Design principles of the MacQ sensors (top) and ASAP1 (bottom). Credit: Gong et al. and St-Pierre et al., Nature Publishing Group.

How brains process information is a daunting question to answer. Genetically encoded sensors help address this problem as activity in neurons of interest can be visualized. Calcium sensors, which measure activity indirectly, have been the tools of choice, but voltage sensors are slowly catching up in popularity. The groups of Michael Lin and Mark Schnitzer, both at Stanford University, have developed voltage sensors that report neural activity at high speed.

“Voltage sensors have long been on the wish list of genetically encoded sensors,” says Lin. The first voltage sensor was reported in 1997. Since then, two classes of sensors with different designs have emerged. In one type, two fluorescent proteins capable of fluorescence resonance energy transfer (FRET) are coupled by a voltage-sensitive domain. Voltage-dependent conformational changes lead to changes in FRET signals. The other type is based on rhodopsins such as archaerhodopsin. In this case, changes in membrane voltage result in changes of the endogenous fluorescence of these proteins. However, these sensors have generally been either too slow to record activity from fast-spiking neurons or too dim to be used in tissues.

The voltage sensor from Yiyang Gong and the Schnitzer lab is based on the latter principle, but it also incorporates a fluorescent protein as a FRET donor (Gong et al., 2014). The researchers fused a modified rhodopsin from Leptosphaeria maculans to either mOrange2 or mCitrine, thereby creating MacQ-mOrange2 and MacQ-mCitrine. This strategy combines the fast kinetic properties of the rhodopsin with the brightness conferred by the fluorescent proteins.

On the other hand, Lin's ASAP1 sensor is designed from scratch (St-Pierre et al., 2014). The researchers inserted a circularly permuted GFP variant between two transmembrane helices of a chicken voltage-sensitive domain such that the fluorescent protein is located extracellularly. Voltage-induced movements distort GFP and change its fluorescence. “ASAP1 is the first truly new design concept for voltage sensors in the last five years or so,” says Lin.

Despite the different mechanisms of these sensors, their performance is comparable. Both sensors can report spike trains at higher speeds than those achieved by previously described sensors while still displaying easily detectable fluorescence changes. Both groups have tested their sensors in brain slices. The MacQ-mCitrine sensor can even report activity from dendrites in live mice. However, Schnitzer concedes that “the detection of action potentials of this type in live mice is really quite on the edge of the capability.”

Voltage sensors have a bright future ahead. “In recent years, we have seen quite some substantial improvement in voltage sensor performance,” says Lin. But both Lin and Schnitzer are trying to improve their sensors by increasing photostability, brightness and dynamic range. “The voltage sensors in general still have some improvements needed to really enable imaging with single-cell resolution in behaving mice, in the way it has become much more commonplace nowadays for calcium imaging,” says Schnitzer.

It will be interesting to see how the performance of these new sensors compares for different applications. Schnitzer thinks that “it depends on the details of the application.... It would be reasonable for people to look at them both in the particular system of interest.”