Main

Neuronal signaling in the brain involves different kinds on neurons interconnected in complex signaling networks. Each signal consists of a transient decrease in the voltage potential across the plasma membrane lasting just milliseconds, and this transient voltage spike, or depolarization, is capable of triggering a spike in other neurons that contact the activated neuron. Determining how individual neurons communicate with each other and mapping their networks are fundamental goals of neuroscience research.

Despite scientists' best efforts, these goals have been hampered by the lack of a method to precisely stimulate defined neurons in a noninvasive manner. Karl Deisseroth at Stanford University has been interested in finding a solution to this problem and says, “We've been watching this literature for a long time but were never satisfied with the slow properties of the available methods.” These methods could not trigger multiple single spikes on timescales of less than a second as needed to faithfully reproduce the rapidity of signaling seen in vivo.

This situation changed when a group of researchers in Germany discovered a light-activated cation channel in green algae called Channelrhodopsin-2 (ChR2; Nagel et al., 2003). These researchers showed that when this protein was expressed in mammalian cells, brief photostimulation was capable of generating inward cation currents. Deisseroth realized that this protein had the potential to fulfill the needs of neuroscientists looking for a selective and noninvasive method to control neuronal signaling.

To see if ChR2 could trigger voltage spikes in neurons similar to those seen in vivo, Deisseroth and colleagues used a virus to express the channel in primary neurons taken from rat brain (Boyden et al., 2005). They attached YFP to ChR2 so they could identify the neurons in which the channel was expressed. Illumination of individual neurons with blue light triggered fast voltage spikes indistinguishable from natural spikes. (See Fig. 1) By using light pulses as short as 10 ms, the authors showed they could precisely trigger fast and reproducible trains of spikes. They could also reliably produce subthreshold voltage transients, which are important in signaling but difficult to produce artificially.

Figure 1
figure 1

Neurons expressing YFP-tagged Channelrhodopsin-2 and a voltage trace showing photostimulation-elicited spikes.

Full size image

Surprisingly the method worked without requiring application of retinal, which is used to form the light-absorbing chromophore in the protein. Presumably this is due to endogenous retinal or the presence of precursors in the culture medium. Deisseroth says, “Some tissues may need retinal supplementation and some may not. The jury is still out on that.” The other surprise was the lack of toxicity seen even after expression for a week. “I think the reason for this is that ChR2 really is closed or mostly closed unless it is excited by light, and ambient light is about a thousand-fold too weak to excite it,” remarks Deisseroth.

The great potential of the technique, however, will only be realized when it is demonstrated to work in acute brain slices and in vivo. It still isn't known whether it will be capable of functioning without the addition of retinal. Ease of use could also be improved with red-shifted variants that could be created by mutagenesis of ChR2. If this method does work well in vivo, the ability to express ChR2 in defined subsets of neurons should open up entire new avenues of research into neuronal signaling and functional integration.