My recent attempts at making a red optogenetics stimulation system (in the 620 nm range) were partially successful, in that I was able to make something that would produce about 3 mW from the end of an implanted 200 µm fibre. Unfortunately, this is less power than I wanted (ideally more like 10 mW), but my research into which opsin to use led me to a recent paper from Karl Diesseroth’s group1, where they develop a new red-shifted super sensitive opsin.
In this work, Marshel et al. begin by data mining a marine microbes genetic database for potential channelrhodopsins (Figure 1A/B). This kind of largescale screening is not my forté, but it was very interesting the way they used previous channelrhodopsin crystal structure to inform their screen as to variants that had novel sequences specifically for the residues that form the ion pore, thereby increasing their chances of finding a variant with very different properties.
They then transfected these channelrhodopsin variants into cultured neurones and used patch clamping to determine the channel properties, which allowed them to narrow down to a the most promising super sensitive opsin, of a previously unknown class. They called it chRmine, supposedly to be pronounced carmine (like the colour), but my brain can’t call it anything other than chromine.
Anyway, naming aside, it seems to be an excellent opsin for use in optogenetics experiments. It shows huge excitatory photocurrents in the orange and red spectra, when compared to 2 other red-responsive opsins, Chrimson and bReaChES (Figure 1C). They then go through a series of opto-electrophysiology recordings to demonstrate the suitability for optogenetics experiments, beginning with the reversal potential, which demonstrates Na+/K+ permeability, which is great for driving neuronal spiking (Figure 1D).
ChRmine induces neuronal firing with very short pulses (Figure 1E) and very low irradiance (Figure 1F), both on an order of magnitude better than the compared opsins. This sensitivity is so important for optogenetics, as it makes it much easier to deliver the amount of light needed and you can limit the phototoxicity damage.
The other important quality of chRmine is that it has very quick rise and fall times – essentially it opens and closes very quickly upon light activation, which allows very fast (40 Hz) spiking (see supplementary info from Marshel et al.).

The figure I’ve shown here is only half of the first of 6 figures (not including supplemental information) from this absolute behemoth of a paper. The complexity of the experiments is far beyond my experience and understanding in neuroscience, but they use their new red super sensitive opsin to control individual cortical neurones with millisecond precision for a series of learning and behaviours in mice. Please do check it out if that’s your bag, but it’s far too “neurosciencey” for me.
Anyway, back to my attempts at making a red-based opto system. I’ve managed to achieve 3 mW of 620 nm light from a 200µm fibre with an easy-to-use LED system, which is on the low end for classic red optogenetics, but should do great for the super-sensitive chRmine.
And to get an idea of how well it would work in a mouse’s brain, I went to Karl Diesseroth’s optogenetics website to use his irradiance prediction tool2. For those who’ve never used this, it’s quite handy – you input various parameters for your study, including the wavelength of light and details of your optic fibre cannula, and it shows you the light spread you can expect through the brain (Figure 2).

So, going back at the Marshel paper, they show 100% activation of chRmine at 0.08 mW/mm2. Looking at the predicted irradiance values, that comes out at a depth of 2.3 mm from the fibre tip, which is immense – in an experiment that would illuminate deeper than the entire hypothalamus. Clearly then, we could drop the intensity of the stimulation light, but by how much? Using Diesseroth’s irradiance calculator, I input a series of decreasing light intensities and noted how deep you can go and maintain 0.08 mW/mm2 (Figure 3).

Given experimental practicalities, I would be happy to have 1.2 mm of activation depth, which translates to about 0.3 mW of power, which is a 10-fold reduction even from my relatively low power LED system. This would obviously need validating in vivo, but it’s very promising, and should negate a lot of the classic difficulties with obtaining high light power for optogenetics. I’ve had a peek, and it seems that Addgene stock chRmine AAV’s, so I am excited to try this out.
1. Marshel et al., Science 365, eaaw5202 (2019) Cortical layer-specific critical dynamics triggering perception.
2. https://web.stanford.edu/group/dlab/cgi-bin/graph/chart.php
2 thoughts on “The Power of Red Shift”