Neuroscientists know well how technically challenging it is to poke a pipette into a living cell to measure its electrical properties. Skilled researchers with a gift for this meticulous methodology and the patience required to perform it are highly sought after in many laboratories, but they are becoming a rare species.

Voltage-sensitive fluorescent dyes have become an alternative to electrical cell recordings using pipettes. But although dyes are fast and sensitive enough to detect single action potentials in spiking neurons, phototoxicity and challenges in delivery have prevented their widespread use.

Fast and sensitive detection of action potentials with a genetically encoded voltage indicator.

A fast and sensitive genetically encoded voltage indicator is high in the list of most-wanted tools by neuroscientists. Initial versions were substantially slower than dyes and not sensitive enough to reveal single action potentials in cells. Major efforts to develop better protein-based voltage sensors have come from Thomas Knöpfel at RIKEN, who uses fluorescence resonance energy transfer and a voltage-sensing phosphatase from a sea squirt as the basis for the design of new voltage-sensitive fluorescent proteins. These sensors have been used to detect action potentials in mammalian neurons in vivo, but their sensitivity is still not high enough to detect single action potentials from single trials (Nat. Methods 7, 643–649; 2010).

The group of Adam Cohen at Harvard University has recently used an entirely different class of proteins to develop fast and sensitive voltage indicators: microbial rhodopsins. The first of such voltage indicators, proteorhodopsin optical proton sensor (PROPS), is based on the endogenous fluorescence of a rhodopsin from marine bacteria, but its use is limited to prokaryotes (Science 333, 345–348; 2011). In this issue, Cohen and colleagues describe the use of Archaerhodopsin 3 (Arch)—a light-driven proton pump better known for its capacity to silence neurons in optogenetic experiments—as a new class of voltage sensor for mammalian neurons (Nat. Methods 9, 90–95; 2012).

Arch and its nonpumping mutant, Arch(D95N), could resolve individual action potentials in cultured mammalian neurons with high signal-to-noise ratio and low phototoxicity. But there is still much room for improvements that will one day lead to high-quality all-optical electrophysiology both in vitro and in vivo.