Sergey Kasparov, Sam Lane & Anja Teschemacher
University of Bristol, UK

https://doi.org/10.36866/pn.93.41

Perhaps no other technology introduced into experimental bio-medicine has ever attracted such a surge of publicity than what is now commonly referred to as ‘optogenetics’. A number of very good reviews on this topic are available and the key facts on molecular and engineering aspects can be found there (Chow et al. 2012; Carter & de Lecea, 2011; Lin, 2011; Mattis et al. 2012). This update is for those who are either considering or only just beginning to use it. It attempts to focus on the key features and history of this technology and share some thoughts on several issues which tend to receive less attention but which are nevertheless important.

In essence, the term ‘optogenetics’ only means a combination of genetic engineering and optical methods. From this perspective this technology can be traced back to the first experiments which applied green fluorescent protein (GFP) or even earlier. Today, two rather distinct families of protein tools are generally included in the term ‘optogenetics’.

The first family comprises numerous fluorescent genetic constructs which are used as reporters or markers. In the simplest scenario they are used to visualise cells of interest, for example specific sub-sets of neurones in the brain, by inducing them to express a fluorescent protein, such as GFP. As we now have so many variants derived from several species with various spectra of excitation and emission, it is possible to combine proteins of different colours and in this way analyse the structure of the brain. This has been used to create the striking ‘brainbow’ effect where simultaneous expression of several differently coloured proteins generates distinct fluorescence hues in adjacent cells, thus facilitating structural analysis of the brain (Livet et al. 2007; Cai et al. 2013). Modifications of fluorescent proteins (mainly of GFP) are also used as indicators of intracellular signalling or trafficking processes. For example, a fluorescent marker protein can be fused to another protein which migrates from cytoplasm to the nucleus upon activation (so called translocation assays). Or, reporter tools combine fluorescent proteins in such a way that the construct changes its fluorescence upon binding with an intracellular second messenger (for example Ca2+) or reacts to another signalling event, for example depolarisation of the membrane potential of a neurone.

It is, however, the second family, the optogenetic ‘actuators’ or ‘effectors’, which were introduced more recently and led to an avalanche of high profile publications. These constructs are derived from proteins which have conferred survival advantages to various simple organisms by allowing them to detect light. Our eyes contain an example of such a protein, rhodopsin, a G-protein-coupled molecule which acts as a photo-transducer. In fact, mammals express several rhodopsin-like molecules and whilst their function in the rods of retina is obvious, the closely related encephalopsins are found in the brain where their role remains a mystery (Blackshaw & Snyder, 1999).

Expressing light-sensitive proteins which can alter cell function in otherwise light-insensitive tissue such as brain or heart offers a remarkable opportunity for inducing changes to the activity of the targeted cells in a non-invasive way. Perhaps the first demonstration that this was actually possible was the study of Zemelman et al. (2002), which used a complex multi-component system of proteins that were co-expressed and used to control neuronal circuits in Drosophila. In that system the signal from the trigger (light) was converted into an intracellular signalling event via G-proteins, which is the natural coupling mechanism for mammalian rhodopsins. While these experiments demonstrated the principle, the system was too complex to be easily adapted to mammalian cells and in addition too slow to allow what was deemed essential – a high degree of temporal control over neuronal action potentials with individual flashes of light.

The breakthrough came with the introduction of channelrhodopsin-2 (ChR2), a protein from the green alga Chlamydomonas reinhardtii (Nagel et al. 2003). ChR2, albeit structurally reminiscent of G-protein coupled receptors (it also has 7 transmembrane domains), works in an entirely different manner. Upon illumination it opens a non-selective cation pore permeable to Na+, K+, H+ and to some extent to Ca2+. In the membrane of a neurone such a conductance leads to a rapid influx of Na+ and depolarisation that, if sufficiently strong, triggers action potentials (Fig. 1). At the same time, at neutral pH, the high H+ permeability of ChR2 is of no significance because there is little drive for protons to move in or out of the cell. Very soon ChR2 was demonstrated to allow efficient control of activity in mammalian neurones (Boyden et al. 2005; Nagel et al. 2005). It came as a surprise that ChR2 did not even require supplementation with all-trans retinal, the obligatory co-factor for opsins, evidently because nervous tissue contains enough retinal which is also present in many culturing media (Boyden et al. 2005). Numerous mutants of ChR2 were then generated with the primary aim to increase channel conductance and speed up channel opening and closing in order to achieve more precise control of neuronal activity (Mattis et al. 2012). Optogenetic silencers (H+ and Cl– pumps) were also developed; see Fig 1 (Gradinaru et al. 2008; Chow et al. 2010).

Increasing conductance is important because in comparison with the native voltage-gated Na+ channels which control action potentials, ChR2 conductance is low and it only works well when expressed at high density. A high level of protein expression can be achieved using powerful genetic targeting systems available today, but this also represents a challenge for the protein-producing and protein-sorting machinery of the cell, which may fail to deliver the majority of the expressed optogenetic actuator to the plasma membrane. After all, ChR2 is a foreign protein to mammals. In addition, fluorescent proteins are usually fused to the actuators and this worsens the situation because these may fold imperfectly or form multimers. This may lead to ER stress and be damaging to the cell. If light-sensitive channels partially remain in the endomembranes, they might be activated upon illumination, possibly releasing Ca2+ stored in intracellular organelles. Whether such retained actuator molecules are functionally relevant and play a role in the effects of ChR2 is not known but should be investigated. Efforts are ongoing to improve maturation of optogenetic proteins and their targeting to the plasma membrane by targeted mutations and by integrating specific trafficking signals.

Adjusting kinetics of ChR2 and similar actuators is important for several reasons. First, faster depolarisations and repolarisations allow more accurate control of neurones with short flashes of light. For some slowly spiking neurones this is less crucial, as typically even the first generation of ChR2-like proteins successfully deliver firing rates of 20–30 Hz. However, for fast spiking cells, frequencies of up to 100 Hz may be desirable and here light-sensitive channels still have room for improvement, in spite of significant recent progress. Ideally one would want to illuminate with light pulses of a duration similar to naturally occurring action potentials (typically around 1 ms) in central neurones, but generating enough current with such short flashes is difficult and may require using high light intensities which may be phototoxic. Hence almost all published studies use light pulses of 5–20 ms or even longer. It is important to realise that the feasibility of driving high frequency bursts with ChR2-like proteins also depends on the type of cell which is to be stimulated. A large neurone would typically have a relatively low input resistance (e.g. significant leak of current across its plasma membrane) and a large membrane capacitance. Hence it is much harder to drive it faster than a small compact cell, which can be efficiently depolarised with relatively modest currents created by ChR2. In addition, comparatively slow ChR2 kinetics means that the membrane remains depolarised slightly longer than the end of the light pulse, thus permitting longer activation of voltage gated Ca2+ channels and stronger Ca2+ influx than during native action potential activity. As a result, neurotransmitter release caused by flashes of light applied to ChR2-expressing synaptic terminals could be stronger than if it was driven by physiological action potentials. The same applies to somato-dendritic release of neurotransmitters when ChR2-expressing somata of neurones are stimulated with light. Powerful somato-dendritic release could have various functional consequences, for example triggering a negative feedback response – consider, for example, noradrenaline acting on α2-adrenoceptors which are coupled to Gi proteins. Overall, there is a fine balance between the speed of activation and inactivation of a light sensitive channel, its unitary conductance, density on the plasma membrane, sensitivity to light and the intrinsic electrical properties of the membrane of the neurone (or other cell type) to be stimulated (Mattis et al. 2012). The cycle of activation and inactivation of light sensitive cation channels can be varied and some mutants, once activated, remain in the open state for many minutes. This property has led to development of the so-called step-opsins (Berndt et al. 2009), which can be used when precise timing with pulsatile light is not essential but prolonged application of light is technically difficult.

A further useful development has been the variation of spectral characteristics of optogenetic actuators. ChR2 and its closest relatives are best activated by blue light, typically by 470 or 445 nm lasers or diodes. However, sometimes it may be desirable to use longer wavelengths for excitation. In vivo longer wavelengths travel further through the tissue thus helping to increase the effective volume of activated tissue. Most of the constructs used today are mutants of ChR2 or its chimeras with ChR1 from the same species and they all have excitation maxima below 500 nm. However, chimeras between ChR1 and VChR1, a light-sensitive channel from Volvox carteri, can be efficiently excited by green or even yellow light (~550 nm (Mattis et al. 2012). The most spectacular development in this direction is the ReaChR, recently published by J. Y. Lin and R. Y. Tsien (Lin et al. 2013).

While numerous mutants of ChR2-like light sensitive channels have been generated, most of the focus has been on either improving their dynamics (faster vs. slower currents) or varying their spectral characteristics (e.g. excitation maxima). By contrast, one rather fundamental property, ion permeability, has received comparatively little attention. Engineering ion-selective versions of ChR2-like channels is, of course, not trivial but could result in some very useful improvements compared to a non-selective cation permeability. First of all, removing the concomitant K+ conductance should make these channels more efficient depolarisers. Second, generating a selective light sensitive Ca2+ pore would be highly advantageous for many studies which focus on the role of Ca2+ but do not require membrane depolarisation (for example for control of non-neuronal cells). Here some efforts have been made (Kleinlogel et al. 2011) but clearly there is room for improvement. Finally, if it was possible to confer K+ selectivity to a ChR2 pore, this would result in a highly efficient light-sensitive ‘off-switch’ for excitable cells. Moreover, theoretically it could be even converted into a step-function off-switch – a construct which once activated by a flash of light remains in that state for a significant length of time. The light sensitive ion pumps Arch and NpHR (Fig. 1) are currently available for this purpose but work on a slightly slower time scale and require constant illumination because they are energy dependent (Chow et al. 2012).

In addition to light-sensitive ion channels, two further classes of optogenetic actuators are increasingly coming into use and deserve mentioning (Fig 2). The first class are G-protein-coupled light-sensitive proteins, one example of which are the opto-adrenoceptors, chimeras of rhodopsin and adrenergic receptors with a choice of intracellular coupling mechanisms. Construction of such chimeras was possible because of the domain organisation of G-protein-coupled receptors, which permits switching of various components of these proteins (Kim et al. 2005; Airan et al. 2009). These tools cannot be used to control neuronal spiking with any temporal precision, but on the other hand they offer a very attractive way for exploring effects controlled via G-protein-mediated signalling cascades. Since the original publication in 2009, there have been no studies using these receptors, which may be suggestive of technical difficulties. A recent study described a new light-sensitive chimeric G-protein-coupled receptor, ‘JellyOp’, which uses parts of a bleach resistant opsin from Carybdea rastonii, the box jellyfish (Bailes et al. 2012). JellyOp couples to adenylate cyclase and promises to be a useful addition to the previously available toolbox.

The second class are the light sensitive enzymes, exemplified by the photoactivated adenylyl cyclase from Beggiatoa (Stierl et al. 2011). The interesting feature of this approach is that a short pulse of light is sufficient to activate the enzyme, which then generates cAMP for some time and thus greatly reduces the risk of photo damage. The difficulty is to make such proteins completely inactive in the dark state, something yet to be achieved.

For obvious reasons the ability to selectively activate selected neuronal populations by light was most appealing to neuroscientists working to disentangle multiple and intermingled neuronal networks within the brain. However, the use of optogenetic actuators in neuroscience is not limited to triggering or inhibiting neuronal action potentials. In fact, optogenetics can be effectively used to study non-excitable parts of the brain supercomputer, the glia cells. When ChR2 is expressed in astrocytes, light flashes transiently depolarise astrocytes, as they would with neurones, but these depolarisations appears to have little immediate effect on astrocytes or adjacent neurones. However, optogenetic activation of astrocytes leads, with some delay, to the release of ‘glio-transmitters’ such as ATP and activation of astrocytic and connected neuronal networks. Given that there is currently no other non-genetic way to selectively activate astrocytes, this imperfect tool can nevertheless be very effective (Gourine et al. 2010; Figueiredo et al. 2011). Opto-adrenoceptors and JellyOP mentioned above could be preferred tools for selective control of non-excitable cells where cellular physiology is often controlled by IP3 or cAMP concentrations. Control of cellular function by light is of course also applicable to tissues outside of the brain (Auslander & Fussenegger, 2012).

Finally, what is left to do for genetics in opto-genetics? In fact, the success of optical interrogation of brain function is critically dependent on the ability to selectively express optogenetic actuators and other molecular tools exclusively in the cells of interest. Two general strategies are currently being used for targeting. Either the experimentalist relies on the use of a viral vector with a cell-selective promoter (e.g. CamKII or GFAP promoters) to target expression of the actuator. Or, the selectivity is determined by transgenic expression of CRE recombinase while the optogenetic gene is delivered by a CRE-dependent viral vector or by breeding a CRE-driver mouse into another strain which contains the construct which needs to be switched on by CRE. Numerous CRE-deriver mice are available and many have very tightly controlled cellular specificity. However, for many types of experiments, the rat remains a better option, and the choice of CRE-driver rats is extremely limited. Achieving a high degree of specificity with short promoters that are suitable for viral vector-mediated expression is difficult and unfortunately only a few such promoters are yet available, but where they are available, this approach offers great flexibility and speed. The downsides are some risk of tissue reactions to viral particles and the fairly small choice of potent and selective promoters, which are very hard to design rationally. Still, we believe that many more cell-selective viral targeting systems can and should be generated. Altogether, improvements in the gene expression systems are essential for future progress in application of optical tools.

In summary, optogenetics has now reached the stage where it has become a recognised and widely used technology in physiology. There are many improvements which can and should be made to the existing tools and their means of delivery. We look forward to further developments in the near future.

References

Airan RD, Thompson KR, Fenno LE, Bernstein H & Deisseroth K (2009). Temporally precise in vivo control of intracellular signalling. Nature 458, 1025–1029.

Auslander D & Fussenegger M (2012). Optogenetic therapeutic cell implants. Gastroenterology 143, 301–306.

Bailes HJ, Zhuang LY & Lucas RJ (2012). Reproducible and sustained regulation of Galphas signalling using a metazoan opsin as an optogenetic tool. PLoS ONE 7, e30774.

Berndt A, Yizhar O, Gunaydin LA, Hegemann P & Deisseroth K (2009). Bi-stable neural state switches. Nat Neurosci 12, 229–234.

Blackshaw S & Snyder SH (1999). Encephalopsin: a novel mammalian extraretinal opsin discretely localized in the brain. J Neurosci 19, 3681–3690.

Boyden ES, Zhang F, Bamberg E, Nagel G & Deisseroth K (2005). Millisecond-timescale, genetically targeted optical control of neural activity. Nature Neuroscience 8, 1263–1268.

Cai D, Cohen KB, Luo T, Lichtman JW & Sanes JR (2013). Improved tools for the Brainbow toolbox. Nat Methods 10, 540–547.

Carter ME & de Lecea L (2011). Optogenetic investigation of neural circuits in vivo. Trends Mol Med 17, 197–206.

Chow BY, Han X & Boyden ES (2012). Genetically encoded molecular tools for light-driven silencing of targeted neurons. Prog Brain Res 196, 49–61.

Chow BY, Han X, Dobry AS, Qian X, Chuong AS, Li M, Henninger MA, Belfort GM, Lin Y, Monahan PE & Boyden ES (2010). High-performance genetically targetable optical neural silencing by light-driven proton pumps. Nature 463, 98–102.

Figueiredo M, Lane S, Tang F, Liu BH, Hewinson J, Marina N, Kasymov V, Souslova EA, Chudakov DM, Gourine AV, Teschemacher AG & Kasparov S (2011). Optogenetic experimentation on astrocytes. Exp Physiol 96, 40–50.

Gourine AV, Kasymov V, Marina N, Tang F, Figueiredo MF, Lane S, Teschemacher AG, Spyer KM, Deisseroth K & Kasparov S (2010). Astrocytes control breathing through pH-dependent release of ATP. Science 329, 571–575.

Gradinaru V, Thompson KR & Deisseroth K (2008). eNpHR: a Natronomonas halorhodopsin enhanced for optogenetic applications. Brain Cell Biol 36, 129–139.

Kim JM, Hwa J, Garriga P, Reeves PJ, RajBhandary UL & Khorana HG (2005). Light-driven activation of beta 2-adrenergic receptor signaling by a chimeric rhodopsin containing the beta 2-adrenergic receptor cytoplasmic loops. Biochemistry 44, 2284–2292.

Kleinlogel S, Feldbauer K, Dempski RE, Fotis H, Wood PG, Bamann C & Bamberg E (2011). Ultra light-sensitive and fast neuronal activation with the Ca(2+)-permeable channelrhodopsin CatCh. Nat Neurosci 14, 513–518.

Lin JY (2011). A user’s guide to channelrhodopsin variants: features, limitations and future developments. Exp Physiol 96, 19-25.

Lin JY, Knutsen PM, Muller A, Kleinfeld D & Tsien RY (2013). ReaChR: a red-shifted variant of channelrhodopsin enables deep transcranial optogenetic excitation. Nat Neurosci 16, 1499-1508.

Livet J, Weissman TA, Kang H, Draft RW, Lu J, Bennis RA, Sanes JR & Lichtman JW (2007). Transgenic strategies for combinatorial expression of fluorescent proteins in the nervous system. Nature 450, 56–62.

Mattis J, Tye KM, Ferenczi EA, Ramakrishnan C, O’Shea DJ, Prakash R, Gunaydin LA, Hyun M, Fenno LE, Gradinaru V, Yizhar O & Deisseroth K (2012). Principles for applying optogenetic tools derived from direct comparative analysis of microbial opsins. Nat Methods 9, 159–172.

Nagel G, Brauner M, Liewald JF, Adeishvili N, Bamberg E & Gottschalk A (2005). Light activation of channelrhodopsin-2 in excitable cells of Caenorhabditis elegans triggers rapid behavioral responses. Curr Biol 15, 2279–2284.

Nagel G, Szellas T, Huhn W, Kateriya S, Adeishvili N, Berthold P, Ollig D, Hegemann P & Bamberg E (2003). Channelrhodopsin-2, a directly light-gated cation-selective membrane channel. Proc Natl Acad Sci U S A 100, 13940–13945.

Stierl M, Stumpf P, Udwari D, Gueta R, Hagedorn R, Losi A, Gartner W, Petereit L, Efetova M, Schwarzel M, Oertner TG, Nagel G & Hegemann P (2011). Light modulation of cellular cAMP by a small bacterial photoactivated adenylyl cyclase, bPAC, of the soil bacterium Beggiatoa. J Biol Chem 286, 1181–1188.

Zemelman BV, Lee GA, Ng M & Miesenbock G (2002). Selective photostimulation of genetically chARGed neurons. Neuron 33, 15–22.