Howard Hughes Medical Institute researchers have designed a new type of molecular switch that can turn neurons on or off. Using the switches, the scientists turned down the appetites of ravenous mice, suggesting that these tools might help researchers better understand the connection between behavior and neural circuitry.

Every time a neuron in your brain fires, ion channels are at work. These protein passageways allow ions such as calcium and sodium to enter or leave a cell. Researchers, including Scott Sternson at HHMI’s Janelia Farm Research Campus, have big plans for ion channels. We hope to take charge of the molecules and use them as “remote control over neuron activity,” says Sternson, a chemist-turned-neuroscientist who has had this project in mind since he joined Janelia Farm in 2006. By using the ion channels to activate or suppress certain cells in an animal’s brain, he explains, scientists might be able to discover how small groups of neurons determine what actions an animal performs.

“This gives us new tools to be able to control electrical activity in a wide range of cells.”
Scott M. Sternson

Researchers have made some progress toward this goal. For example, HHMI early career scientist Karl Deisseroth of Stanford University and colleagues have isolated an ion channel called channelrhodopsin, which some algae use to sense light, and installed it in the brains of mice. Deisseroth and others have used the tool to control the activity of specific neurons and alter the animals’ behavior. By delivering light through fiber-optic filaments to sleep-controlling cells in the brain, for example, the researchers can activate the channel and awaken dozing mice.

Now, Sternson and colleagues have gone a step further by creating a toolbox of custom ion channels and a matching set of small molecules, or ligands, to activate them. “This gives us new tools to be able to control electrical activity in a wide range of cells,” says Sternson.

In the September 2, 2011, issue of the journal Science, the researchers revealed that they had built hybrid ion channels that carry part of a receptor for the neurotransmitter acetylcholine. When this portion of the channel is stimulated, it triggers the other part of the channel to open and allow a particular type of ion, such as sodium, calcium, or chloride, to cross. Depending on the ion that passes, neurons are either activated or inhibited. However, due to acetylcholine’s natural activity in the nervous system, researchers cannot use it to manipulate specific neurons without interfering with nerve function more broadly, Sternson says. So the team modified their ion channels to respond instead to chemical relatives of a drug called PNU-282987.0

To create their designs for new channels, Sternson and colleagues started with a computer model, teaming up with protein engineer Loren L. Looger, who’s also a group leader at Janelia Farm. What researchers don’t have, Looger says, is the acetylcholine receptor’s crystal structure, a detailed three-dimensional representation of the molecule that would allow scientists to determine how it binds to a particular ligand and adjust the architecture of both accordingly. So the researchers used a substitute, modeling part of the receptor based on a similar protein in snails whose crystal structure has been worked out.

With this virtual version, the researchers could infer how tweaking the receptor’s structure would affect its ability to bind to variants of PNU-282987. “We could predict that certain amino acids [in the receptor] would interact with the drug and what would happen if you change them,” says Looger.

The researchers then put the channel designs they developed with the computer model through their paces in the lab. They tested which of 43 channels they had built were activated by 71 chemical relatives of PNU-282987 that they synthesized in their lab. Sternson and colleagues found a bonanza: three classes of modified channels that could be activated selectively by complementary ligands in the presence of the native channel. They also showed that the channels could be activated selectively relative to each other, a critical feature for their use to independently control separate neuron populations in the same organism.

They used the ion channel-ligand combinations to manipulate neuron activity in brain tissue from mice. They found that when the relevant molecule was applied, some of their channels would turn neurons on, others would turn neurons off, and still others would allow the transit of calcium, an important ion for neural functions.

To show that the channels and the small molecules worked in living animals, the researchers used mice they had engineered to carry channelrhodopsin in some cells of the hypothalamus, a structure deep in the brain that controls basic functions such as appetite and body temperature. Light activates appetite-controlling neurons in these mice, spurring them to eat ravenously. The researchers modified the mice to also carry one of the hybrid channels for neuronal inhibition in the hypothalamus. An injection of the ligand that opens the channel prevented the animals from over-eating, the researchers found, indicating that their designer channels can shape behavior.

Sternson says that his group and other scientists have already started to apply these tools to explore questions such as how the brain regulates appetite. And he expects that this is just the beginning. “We can generate a diversity of ion channels by this approach,” he says.