The new tool can help you monitor serotonin transfer with greater fidelity than current methods

Serotonin is a neurochemical that plays a crucial role in the way the brain controls our thoughts and feelings. For example, many antidepressants are designed to alter the serotonin signals sent between neurons. In the article in Cell, Researchers funded by the National Institutes of Health have described how they used advanced genetic engineering techniques to transform a bacterial protein into a new research tool that can help monitor serotonin transfer with greater reliability than current methods.

Preclinical experiments, primarily in mice, have shown that the sensor can detect subtle changes in serotonin levels in the brain in sleep, fear, and social interactions in real time, as well as test the effectiveness of new psychoactive drugs. The study was partially funded by the NIH Brain Research through the Innovative Neurotechnology Advancement Initiative (BRAIN) which aims to revolutionize our understanding of the brain in healthy and diseased conditions.

The study was led by researchers from the laboratory of Dr. Lin Tian, ​​chief researcher at the University of California, Davis School of Medicine. Current methods can only detect wide-ranging changes in serotonin signaling. In this study, researchers transformed a fly-shaped bacterial protein in the form of a nutrient trap into a highly sensitive sensor that glows fluorescently when it captures serotonin.

Previously, scientists from the laboratory of Loren L. Looger, MD, Howard Hughes Janelia Research Campus Medical Institute, Ashburn, Virginia, used traditional genetic engineering techniques to convert a bacterial protein into an acetylcholine neurotransmitter sensor. The protein, called OpuBC, usually captures the nutrient choline, which has a similar shape to acetylcholine. For this research, Tian Laboratory worked with a team of dr. Looger and the laboratory of Dr. Viviane Gradinaru, Caltech, Pasadena, California, to show that they need the extra help of artificial intelligence to completely redesign OpuBC as a serotonin catcher.

The researchers used machine learning algorithms to help the computer “come up with” 250,000 new designs. After three rounds of testing, the scientists decided on one. Initial experiments suggested that the new sensor reliably detected serotonin at various levels in the brain, while having little or no response to other neurotransmitters or drugs of a similar shape. Experiments on mouse brain sections showed that the sensor responds to serotonin signals sent between neurons at synaptic communication points. Meanwhile, experiments on cells in Petri dishes suggest that the sensor can effectively monitor changes in those signals caused by drugs, including cocaine, MDMA (also known as ecstasy), and several commonly used antidepressants.

Finally, experiments on mice have shown that the sensor can help scientists study the neurotransmission of serotonin in more natural conditions. For example, the researchers witnessed an expected increase in serotonin levels when the mice were awake and falling while the mice were asleep. They also noticed a larger drop when the mice eventually entered a deeper, REM sleep state.

Traditional methods of monitoring serotonin would miss these changes. In addition, the scientists saw serotonin levels increase differently in two separate circuits of brain fear when the bell alerted mice to foot shock. In one circle – the medial prefrontal cortex – the bell activated serotonin levels to rise rapidly and high, while in the other – the basolateral amygdala – the transmitter crawled to slightly lower levels. In the spirit of the BRAIN initiative, researchers plan to make the sensor available to other scientists. They hope this will help researchers better understand the key role of serotonin in our daily lives and in many psychiatric conditions.


NIH / National Institute of Neurological Disorders and Stroke

Journal reference:

Unger, EK, and others. (2020) Directed evolution of selective and sensitive serotonin sensor through machine learning. Cell.