of the more than 1,000 genes needs to be activated in any given cell.

It was already known that the genes encoding odorant receptors cluster together in large silenced regions toward the center of the nucleus, so the team expected that TAAR genes might show a similar clustering. But instead they found “a striking difference,” Ragoczy said. TAAR genes are sequestered at the edge of the nucleus instead, a region not thought to be a dead zone for gene activity in olfactory neurons (although it is in many other cell types). They also saw that when a single TAAR gene is activated in a neuron, that gene moves away from its neighbors at the periphery into a different part of the nucleus.

It’s not clear why the cells would choose two different organizational systems to silence the different olfactory receptor genes, Ragoczy said, but their findings indicate that olfactory neurons have another layer of organization than scientists had previously appreciated.

Next up, Buck and her team want to uncover how TAAR-containing neurons trigger instinctive reactions by uncovering the signaling pathways in the brain triggered when a TAAR protein recognizes a scent. They’re already looking at neural pathways that govern other scent-driven behaviors in mice, behaviors like fear, mating and changes in appetite.

“What about those neural pathways causes attraction or aversion?” Buck asked. “Are signals from those TAARs going down different railroad tracks in the brain?”

There’s little known about how the brain prompts such behavior, Buck said. And that’s her team’s next challenge.