A molecular memory keeper in the noses of mice

Catherine Dulac and Stephen Santoro Courtesy of Molecular and Cellular Biology. See Related Story: Novel Histone Variant Shaping Olfactory Neurons Published In eLife at its Official Launch

Catherine Dulac and Stephen Santoro
Courtesy of Molecular and Cellular Biology. See Related Story: Novel Histone Variant Shaping Olfactory Neurons Published In eLife at its Official Launch

Novel histone variant enforces ‘use it or lose it’ policy in olfactory neurons

by Parizad M. Bilimoria

Our experiences shape our identities. It matters if we grew up on a farm or in the city. If we’ve endured poverty or had material comforts. Even whether we’re exposed more to Disney films or horror movies, hip-hop versus classical music.

Experiences both big and small personalize our nervous systems beyond what’s written in our DNA, and can lead to lasting physical changes in the brain. London cab drivers, for instance, after years of studying the intricate layout of streets in their city, develop larger hippocampi—the seahorse-shaped inner areas of our temporal lobes, important for memory and spatial navigation.

But exactly how our nervous systems record and respond to our complex environmental experiences is still one of the biggest questions in neuroscience. And while we have some knowledge of how the brain structurally remodels itself in response to sensory experience, how that experience impacts the sensory organs themselves remains unknown.

Recently new insights at the molecular level have been gained in the mouse olfactory system, thanks to research conducted in the laboratory of Catherine Dulac, professor in the Department of Molecular and Cellular Biology and Conte Center at Harvard, and an investigator of the Howard Hughes Medical Institute.

Mice have in the back of their noses a DNA-associated protein called histone H2BE which acts as a molecular memory keeper for olfactory experience, finds postdoctoral fellow Stephen Santoro. Unique to chemosensory olfactory neurons, H2BE seems to tune the lifespan of these cells based on the frequency with which they’re used—so ones that are activated often live longer than those seldom activated.

“The findings from this study provide insights into a new way in which the sensory systems that we use every day to observe the world are themselves altered by those observations,” says Santoro.

One quest launches another

Santoro didn’t begin his experiments thinking about the life span of neurons. He set out on one quest, and along the way, as often happens in science, made discoveries unearthing a whole new set of questions.

When odors hit the main olfactory epithelium at the back of our noses, the chemicals within are detected by molecular sensors on the membranes of chemosensory olfactory neurons. In most mammals these sensors, known as olfactory receptors (ORs), come in hundreds of different varieties. But each olfactory neuron “expresses” or turns on the gene for only one OR type, and all the olfactory neurons expressing that type project to one specific spot, or glomerulus, in the olfactory bulb of the brain. This floor plan, with each odor having its own signature pattern of glomerular activation, allows the olfactory bulb to decode odor identity.  

DNA wrapped around histone proteins.
Image (without original text) reprinted by permission from Macmillan Publishers Ltd: NATURE. Qiu J, 441(7090):143-5, copyright 2006.

Since the whole system is built on the basis of non-overlapping OR gene expression, neuroscientists are eager to understand what controls this process. In other words, how does an olfactory neuron know which single OR gene to turn on, and how does it make sure that all the rest are turned off?

This question of OR choice intrigued Santoro and Dulac. They wondered whether a chromatin protein only expressed by olfactory neurons controls the process. Chromatin proteins such as histones—the spools around which long threads of DNA are wrapped—are known to regulate gene expression, often by influencing chromatin compactness.

Conducting a screen for proteins unique to chemosensory olfactory neurons in mice, Santoro came across an uncharacterized variant of the histone H2B protein, dubbed H2BE. One of the most striking initial observations about this protein was the way it is expressed in the olfactory epithelium. Some neurons have tons of it, while others have barely a trace. And somehow the OR type of a neuron matters, such that neurons of the same OR type have similar H2BE levels.

To try to understand what H2BE does, the researchers designed mutant mice. One line of mutants lacked the H2BE gene, while the other had an extra, ectopic copy. This ectopic copy, unlike the naturally-existing copy, was turned on full blast in all but a few olfactory neurons.

It soon became clear that H2BE—as one might expect of a protein unique to olfactory neurons—is important for the ability to smell.  Without H2BE, mice had problems learning to distinguish odors in behavioral tests designed to reward good performance.

The mutants also revealed that H2BE regulates the expression frequencies of OR genes in a complex manner. When H2BE is deleted, neurons of OR types that normally have lots of H2BE protein show an increase in abundance, while neurons of OR types that normally have very little H2BE protein are relatively reduced.

When the converse experiment is done, with H2BE being ectopically expressed at high levels in a large subset of olfactory neurons, the expression frequencies of the ORs associated with that subset are dramatically reduced compared to those few neurons of OR types that escaped the ectopic H2BE.

“Seeing the shifts in OR expression in the mutant mice was incredibly exciting, especially after we realized that the changes depended on the normal level of H2BE associated with each OR,” says Santoro. “At the same time, the results posed more new questions than answers. Why, for example, would deletion of H2BE from neurons in which it is highly expressed cause those neurons to become more abundant?”

He and Dulac wondered, did the changes in OR expression frequencies in mutant mice mean that H2BE orchestrates the process of OR choice? Maybe if there’s too much or too little H2BE, neurons don’t know what OR type to become.

But one observation didn’t add up: Age was critical for the changes in OR expression frequencies seen in mutants. Differences were much more pronounced at six months of age than at five weeks, even though the olfactory epithelium at both ages is mature, full of neurons that have decided which OR gene to turn on.

Digging deeper, Santoro performed time course experiments comparing the expression of H2BE to that of various OR genes and markers of neuron maturation in embryos, juvenile, and adult mice. It turned out the H2BE protein was not expressed until after OR choice. It had to be regulating OR frequencies in some other way.

Olfactory neurons are one of the few varieties of neurons known to be continually replaced throughout an organism’s life. So the late-onset changes in OR frequencies could be due to changes in the longevity of neurons.

Supporting this, Santoro and Dulac found that olfactory neurons in mice without H2BE had an elevated survival rate, while those of mice with too much H2BE experienced more cell death. In fact, it seemed the lifespan of olfactory neurons might generally be correlated with their H2BE levels, such that even in normal mice, neurons with OR types linked to high H2BE levels die sooner than the rest.

But why? The new mystery was what leads to the different H2BE levels to begin with.

“While setting out to demystify the process of OR choice, we became mystified by a completely unexpected process, which was in some ways even more exciting,” says Santoro.

A mark of experience

Changes in the landscape of the nervous system—especially those occurring long after the system has developed—can be signs of adaptation to the environment. So Santoro and Dulac wondered whether the ever-changing composition of OR types in the olfactory epithelium might have something to do with olfactory experience.

To investigate, they surgically shut nostrils on one side of the nose in juvenile mice for ten days and then compared H2BE levels in the olfactory epithelium of both sides. The odor-deprived side had much more H2BE. To do the reverse experiment, looking at effects of sensory stimulation, mice were exposed for three weeks to odors with known OR targets. The neurons activated by these odors exhibited a clear decrease in H2BE levels.

It seems olfactory neurons follow a ‘use it or lose it’ policy, enforced by H2BE. Through its shifting levels this protein acts as a molecular memory keeper, recording in chromatin the smelling experiences of the mice. Neurons whose ORs are not activated much die off quickly, while those whose ORs are activated constantly—presumably more valuable for responding to the environment—live longer.

“This work took a completely unexpected turn when we realized that our main hypothesis—a role of H2BE in olfactory receptor choice—was likely to be wrong,” says Dulac. “What we discovered instead is a mechanism that had never been described before, by which a chromatin protein acts a sensor of neuronal activity, and in turn, limits or lengthens the life span of the cell.”

The researchers are still exploring how H2BE controls neuronal lifespan, but believe it may be tied to small yet significant differences between H2BE (unique to olfactory neurons) and the standard H2B molecule (found in many cell types, including olfactory neurons). While largely similar in their protein sequences, the two histone variants appear to have some differences in the chemical groups that decorate their surfaces, known as post-translational modifications (PTMs).

Because these PTMs are recognized by other important nuclear proteins that help turn genes on or off, their presence or absence can have widespread effects on a cell’s health. Santoro and Dulac hypothesize that the replacement of H2B molecules with H2BE molecules in neurons of unused OR types might dial down their overall gene activity and hasten their death.

Humans appear to have a gene similar to the mouse H2BE gene. But its expression patterns have yet to be studied. Another major question for future studies is whether histone proteins in the brain do anything similar to what H2BE does in the nose. While the supply of most neuronal types in the mammalian brain cannot be replenished, there are some types in the olfactory bulb and hippocampus that continue to be born throughout adulthood.

“We discovered a new form of activity-dependent process in the nervous system, more specifically in the olfactory system. Now the thousand dollar bet is whether or not such an elaborate mechanism is olfactory specific, or whether it exists in other parts of the brain,” Dulac notes.

She has previously written about the importance of chromatin remodeling proteins such as histones in brain function, both in normal and disease conditions. One classic example of a brain disorder where chromatin structure is believed to be disrupted, studied by another group in the Conte Center, is Rett syndrome—a rare but often severe genetic disorder marked by developmental regression in expressive language and motor skills.

“There is growing interest in the role of chromatin in neurodevelopment, learning, and memory,” says Santoro. “This is partly due to the growing number of neurodevelopmental diseases known to be associated with defective chromatin-associated molecules—many of which have largely uncharacterized functions.”

“Neuronal plasticity is an essential feature for the function and development of the brain, and any impairment is likely to be associated with a mental disorder,” says Dulac. “Finding a new mechanism, like our discovery of H2BE function, is a key mission of the Conte Center, and I am eager to pursue this line of analysis throughout the nervous system.”

Special thanks to Luke Bogart of the Conte Editorial Team for reading and commenting on this article prior to publication.

>>> Download as a PDF file