Advances and prospects of rhodopsin-based optogenetics in plant research

Microbial opsins

After two decades of optimization and rhodopsin engineering, a large set of microbial rhodopsins (opsin + covalently bound retinal = rhodopsin) with different functions is routinely applied in neuroscience and other animal research. Rhodopsins function as light-gated ion channels, ion pumps, or enzymes and some of them have been engineered into voltage sensors (Figure 1). We can assume that the in planta production of retinal, established recently, will allow the use of a wide range of these available rhodopsins, which are briefly described below.

Light-gated cation and anion channels

Cation ChRs are used to excite neurons by depolarizing the membrane (Figure 1A) and originate mostly from algae, first discovered in Chlamydomonas reinhardtii (Nagel et al., 2002, 2003; Kateriya et al., 2004). Improvement of optogenetic tools started early on, for example, the first mutant of ChR2 with larger stationary photocurrents—ChR2-H134R—was already applied in 2005 (Boyden et al., 2005). At present, many additional ChRs were described from different species or optimized by genetic engineering. These new ChRs display different characteristics of channel ion selectivity, kinetics, light sensitivity, action spectrum, membrane targeting, and expression level (Lin et al., 2009; Gunaydin et al., 2010; Govorunova et al., 2017). A representative list of ChR variants most frequently used in animal research is summarized in Table 1.

The desire for repressing neuronal activity by light-induced hyperpolarization led to the application of a light-activated chloride pump, see below (Zhang et al., 2007). Functionally much more effective, natural anion ChRs (ACRs, GtACR1, and GtACR2) were discovered from the cryptophytic alga G. theta, having a strict anion-selectivity and high photocurrents (Figure 1A;Govorunova et al., 2015). ACRs enable effective inhibition of action potentials (Govorunova et al., 2016) at low cytoplasmic Cl⁻ concentration (Mohammad et al., 2017), more powerful than the hyperpolarizing ion pumps used before (Zhang et al., 2007). More ACRs with different characteristics have since been engineered and identified (Govorunova et al., 2017; Oppermann et al., 2019), summarized in Table 1.

Light-gated ion pumps

Already in 1971 and 1973, a light-activated proton pump termed bacteriorhodopsin from Halobacterium salinarum (Oesterhelt and Stoeckenius, 1971, 1973), and the proteorhodopsins (Béja et al., 2000) were discovered, respectively. Light-activated proton pumps, especially the recently discovered ones, like Arch (Chow et al., 2010) and Mac (Husson et al., 2012), and their improved versions (Gradinaru et al., 2010; Mattis et al., 2011) inhibit neural activity by H⁺ efflux and membrane hyperpolarization upon illumination (Figure 1A). Genetic engineering converted proton pump rhodopsins recently into proton channels for bidirectional manipulation of H⁺ (Fudim et al., 2019). Light-gated proton channels are of minor interest for animal research due to the lack of pH differences between cytosol and extracellular space. However, in plants, the pH gradient of up to 2 pH units between vacuoles and cytosol and between apoplast and cytosol represents a perfect application for rhodopsin-based proton channel-mediated pH changes.

The microbial rhodopsin initially used for hyperpolarizing the membrane potential and inhibition of neuronal firing was the chloride pump Halorhodopsin (HR; Zhang et al., 2007). The HR found in Natronomonas pharaonis (NpHR) has an excitation maximum at about 590 nm and induces inward Cl⁻ flux across the plasma membrane (Figure 1A;Lanyi, 1986). Under suitable light illumination, HR therefore hyperpolarizes the membrane. However, the wild-type NpHR shows a limited expression level and produces low photocurrents in neurons, which was partially improved by enhanced versions like eNpHR 2.0 and eNpHR 3.0 (Gradinaru et al., 2008, 2010; Arrenberg et al., 2010). As HR can substantially inactivate with prolonged illumination at its absorption maximum, additional blue or violet light is required for efficient Cl⁻ transport (Zhang et al., 2019).

In 2013, a light-driven sodium ion pump rhodopsin, termed KR2, was discovered from the marine flavobacterium Krokinobacter eikastus, which could pump Na⁺ out of the cell (Figure 1A;Inoue et al., 2013). Based on the analysis of the crystal structure, a mutant with light-driven K⁺ pump activity was engineered (Kato et al., 2015; Gushchin et al., 2016). With the KR2 P219T/S254A mutant, this nonproton pump could be a good starting point for the design of interesting cation pumps for plant optogenetics (Inoue et al., 2019; Kovalev et al., 2019). A list of ion pump rhodopsins is summarized in Table 1.

Light-activated enzyme rhodopsins

Rhodopsins with light-sensitive enzymatic activity have properties different from the classical microbial rhodopsins. In contrast to channel- and pump rhodopsins with seven transmembrane helices (7 TM), all enzyme rhodopsins, identified so far, have an 8 TM structure with both the N- and C-terminal end in the cytosol (Figure 1B), as first demonstrated for the fungal guanylyl cyclase opsin Cyclop (Gao et al., 2015). They contain an enzymatic domain at the C-terminus, such as a cyclic nucleotide phosphodiesterase (Yoshida et al., 2017; Tian et al., 2018, 2020) or a nucleotidyl cyclase (Gao et al., 2015; Scheib et al., 2018; Tian et al., 2018; Figure 1B). Representative enzyme rhodopsins that can be used to manipulate second messenger cAMP and cGMP are summarized in Table 1.

Rhodopsin-based voltage sensors

The activity of a large number of neurons can be routinely recorded using genetically encoded calcium indicators (GECIs) (Zhong and Schleifenbaum, 2019; Dalangin et al., 2020; Gasterstadt et al., 2020; Subach et al., 2020; Zarowny et al., 2020; Shemetov et al., 2021). However, calcium signals are only a proxy to investigate signaling in neurons (Emiliani et al., 2015; Bando et al., 2019) and therefore organic voltage-sensitive dyes are often applied instead to directly measure electric activity with faster kinetics than the Ca²⁺ indicators (Loew, 2010; Strack, 2019; Fiala et al., 2020; Kuhn et al., 2021). However, the voltage-sensitive dyes exhibit higher phototoxicity and allow neither genetically targeted delivery nor long-term imaging studies of single cells in mammalian brain. Genetically encoded voltage indicators (GEVIs) can overcome these challenges, enabling imaging of fast voltage spiking and subthreshold dynamics (Bando et al., 2019, 2019; Hashemi et al., 2019; Leong et al., 2021).

Early generations of GEVIs operated by a voltage-sensitive domain (VSD; Murata et al., 2005; Tsutsui et al., 2013) fused to fluorescent proteins (FPs) (Peterka et al., 2011; Mutoh et al., 2012). However, these VSD-based sensors had limited ability to capture sub-threshold voltage dynamics for precise action potential recordings. Later, GEVIs based on rhodopsins were designed (Figure 1C), exhibiting two desirable features for a voltage indicator—fast and linear. The first generation of rhodopsin-based GEVIs, based on the Proteorhodopsin optical proton sensor (Béja et al., 2000, 2001; Kralj et al., 2011), showed poor membrane targeting in eukaryotic cells. A significant improvement was achieved with Arch (Chow et al., 2010), as it could resolve individual action potentials in mammalian neurons in vitro with high signal-to-noise ratio (SNR), low phototoxicity, sub-millisecond temporal resolution, and subcellular spatial resolution. Later, the Arch/D95N mutant was engineered to eliminate the photocurrent induced by wild-type Arch (Kralj et al., 2012). In recent years, the improvements of rhodopsin-based GEVIs are mainly based on mutations of Arch, including the Arch mutant sensors (Arch-EEN and -EEQ) with faster kinetics and greater fluorescence dynamic range than Arch/D95N, improving nearly three-fold the fidelity of spike detection (Gong et al., 2013; Flytzanis et al., 2014; Piatkevich et al., 2018). QuasAr1 and QuasAr2 showed improved brightness and voltage sensitivity with microsecond response times (Hochbaum et al., 2014). QuasAr3 shows high expression and excellent trafficking to the plasma membrane in vivo, and can report action potentials with fast kinetics and a high SNR (Piatkevich et al., 2018; Adam et al., 2019).

To correct for motion artifacts and to increase SNR, the electrochromic fluorescence resonance energy transfer GEVIs were developed, in which the rhodopsin serves as a VSD and a FP is fused after the 7th helix (Figure 1C; Gong et al., 2014). Following this approach, many rhodopsin-based GEVIs were optimized, such as QuasAr2 with several FPs (Fluorescent proteins, eGFP, Citrine, mOrange2, mRuby2; Zou et al., 2014), the faster Acetabularia rhodopsin (Ace) with mNeonGreen (Gong et al., 2015; Abdelfattah et al., 2019) and the Ace with mRuby3 (red-shifted VARNAM; Kannan et al., 2018). The rhodopsin Ace2 was also fused to a self-labeling tag domain (HaloTag) with different fluorophores to provide a set of bright and photostable chemigenetic indicators (Voltron) for in vivo voltage imaging (Abdelfattah et al., 2019).

Very few studies in plants have used voltage-sensitive probes for quantitative detection of membrane potential via spectroscopy, so far (Konrad and Hedrich, 2008). The use of a range of applicable tools based on rhodopsins could be a major boost to plant science to study electric phenomenon and their physiological role.

Functional rhodopsins for plant research

In contrast to most animal cells, retinal is not produced in plant cells. However, opsins need the essential cofactor retinal to form the chromophore-containing rhodopsin and to function (Nagel et al., 2003; Ullrich et al., 2013). ChR2-XXL is a powerful ChR2 mutant that showed robust expression and enhanced channel activity even at low retinal concentration in several animal cell systems (Dawydow et al., 2014). The cation conducting ChR2-XXL is characterized by a high light-sensitivity and was recently expressed in plants (transient expression in Nicotiana benthamiana and stable expression in Arabidopsis; Reyer et al., 2020). In this study, blue light induced plasma membrane depolarizations in transgene Arabidopsis plants when retinal was applied to the leaf surface (Figure 2A). This approach enabled identifying the plasma membrane H⁺-ATPase as a major driver for membrane potential repolarization.

So far, transgene Arabidopsis with other ChR2 mutants did not result in successful optogenetic applications, probably due to the ChR degradation in the absence of retinal (Ullrich et al., 2013). Thus, the presence of retinal is necessary for broad applications of many microbial rhodopsins, even in animal cell systems (Nagel et al., 2003; Boyden et al., 2005; Ullrich et al., 2013; Dawydow et al., 2014). In our recent study (Zhou et al., 2021a), a dioxygenase from a marine bacterium (MbDio; Kim et al., 2009) was expressed in the chloroplast by N-terminal fusion with a synthetic chloroplast-targeting peptide (Shen et al., 2017). The MbDio could then convert β-carotene into retinal in vivo, providing the essential chromophore to form functional rhodopsins (Figure 2B). The endogenous retinal production boosted the targeted expression of the light-sensitive anion channel GtACR1 to the plasma membrane (Figure 2B). The same strategy recently improved also the expression of two other anion ChRs in plant cells (Zhou et al., 2021b). In line with its activation spectrum (Govorunova et al., 2015, 2017), GtACR1, when illuminated by green light, induced pronounced membrane potential depolarizations and enabled noninvasive manipulation of plant growth and development (Figure 2B; Zhou et al., 2021a). After GtACR1 expression in pollen tubes, global illumination inhibited polar growth, while local light application at the flanks of the apical dome enabled manipulating the growth direction (Figure 2B; Zhou et al., 2021a). This presented direct proof for pollen tube growth to be steered by anion flux and/or the resulting membrane depolarization, for which there was already good evidence (Gutermuth et al., 2013, 2018; Herbell et al., 2018). A membrane depolarization is discussed as the initial event for stomatal closure, which is normally mediated by abscisic acid (ABA)- and pathogen-dependent signaling pathways to activate anion channels (Konrad et al., 2018). Whether anion channel activation represents the key event in stomatal closure has recently been elucidated by the application of GtACR1 in guard cells (Huang et al., 2021). The light-driven activation of GtACR1 alone led to rapid stomatal closure in response to green light without the need to trigger complex and partly Ca²⁺-dependent signal transduction cascades (Huang et al., 2021). Just as in pollen tubes and mesophyll cells, activation of GtACR1 in guard cells leads to huge anion-driven photocurrents causing a membrane depolarization. This anion efflux dependent depolarization activates the voltage-dependent guard cell outward-rectifying K⁺-channel (GORK), together resulting in a net KCl efflux causing stomata closure (Figure 2C).

The prospects of rhodopsin-based optogenetics in plants

To survive, plants have developed physical and biochemical strategies for adaptation, based on chemical and electrical signals (Hedrich, 2012; Gilroy et al., 2016). These chemical and electrical signals act locally but may also be conducted throughout the plant to communicate local stress, such as pathogen attack, so that a systemic response can be triggered (Farmer et al., 2020). Ion transport and the corresponding membrane potential changes underpin important components in cellular and long-range plant signal transduction, but the exact role of the electrical signals, Ca²⁺ or pH changes is largely unclear (Hedrich et al., 2016; Choi et al., 2017; Farmer et al., 2020).

One of the first responses of leaves to various stress stimuli is an electric response, namely a membrane depolarization, together with changes in the cytosolic Ca²⁺ concentration ([Ca²⁺]_cyt)- and pH homeostasis. These chemo-electric signals in the mesophyll are thought to underlie the same mechanism as in guard cells (Roelfsema et al., 2012; Güzel Deger et al., 2015) albeit they can be triggered by other stimuli (Elzenga et al., 1995; Elzenga and Volkenburgh, 1997; Shabala and Newman, 1999). Early ion signaling events in the mesophyll upon pathogen stress include influx of Ca²⁺, cytosolic acidification and a plasma membrane depolarization (Jeworutzki et al., 2010; Li et al., 2021) likely mediated by S-type anion channels (Güzel Deger et al., 2015). In guard cells, the phytohormone ABA triggers a similar Ca²⁺-transient, an anion channel-dependent depolarization, but a cytosolic alkalization (Levchenko et al., 2005; Li et al., 2021). On the tissue scale, it is believed that [Ca²⁺]_cyt and electric signals are interconnected (Gilroy et al., 2016) and travel together from a stressed leaf to distal leaves in order to control bulk stomatal transpiration, a process called systemic acquired acclimation (Devireddy et al., 2018). Rapid transmission of the [Ca²⁺]_cyt and electric wave is known to take place in the phloem and it is passed on by plasmodesmata connections to and within the mesophyll tissue (Nguyen et al., 2018; Toyota et al., 2018).

Leaf-to-leaf signals or cell-to-cell signals within leaves communicated by electric depolarization waves have recently been associated with the activity of Ca²⁺-permeable glutamate-receptor-like channels (Choi et al., 2014; Nguyen et al., 2018; Shao et al., 2020) while the membrane repolarization is linked to H⁺-ATPase activity (Kumari et al., 2019; Reyer et al., 2020). Microbial rhodopsins are ideally suited to dissect the role of individual components of these cooperative chemical–electrical signals. Noninvasive light-induced membrane potential manipulation in defined cell types will gain insight in these putative signaling mechanisms. The anion ChRs were already shown in different plant cell systems to induce rapid green light-dependent depolarization with fast on- and off-kinetics (Zhou et al., 2021a; Zhou et al., 2021b; Huang et al., 2021). By using photosynthetically inactive green light, it is possible to produce depolarizations with defined time and intensity regimes without triggering the intrinsic blue- and red light photoreceptors, known to cause electric responses (Stoelzle et al., 2003; Ando and Kinoshita, 2018). So called systemic potentials (Szechynska-Hebda et al., 2017), which represent hyperpolarizing waves in plants, could be optogenetically generated by defined activation of suitable H⁺-pumping rhodopsins, such as eArch 3.0 (Mattis et al., 2011) or the inward-directed Cl⁻ pump eNpHR 3.0 (Gradinaru et al., 2010). This would clarify whether this type of electrical wave acts as a signal or whether it is just a passive reaction to changes in membrane transport processes without a signal function. The expression of either anion or cation channel rhodopsin variants, or even co-expression of both in the phloem tissue will shed light on the contribution of membrane potential depolarization and [Ca²⁺]_cyt signals to trigger/transmit long-distance stress signals for systemic acquired resistance or systemic acquired acclimation (Szechynska-Hebda et al., 2017; Hilleary and Gilroy, 2018; Hake and Romeis, 2019; Farmer et al., 2020).

It is widely accepted that [Ca²⁺]_cyt plays an important role in stomatal movement and a Ca²⁺-dependent signaling pathway for the activation of anion channels has been entirely reconstituted in Xenopus oocytes (Geiger et al., 2011; Hubbard et al., 2012). However, Ca²⁺-independent stomatal closure mechanisms exist in parallel (Levchenko et al., 2005; Marten et al., 2007). A certain frequency of repetitive [Ca²⁺]_cyt increases was proposed to encode for stomatal closure (Allen et al., 2000, 2001) and it is widely accepted that Ca²⁺ signaling is associated with deflating of guard cells (Hubbard et al., 2012). However, [Ca²⁺]_cyt increases during stoma opening are also observed (Irving et al., 1992; Young et al., 2006; Harada and Shimazaki, 2008). Which role Ca²⁺ exactly plays and which Ca²⁺ signatures (if at all) contribute to control stomatal opening or closing can now be precisely studied by optogenetic means using ChR2 variants with high Ca²⁺ permeability (Duan et al., 2019). Employing ChR2-XXM, distinct [Ca²⁺]_cyt profiles via precise application of defined light regimes could be triggered in guard cells to shed light on those unanswered questions.

In addition to controlling guard cell behavior, steady or oscillating Ca²⁺ elevations have fundamental roles during normal plant growth and development (Dodd et al., 2010; Stael et al., 2012) and are involved in many plant stress responses to pathogens, drought or salinity (Kudla et al., 2010, 2018; Romeis and Herde, 2014; Wilkins et al., 2016). How [Ca²⁺]_cyt dynamics, which are often associated with electric and pH alterations, result in specific physiological responses has been the subject of intense research in the last two decades. The long-standing hypothesis that Ca²⁺ signatures with different frequencies and intensities encode defined responses (Webb et al., 1996; Dodd et al., 2010; Whalley et al., 2011; Whalley and Knight, 2013) can now finally be tested by optogenetic means using ChR2-XXM. Both ChRs and ACRs are perfectly suited to induce defined Ca²⁺ or voltage dynamics, respectively, in various plant tissues and even in defined cell types (Figure 3A).

Subcellular-targeted manipulation with optogenetic methods is also possible by targeting the rhodopsins to subcellular compartments, like the vacuole or endoplasmic reticulum (Figure 3B), known to be Ca²⁺-stores (Stael et al., 2012). It is likely that rhodopsins targeted to subcellular compartments benefit from the production of retinal in chloroplasts as well, because the retinal is membrane-permeable and might already facilitate the protein folding and stability immediately after the translation step.

Similar to the ChRs, which control [Ca²⁺]_cyt dynamics, optogenetic tools for manipulating cellular pH will be groundbreaking to understand the role of H⁺ in plant signaling, since H⁺ have been identified as bona fide second messengers in animals and yeast (Beg et al., 2008; Dechant et al., 2010). The known rhodopsin-based H⁺-pumps, or H⁺-channels that could derive from those by genetic engineering, would be perfectly suited to investigate the physiological role of pH-changes, a rather unexplored field in plant physiology. Optogenetic tools that could specifically induce alkalization or acidification would help to elucidate the perhaps cooperative role of Ca²⁺ and H⁺ as second messengers. Recently, simultaneous Ca²⁺ and pH measurements revealed pH changes to play a crucial role in stomatal physiology, pathogen responses and in pollen tube growth, possibly even upstream of Ca²⁺ (Li et al., 2021). An increasing interest in the physiological role of cellular pH changes (Behera et al., 2018; Waadt et al., 2020; Li et al., 2021) and pH sensing (Geilfus, 2017; Kesten et al., 2019) would benefit from optogenetic tools to approach unanswered questions related to the role of pH in a plant’s life.

The open questions listed here, which could be clarified by means of different rhodopsins, represent only the tip of the iceberg of possibilities, due to the recent technical achievement of in planta retinal production and improved rhodopsin targeting (Zhou et al., 2021a; Zhou et al., 2021b).

Combining optogenetics with sensing cellular parameters

Imaging technologies with an increasing set of fluorescence sensors, either based on synthetic dyes or genetically encoded reporter proteins are being engineered for combining optogenetics with ion- or voltage imaging (Gonzalez et al., 2021). Thus, cellular processes in intact cells and tissues can be manipulated and visualized at the same time, allowing causal relationships in cell signaling to be understood. The combination of optogenetic actuators with sensors of cellular parameters gave rise to a field called all-optical physiology (Emiliani et al., 2015; Doering et al., 2018). This combined approach requires cross-talk free spectral properties of the rhodopsin actuators and fluorescent sensors, which will be particularized here. We list a number of possible actuator/sensor pairs successfully used in animal research and discuss their applicability in plants, having the presence of native plant photoreceptors in mind.

Actuator/sensor pairs for possible plant applications

Simultaneous application of optogenetics and fluorescence sensors allows to decipher causal relationships between a physiological response and optogenetic manipulation noninvasively in space and time. However, the excitation wavelength for the sensor should not be in the range of the excitation spectrum of the optogenetic tool, otherwise an uncontrolled change of the cell already occurs ahead of time. Many of the rhodopsin-based optogenetic tools used in animal cell systems have excitation spectra in the blue or blue-green spectrum. This has prompted the use and development of fluorescence sensors in the orange, red, and even near-infrared (NIR) spectrum that can be simultaneously applied without interfering with the absorbance spectrum of the optogenetic tool. The most widely used actuator to trigger light-dependent [Ca²⁺]_cyt increases and a membrane potential (V_m) depolarization is ChR2 and its variants or homologs from different species (see Table 1).

Together with ChR2, synthetic dyes are used as sensors to image [Ca²⁺]_cyt during optogenetic manipulation. ChR2 (C128S)-induced [Ca²⁺]_cyt-changes have been studied e.g. with Rhodamine-2 (Rhod-2; Balachandar et al., 2020) or Calcium Orange in different cell types (Mohanty et al., 2008; Bernardinelli et al., 2011). Cultured neuron cells loaded with the Ca²⁺-indicator Fluo-4 showed ChETA (a ChR2 variant)-induced [Ca²⁺]_cyt transients when optogenetically stimulated (Zabolocki et al., 2020). While the spectra of Rhod-2 and Calcium Orange are essentially red-shifted to prevent optical cross-talk with the ChR2 (variant; Table 1;Supplemental Table S1), the excitation spectrum of Fluo-4 largely overlaps with the ChR2-activation spectrum (Renault et al., 2015; Klimas et al., 2020). The strategy to use latter dyes was to excite the dye with very low excitation light to minimize ChR2 activation; however, some residual activation of the optogenetic tool remains in the course of Ca²⁺-imaging (Zabolocki et al., 2020). The synthetic dyes Fluo-4, Rhod-2, and Calcium Orange are loaded to the animal cell lines by incubating with their acetometoxymethyl (AM) ester form, a routine technique nowadays (MacLean and Yuste, 2009). Unfortunately, the use of dye loading to cells with AM esters in plant science is hampered by the presence of esterases in the cell wall space that cleave off the AM ester before the dye can enter the cells in its cell-permeable form. Then again, not all synthetic dyes are administered via the AM-ester route into cells. Voltage-sensitive dyes, for example, do not have to be loaded into the cells, but are often combined with surfactants to overcome their low solubility.

Detection of the electrical activity by means of voltage-sensitive dyes that report changes in V_m is of great importance and meaningful for neuroscience, as it bypasses the need for electrode-based recordings. Voltage imaging with a NIR-absorbing phosphine oxide rhodamine voltage reporter called poRhoVR 14 combined with ChR2 allowed all-optical electrophysiology with neurons (Gonzalez et al., 2021). The excitation and emission spectrum of the poRhoVR 14 dye allows multicolor experiments in concert with commonly employed optogenetic actuators but also other sensors with excitation/emission spectra <650 nm (Figure 4). The long-wavelength voltage-sensitive dye PGH1 has a high water solubility compared to other styryl dyes and an absorption spectrum (peak at ∼710 nm; Figure 4) separate from that of ChR2 and eNpHR3.0 (Salama et al., 2005; Patrick et al., 2007). PGH1 monitored electric excitation or silencing of cardiac cells with ChR2 or eNpHR3.0, respectively (Park et al., 2014). Two other NIR voltage probes called BeRST1, a sulfonated silicon rhodamine-based dye (Huang et al., 2015) and Di-4-ANBDQBS, a styryl dye (Matiukas et al., 2007) were employed recently to record neuronal electric dynamics when photo-stimulated with ChR2 (Klimas et al., 2020). The approach by Klimas et al. (2020) combined optogenetic stimulation of ChR2 with dual imaging of both, Ca²⁺ and V_m with either paired imaging of Rhod4/BeRST1, Rhod4/Di-4-ANBDQBS, or R-GECO1/BeRST1 (Klimas et al., 2020). This simultaneous imaging approach with well-separated excitation spectra using V_m and [Ca²⁺]_cyt probes, including the genetically encoded [Ca²⁺]_cyt sensor R-GECO1, showed fluorescence signals with minimal crosstalk. It demonstrated the applicability of the platform for multiparameter evaluation of chemo-electric signaling in neuronal networks, a benchmark for all-optical physiology. Whether the Di-4-ANBDQBS dye with excitation/emission peaks of 560/760–770 nm (Figure 4) can be used in plant cells to reveal membrane potential changes optically remains to be shown.

In contrast to the application of dyes, the use of genetically encoded sensors is more labor-intensive, possibly associated with a poorer SNR, but they can be co-expressed with the optogenetic tools. GECIs and GEVIs are being developed constantly and rapidly, especially in neuroscience. Some basic features to be considered when using GECIs are the photophysical properties of the fluorescent sensor proteins and their effective calcium-binding kinetics (see Supplemental Table S1).

An unfavorable issue raised with the red-shifted GECI R-GECO1 is a blue-light-dependent photoswitching effect based on photophysical properties that may generate artifacts when combined with optogenetics (Akerboom et al., 2013). A transient photo-switching effect by blue light excitation was described to result in a ∼50% higher R-GECO1 fluorescence for up to 900 ms (Akerboom et al., 2013). This issue, however, only needs to be considered in Ca²⁺-imaging faster than 1 Hz when blue light is simultaneously applied, for instance in confocal laser-scanning microscopy. Plant [Ca²⁺]_cyt dynamics are usually much slower than e.g. in neurons, and thus these R-GECO1 photoswitching effects are less acute to consider, especially if imaging experiments are designed sequentially (Li et al., 2021). The jRCaMP1a and jRCaMP1b Ca²⁺-sensors (originating from mRuby) do not show photoswitching after illumination with blue light, as R-GECO1 and jRGECO1 (based on mApple) do (Molina et al., 2019). Since some of the common optogenetic tools (e.g. ChR2) are triggered by blue light, the properties of jRCaMP1a/b makes them good candidates for combined [Ca²⁺]_cyt imaging and optogenetics.

In principle, optogenetic stimulation and [Ca²⁺]_cyt imaging with minimized crosstalk is possible with: (1) blue excitation/green emission GECIs of the GCaMP series (Chen et al., 2013; Lin and Schnitzer, 2016) and red-shifted optogenetic tools with spectral peaks near and above 600 nm like VChR1 (Zhang et al., 2008), C1V1 (Yizhar et al., 2011), ReaChR (Lin et al., 2013), Chrimson (Klapoetke et al., 2014), and Jaws (Chuong et al., 2014); or (2) green excitation/red emission with red-shifted GECIs such as RCaMPs, jRGECO1a (Dana et al., 2016), KGECO1 (Shen et al., 2018), or NIR-GECOs (Qian et al., 2019) and rhodopsins with blue excitation spectra like CheRiff (Hochbaum et al., 2014) or PsChR D139H (Duan et al., 2019; Figure 4); or (c) highly sensitive GECIs combined with less sensitive optogenetic tools, even sharing similar spectra, while low light power is used for imaging.

Though, it should be noted that the combination of green/blue GECIs (like GCaMP6s) with red-shifted rhodopsins is still proving difficult in practice because all published red-shifted rhodpsins show some degree of absorption in the blue spectral range, in which, for example, the GCaMPs are excited (Chuong et al., 2014; Stamatakis et al., 2018). This limitation could be partly overcome using multi-photon excitation of GCaMP6s, an “ultrasensitive” GECI, and C1V1- or ChR2-H134R based manipulation with negligible optical cross-talk (Packer et al., 2015; Baker et al., 2016). One of the first convincing experiments on the simultaneous use of red-shifted GECIs together with ChR2 was reported in 2013 (Akerboom et al., 2013). This work not only showed the limitations of R-GECO1 but also introduced the red GECI series called RCaMPs, albeit having a lower brightness exhibiting no photoswitching effect, which is favorable during simultaneous ChR2 applications (Akerboom et al., 2013). Further development of intensiometric red GECIs revealed the utility of jRCaMP1a and NIR-GECO1 (Dana et al., 2016; Qian et al., 2019), each paired with ChR2 (Zhang et al., 2007), to monitor [Ca²⁺]_cyt dynamics during optogenetics stimulation (Mao et al., 2019). Using two-photon excitation, the combination of jRCaMP1a [Ca²⁺]_cyt imaging with either ChR2- or ACR2 was successfully demonstrated as well in vivo (Forli et al., 2018). The chromophore from the intensiometric NIR-GECO1 probe originates from a bacterial phytochrome, causing the fluorescence brightness to be rather dim. However, a recent work accomplished to generate a fully NIR FRET-based ratiometric Ca²⁺ sensor termed iGECI, utilizing fluorescence proteins harboring bacterial phytochrome chromophores as well (Shemetov et al., 2021). Bacterial phytochromes use biliverdin as a cofactor, which should be readily available in plants as it is a by-product of heme catabolism to produce phytochromobilin, the chromophore of light-sensing phytochromes (Gohya et al., 2006; Dammeyer and Frankenberg-Dinkel, 2008). Thus, the ratiometric iGECI Ca²⁺-sensor is likely a suitable candidate to combine with spectrally separated optogenetic tools; its functional expression in plants, however, remains to be shown.

In neurophysiology, the readout of [Ca²⁺]_cyt signals represents an indirect readout of nerve cell activity. In order to record neuronal electrical activity, GEVIs are preferably used after their successful optimization (Knöpfel and Song, 2019). The rhodopsin-based toolbox not only contains proteins to manipulate cellular processes but also tools to monitor cellular events (Figure 1;Supplemental Table S1). These rhodopsin-based sensor tools for V_m sensing fall into the class of GEVIs. The slightly blue-shifted ChR2 version CheRiff was combined with the GEVIs QuasAr1 or QuasAr2 that have an excitation peak in the orange-red spectrum (580–590 nm; Figure 4) for simultaneous blue-light-driven ChR-based manipulation and optical voltage readout (Hochbaum et al., 2014). The so called Optopatch1/2 vectors harbored the coding sequence of CheRiff and QuasAr1/2, separated by a P2A self-cleavage site, enabling the co-expression of optogenetic actuator and fluorescence voltage sensor (Hochbaum et al., 2014). Further optimization of this Optopatch vector scaffold resulted in Optopatch3 and 4 constructs using QuasAr3 and Archon, respectively, showing higher expression, better trafficking in vivo, faster kinetics, and better SNR (Adam et al., 2019; Fan et al., 2020).

Optogenetic inhibition and voltage imaging with i-Optopatch used stGtACR2, a blue-shifted anion ChR together with SomArchon (Fan et al., 2020). Both, CheRiff and stGtACR2 showed negligible activation at 639 nm excitation wavelength used to excite the Arch-derived GEVIs (Hochbaum et al., 2014; Fan et al., 2020). A non-opsin-based GEVI called FlicR1, consisting of an integral membrane protein with a Ciona intestinalis voltage-sensing domain attached to the red fluorescent mApple protein that is compatible with blue-light-excitable ChRs and was shown to report action potentials in neurons (Abdelfattah et al., 2016). Unfortunately, FlicR1 and other available nonrhodopsin-based voltage sensors such as the ASAP series (St-Pierre et al., 2014) or GEVIs like VSFPs/Butterflies (Knöpfel and Song, 2019) show little fluorescence changes (ΔF/F) upon physiological V_m changes and off-targeting that decreases the SNR (Villette et al., 2019; Milosevic et al., 2020). To date, the application of GEVIs has not entered plant science, neither with rhodopsin-based nor with nonrhodopsin-based GEVIs. Whether this will change, once the problems of nonspecific targeting of GEVIs and their low fluorescence change are resolved, remains to be seen.

Originally developed by neuroscientists but little used in vivo is a red-fluorescent pH sensor, called pHuji, which could potentially be very interesting for plant research when combined with optogenetics. In contrast to animal cells, plant cells have a strong inward H⁺ gradient of almost 2 pH units and pH signals play an important role in plant physiology (Felle, 2001; Li et al., 2021; Tsai and Schmidt, 2021). The spectral properties of pHuji (Figure 4) are compatible with the use of blue-engineered optotools such as CheRiff or ChR2.

Limitations through plant photoreceptors and chlorophyll fluorescence

In contrast to many animal cell systems, pollen tubes, or plant roots, which have a high degree of transparency, green plant tissues harbor chlorophylls with strong autofluorescence in the red spectrum (Pedrós et al., 2008). This complicates the use of sensor fluorescent proteins with emission in the red spectrum between ∼650 and 750 nm, as it prevents to distinguish between sensor- and chlorophyll fluorescence. Nevertheless, with regard to the simultaneous use of sensors and optogenetic tools, several possibilities remain. The orange/red fluorescent R-GECO1, for example, has already found successful application in Ca²⁺-imaging, even in “green” tissue and its fluorescence can be separated from the chlorophyll fluorescence (Keinath et al., 2015; Waadt et al., 2017; Li et al., 2021). From a spectral point of view both R-GECO1 and jRCaMP1a are suitable for simultaneous use with optotools that can be excited in the blue and possibly in the green spectrum. Potentially, other red-fluorescent Ca²⁺ sensors with distinct photophysical properties such as CAR-GECO1, K-GECO1 or (j)REX-GECO (Molina et al., 2019) are also suitable for such a combined approach (see Supplemental Table S1), but unfortunately no data are available on their performance in plant cells. The NIR-GECO1 and poRhoVR 14 fluorescence spectra mostly overlay with the chlorophyll fluorescence spectrum (Figure 4) are therefore unlikely to be used for Ca²⁺ or voltage monitoring in photosynthetic tissue, respectively. In contrast, probes with further red-shifted emission spectra from sensors such as QuasAr, di-4-ANBDQBS, or Archon are again compatible for simultaneous detection of membrane voltage and use of blue- or green absorbing actuators (Figure 4).

In animal research, wavelengths >800 nm are used for the excitation of fluorophores by multiphoton microscopy, also for the simultaneous excitation of optotools and sensors. Light of higher wavelengths has the advantage of penetrating deeper into the tissue and scattering less. However, these wavelengths cause strong autofluorescence in chloroplasts (Breunig et al., 2013), which must be taken into account when using fluorescence-based sensors in photosynthetically active tissue.

The use of multiphoton excitation in the NIR or excitation with light in the visual spectrum involves the risk of addressing existing plant photoreceptors. It is known that the excitation of phototropins with blue light can lead to electrical and chemical signals (Stoelzle et al., 2003; Harada and Shimazaki, 2008; Babla et al., 2019) that may be similar to those generated via light-gated ion channels such as ChRs. Optogenetics in plants therefore presupposes that one has to consider corresponding control plants as a reference for monitoring electric and chemical signals.