Unique Light-gated Ion Channel Properties of a Novel Modular Cation Channelrhodopsin from an Evolutionary Important Terrestrial Green Alga

Channelrhodopsins are a family of microbial rhodopsins that function as a light-gated ion channel. We report the molecular properties of a novel channelrhodopsin KnRh3 from an evolutionary important lamentous terrestrial alga Klebsormidium nitens. KnRh3 is constituted of a 7-transmembrane domain, followed by a long C-terminus moiety that encodes a peptidoglycan binding domain (FimV). When functionally expressed in mammalian cells, KnRh3 showed light-induced cation channel currents. The maximum action spectrum exhibited was at 430 nm and 460 nm, the former making KnRh3 one of the most blue-shifted channelrhodopsins characterized thus far. The channel closure rate was relatively fast (τ 0ff = 10 ms). Surprisingly, photocurrent kinetics were affected by the C-terminus moiety of KnRh3. When this moiety was truncated to various lengths, this prolonged the channel open lifetime by more than 10-fold. We identied two arginine residues, R287 and R291, those are crucial for altering the kinetics. We propose that electrostatic interaction between the 7-TM domain and the C-terminus domain accelerates the photocycle. The most blue-shifted action spectrum of KnRh3 serves as a novel prototype of channelrhodopsin for studying the molecular mechanism of color tuning. In addition, KnRh3 would expand the optogenetics tool kit, especially for when short wavelength excitation is required.


Introduction
Microbial-type rhodopsins are made up of 7-transmembrane helices with a covalently bound all-trans retinal as a chromophore 1 . These are found in archaea, bacteria, eukaryotes (such as fungi and algae) and viruses and are physiologically responsible for energy production and phototaxis responses.
Molecular functions of microbial rhodopsins involve ion transporters, sensors and light-regulated enzymes 2,3 . Ion-transporting rhodopsins are divided into pumps and channels. Bacteriorhodopsin was the rst identi ed outward directed proton-pumping rhodopsin 4 . The Cl − pump, the Na + pump, and inwarddirected light-gated proton pumps have also been characterized [5][6][7][8] . Detailed structure-function studies together with electrophysiology and molecular dynamics studies revealed the detailed molecular mechanisms of bacteriorhodopsin and other ion-pumping rhodopsins. Channelrhodopsin-1 and − 2 (CrChR1 and CrChR2) from Chlamydomonas reinhardtii were the rst light-gated ion channels to be discovered and characterized 9,10 . These proteins permeate cations in which the permeability ratio of H + , Na + and K + is 10 6 , 1, and 0.5, respectively. High-resolution X-ray structures revealed details of the molecular architecture and provided insight into the photoactivation and ion conduction pathway 11,12 .
Successful expression of CrChR2 in neurons allowed the action potential to be manipulated by light, which initiated a new research eld, optogenetics 13,14 . Channelrhodopsin variants have been engineered to improve the functionality of ChR, and homologous ChRs were also reported accordingly 15 . Color-tuning variants of ChRs cover almost the entire visible range of the light spectrum. CrChR2 from C. reinhardtii exhibits an action spectrum maximum at 470 nm 10 . ChR variants such as C1V1, which is the chimeric version of ChR1 from C. reinhardtii and Volvox carteri, or C1C2 (green light receiver) absorb around 530 ~ 545 nm light 16,17 . A further red-shifted ChR, Chrimson from Chlamydomonas noctigama, exhibits an Page 3/23 absorption maximum at 590 nm which allows reliable neuronal stimulation over 600 nm light when applied to neurons 18 . On the other hand, TsChR or PsChR absorb shorter wavelengths, making it possible to excite neurons at 440 nm 19 . The open channel lifetime can be extended by mutations at C128 and D156 (DC pair) which form a hydrogen bond bridge in CrChR2. A mutation at C128 to T, A and S slowed channel closure kinetics by 1000-fold 20 . The CrChR2 D156C has an even stronger effect, with high light sensitivity and a prolonged open channel lifetime, up to 30 min 21 . Recent genomic database research reveals that various lower eukaryotes contain opsin-related genes. Consequently, a number of ChR genes were identi ed and their molecular properties, such as ion selectivity and absorption maxima, were shown to be different.
Recently, the genome of an evolutionary important terrestrial green alga (Klebsormidium accidum) was sequenced, and analysis of the genome delineated the molecular basis of the adaptation of land plants 28 . K. accidum was renamed as Klebsormidium nitens (http://www.plantmorphogenesis.bio.titech.ac.jp/~algae_genome_project/klebsormidium/). K. nitens adapts to diverse environmental conditions such as desiccation, light, and temperature gradients 29 .
However, the sensory photoreceptor landscape of K. nitens has not yet been established. For the rst time, we identi ed a ChR homologue that encodes a novel modular channelrhodopsin from an evolutionary important lamentous terrestrial alga, K. nitens. A unique feature of KnRh3 is that it includes the FimV domain at the C-terminus of rhodopsin. FimV is involved in peptidoglycan-binding protein which is responsible for twitching motility in Pseudomonas aeruginosa and other microbes 30 . Thus, it is anticipated that light signals are transduced via ChR into motility in K. nitens.
In this report, we performed an electrophysiological experiment using KnRh3 to reveal its ion channel function. Its photocurrent properties such as channel kinetics, absorption maximum, light sensitivity and ion selectivity was compared to those of CrChR2. For the rst time, we show an unexpected modulation of the photocurrents of KnRh3 by its extended C-terminus region. Novel KnRh3 has the potential to expand the optogenetics tool kit, especially for dual light (two photon) applications when short wavelength excitation is required. Moreover, functional characterization of channelrhodopsin from an evolutionary terrestrial alga indicates the existence of channelrhodopsin-mediated physiological responses beyond aquatic habitats (marine and freshwater systems).

Mammalian cell culture
Electrophysiological assays and immunostaining of KnRh3 and CrChR2 were performed on ND7/23 cells, which are hybrid cell lines derived from neonatal rat dorsal root ganglia neurons fused with mouse neuroblastoma. ND7/23 cells were grown on a coverslip in Dulbecco's modi ed Eagle's medium (DMEM; Wako, Osaka, Japan) supplemented with 2.0 mM of all-trans retinal and 5% fetal bovine serum, and under a 5% CO 2 atmosphere at 37°C. The expression plasmids were transiently transfected by using the  Table 1. The liquid junction potential was calculated and compensated by pClamp 10.7 software. Time constants were determined by a single exponential t unless indicated otherwise. Table 1 Measurment solution component for ion selectivity Optics For whole-cell voltage clamp, irradiation at 480 nm was carried out using collimated LED (parts No. LCS-0470-03-22, Mightex, Toronto, Canada) controlled by computer software (pCLAMP10.7, Molecular Devices). Light power was measured directly by an objective lens of a microscope by a power meter (LP1, Sanwa Electric Instruments Co., Ltd., Tokyo, Japan). All action spectra were measured at the same light intensity in the range of 410 nm to 650 nm by a xenon light source OSG (Hamamatsu photonics, Hamamatsu, Japan).

Statistical validation of the data
Data in the text and gures are expressed as mean ± SEM and were evaluated with the Mann-Whitney U test for statistical signi cance, unless noted otherwise. Means were judged as statistically insigni cant when P > 0.05.

Genome of terrestrial alga encodes a novel modular channelrhodopsin (KnRh3)
Extensive and targeted mining of the genomic database of K. nitens revealed the existence of several rhodopsin-encoding genes (http://www.plantmorphogenesis.bio.titech.ac.jp /cgibin/ blast/blast_www_klebsormidium.cgi). Among them, we found one gene that is highly homologous to chlorophyte cation channelrhodopsins such as CrChR2. We named it KnRh3 in this study. KnRh3 possesses 831 amino acids and harbors a membrane-embedded 7-TM rhodopsin domain (amino acids 1-272) followed by a long C-terminus region (273-831 amino acids) (Fig. S1). Sequence comparison of the rhodopsin domain revealed 34.5% identity with and 69.3% homology to CrChR2. Several important amino acid residues were aligned when compared to four ChRs and bacteriorhodopsin (BR) (Fig. 1A). Three (Glu82, Glu90, Glu97) out of ve glutamate residues at positions 82, 83, 90, 97 and 101 in CrChR2 that are important for channel activity were conserved in KnRh3. Glu83 was replaced by Val. Glu101 was substituted by Pro in KnRh3 whereas in GtACR1 (light-gated anion channel), Gln was observed at this position. His instead of Asp at position 96 was regarded as a characteristic of channelrhodopsins, but in KnRh3, this position was occupied by Ala, by D156 in GtACR1, while C128 in CrChR2 formed a hydrogen bridge (D-C pair or D-C gate) which altered the channel open lifetime 20,21 . This pair is also conserved in KnRh3 whereas chlorophyte cation channel rhodopsin (GtCCR4) lacks it. Thus, we predicted that KnRh3 might function as a cation channelrhodopsin when functionally expressed in a suitable system. Notably, KnRh3 has a long C-term domain that encodes a putative peptidoglycan binding domain (FimV) between 410-690 amino acids (Fig. S1). In general, the peptidoglycan layer is located outside of the cytoplasmic membrane in bacterial cell walls. In fact, the homologous region of KnRh3 is predicted to be located on the extracellular side, whereas the C-terminus region is normally located on the cytoplasmic side in the case of microbial-type rhodopsin (Fig. S2). Thus, we considered why the FimV domain is connected to the C-terminus of the 7-TM domain. We anticipated that the membrane topology of KnRh3 is inverted, as in the case of heliorhodopsin (HeR), which is a recently identi ed subfamily in microbial rhodopsins 31 . We synthesized a codon-adapted full-length gene of KnRh3 (831 amino acids) for functional expression in a mammalian system and fused it with the p3.0-eYFP vector with eYFP at the C-terminus of KnRh3. After transfection of the fusion expression construct, we observed weak YFP uorescence on the cellular membrane, indicating poor expression in ND7/23 cells. A voltage-clamp measurement was then performed. However, no photocurrent was observed (data not shown). We anticipated that low expression was due to a long C-terminus domain (~ 540 amino acids) after the 7-TM domain. Thus, we systematically truncated the C-terminus to various lengths. We created ve variants with different lengths, carrying 697, 397, 310, 290, and 272 amino acids. Note that only the variant with 697 amino acids included the FimV domain. In addition, we fused a cMyc epitope tag at the C-terminus of the 697 amino acids variant to detect immune uorescence. Figure 1B-D shows uorescent images of the transfected cells. eYFP uorescence indicates successful expression of KnRh3. Although no alexa595 signal was observed in the absence of detergent (Fig. 1B), treatment with detergent to permeabilize the membrane generated a signal (Fig. 1C). No alexa595 signal was observed when KnRh3 without cMyc-tag was expressed even after the detergent treatment (Fig. 1D). These results indicate that the C-terminus domain is on the cytoplasmic side. Thus, KnRh3 exhibits typical membrane topology of a channelrhodopsin.
The C-terminus of KnRh3 modulates photocurrents and photocycle of the channelrhodopsin domain We then performed patch clamp recordings of each variant. All the constructs having various amino acid lengths are depicted in Fig. 2A. As shown in Fig. 2B-F, large photocurrents were observed in all the ve KnRh3 variants upon blue light illumination for 1.0 second. Photo-current amplitudes were comparable to that of CrChR2 (Fig. 2G). Current direction reversed as clamped voltage shifted from − 90 to + 50 mV in 20 mV steps, which is a typical ion channel property. Illumination induced a large peak current (Ip) which slowly decayed into smaller amplitudes during illumination. However, the current did not reach a plateau within 1 sec, whereas the photocurrent of CrChR2 rapidly dropped to a steady state level (Fig. 2G). When light was shut off, this diminished the photocurrent into the baseline, indicating channel closure. Notably, the phases of current decay were signi cantly different among the ve variants, two of which, those with 697 and 397 amino acids, rapidly decayed into the baseline whereas the other three (300, 290 and 272 amino acids) showed a relatively slow decay (Fig. 2B-F). This indicates that the channel kinetics and the photocycle are affected by the cytoplasmic domain of KnRh3. To further investigate the observation above, we created in addition four variants with different lengths of amino acids: 321, 317, 310 and 280.
Current decay kinetics (τ-off) at -70 mV from all nine variants were compared (Fig. 2H). The variants longer than 317 amino acids showed a fast off-kinetics (~ 10 ms). The variant with 310 amino acids showed an off-kinetics of about 20 ms. Furthermore, the time constants of the shorter variants became even larger and the shortest variant with 272 amino acids reached 130 ms, more than 10 times slower than the original construct. This result clearly shows a correlation between the length of amino acids and the time constant of channel kinetics in KnRh3, indicating that the C-terminal cytoplasmic domain affects both channel kinetics and the photocycle. In particular, amino acids between 272 and 310 largely contribute to the modulation of light-gated ion channel activities in KnRh3.
Interestingly, the photocurrent amplitudes also differed among the nine variants (Fig. 2I). The variants shorter than 300 amino acids reached about 3 nA at -70 mV, while the longer variants > 310 amino acids showed smaller currents under the same condition in which only about 1 nA was observed from cells expressing the 397 and 697 amino acids variants. We reasoned that current amplitude depends on the open-life time of the channels. As shown in Fig. 2H, 32 . Thus, the apparent current amplitude of the short variants was elevated as the M-state accumulated during illumination.
The current-voltage relations (I-V plots) of KnRh3 and CrChR2 are depicted in Fig. 2J (peak component) and Fig. 2K (steady state component). Current direction and amplitude depend on the applied voltage. The reversal potential of KnRh3 is about − 10 mV while that of CrChR2 is 0 mV under the same condition. As is already known, ion conductance of CrChR2 is inwardly-recti ed and thus the outward current was largely suppressed (Fig. 2J and K, black). The I-V plots of KnRh3 show a rather weak recti cation and relatively large outward currents were observed (see + 50 mV in Fig. 2J and K). The I-V plots of all the variants are summarized in Fig. S3. The photocurrent properties of KnRh3 397 amino acids (excluding FimV) and 697 amino acids (including FimV) are shown in Fig. S4. The current shape, I-V relation, and current amplitudes of these two variants were identical. These results indicate that FimV did not alter channel function.
All nine variants that were tested were fused to eYFP at the C-terminal. Thus, one might argue whether the length-dependent effect in Fig. 2I might have originated from eYFP, i.e. if an interaction between rhodopsin and eYFP may have altered the channel kinetics for an unknown reason. To exclude this possibility, we created 272 amino acids and 397 amino acids KnRh3 variants without eYFP. The variants were transfected, and the photocurrents were recorded to compare the channel-off kinetics and amplitude (Fig. 3). The time constants (t-off) were almost identical between eYFP-tagged and untagged 272 amino acids and 397 amino acids variants (Fig. 3A). This clearly indicates that a length-dependent effect was observed, regardless of the presence of eYFP at the C-terminus of KnRh3. Photocurrent amplitudes of untagged variants were smaller than those of eYFP-tagged constructs (Fig. 3B). This could be due to a lower expression level of the former. In fact, eYFP constructs contain the ER-exporting signal peptide and the membrane tra cking signal peptide to improve membrane localization, whereas no additional sequence was tagged in the untagged variants.

Positively charged residues in the C-terminus domain of KnRh3 regulated intramolecular interactions
Based on the result in which the cytoplasmic domain altered the channel kinetics in KnRh3 (Fig. 2H), it could be hypothesized that the cytoplasmic domain interacts with the rhodopsin domain (7-TM domain).
Since no signi cant change in t-off was observed among variants longer that 317 amino acids, the region from 317 amino acids onwards is not important for the interaction. Thus, we anticipated that one or several amino acids in the cytoplasmic region between 272 and 317 amino acids would be crucial for the interaction. We noticed two characteristic negatively charged residues, E285 and E293, and three positively charged residues, R287, K289 and R291, some of which might form an electrostatic interaction with the 7-TM domain (Fig. 4A and Fig. S1). Thus, we replaced each residue with alanine in the KnRh3-397 amino acids variant which exhibits fast kinetics, similar to the longest variant (697 amino acids). As shown in Fig. 4B-D, photocurrent decay after switching off light apparently became slower after mutating R287A and R291A. The decay was further slowed in the double mutant R287A/R291A (Fig. 4E). Figure 4F summarizes the time constants of all the mutants tested. The R287A mutant shows a t-off of 50 ms which is ve-fold slower than that of the KnRh3-397 variant, while R291A was slowed down to 25 ms. The double mutant R287A/R291A reached about 100 ms which is more than 10 times slower than KnRh3-397 amino acids (no amino acid substitution) and is close to the 272 amino acids (no amino acid substitution) variant (130 ms). The three remaining mutants (E285A, K289A and E293A) showed no effect on kinetics. These results clearly indicate that the positively charged residue, R287 and R291, interact with the 7-TM domain and contribute to altered channel kinetics and the photocycle (Fig. 4I). These three mutants exhibited signi cantly larger photocurrents than KnRh3-397 (Fig. 4G) without any change in their reversal potential (Fig. 4H).

Action spectrum and light sensitivity of KnRh3
The action spectrum, a wavelength dependency of photocurrent amplitude, was measured with a multicolor light source (Fig. 5). The 272 amino acids and 397 amino acids variants showed a spectrum with two λ max at 430 and 460 nm which are shorter than that of CrChR2 (λ max = 470 nm). To our knowledge, the λ max of KnRh3 is one of the shortest among naturally occurring ChRs.
We then measured the light-power dependency of two KnRh3 variants (272 amino acids and 397 amino acids), comparing them to CrChR2 (Fig. 6). Photocurrent amplitude, including Ip (peak component) and Is  (Fig. 6A lled circle). On the other hand, light sensitivity of the 397 amino acids variant was markedly lower than that of the 272 amino acids varint. Both Ip and Is became saturated at about 3-4 mW/mm 2 (Fig. 6B). Note that the x-axis of each panel differs by a single order of magnitude. The lightpower dependency of CrChR2 falls in the same range as KnRh3 397 amino acids (Fig. 6C). The halfsaturation light intensity (EC 50 ) was determined to be 0.051 mW/mm 2 (KnRh3-272), 0.712 mW/mm 2 (KnRh3-397) and 0.731 mW/mm 2 (CrChR2). This indicates that the C-terminal length affected the light sensitivity of KnRh3.

KnRh3 permeates monovalent and divalent cations
We predicted that KnRh3 functions as a cation channelrhodopsin considering its sequence similarity to chlorophyte channelrhodopsins (e.g. CrChR2). Ion selectivity was tested by patch-clamp recording. We used the KnRh3 272 amino acids variant for all experiments. The extracellular solution was systematically replaced with various cations, while the intracellular solution was xed (refer to Table 1 for more details). Figure 7A-F shows representative photocurrent traces under various ionic conditions. In the presence of NMG, we assumed that H + was the conducting ion. The I-V plot under several ionic conditions is depicted in Fig. 7G (KnRh3) and Fig. 7H (CrChR2). Note that the peak component (Ip) of the photocurrent was plotted for all the relevant measurements, and essentially the same result was obtained from the analysis of Is, which is summarized in Fig. S5 and S6. When Fig. 7A and 7B are compared, voltage dependency of current direction was obviously different. The reversal potential was shifted from − 30 mV to + 20 mV (Fig. 7G red and blue), indicating H + conductance across KnRh3 (Fig. 7G, I, J and K).
Large Na + conductance was observed, as shown in Fig. 7C, with a reversal potential of + 20 mV (Fig. 7G green, J and K). The shift in reversal potential of the K + solution indicates that KnRh3 permeates K + (Fig. 7I, J and K Fig. S5). We then tested divalent cations, Ca 2+ and Mg 2+ . We observed an inward photocurrent in the Ca 2+ solution (Fig. 7E) and reversal potential shifted signi cantly from − 30 mV (NMG pH 9.0) to -5 mV (Ca 2+ pH 9.0) (Fig. 7G red and black, J, K). A similar observation was observed in the Mg 2+ solution (Fig. 7F, I, J and 7K, Fig. S5). Altogether, we demonstrated that KnRh3 is a cation channelrhodopsin that permeates monovalent and divalent cations. Figure 7H shows the I-V plot of Cr_ChR2 under the same condition as Fig. 7G. Reversal potentials and a shift in reversal potential (ΔErev) of KnRh3 and CrChR2 are depicted in Fig. 7J, 7K and 7L, respectively. These values allow the permeability ratio of each cation to be estimated. The ΔErev of H + , Na + , K + , Ca 2+ and Mg 2+ from CrChR2 were determined as 90, 70, 65, 40 and 35 mV, respectively (Fig. 7L). This indicates that the rank in permeability ratio is H + > Na + > K + > Ca 2+ > Mg 2+ . On the other hand, KnRh3 showed no signi cant difference in ΔErev among H + , Na + and K + (45 ~ 48 mV), indicating that permeability of these cations was almost the same (Fig. 7K). The ΔErev of Ca 2+ and Mg 2+ were about 30 and 25 mV, respectively, indicating lower permeability for divalent cations than monovalent cations (Fig. 7K). As mentioned above, we analyzed the peak component of each photocurrent. In addition, Is was analyzed. Essentially the same results were found (Fig. S6).

Discussion
In this study, we report on the electrophysiological characterization of a newly identi ed multi-domain cation channelrhodopsin from Klebsormidium nitens (KnRh3). KnRh3 possesses a large cytoplasmic domain made of about 540 amino acids which encode a peptidoglycan binding moiety (FimV). Several ChRs also have a long C-terminus tail. Full-length ChR1 and ChR2 from Chlamydomonas reinhardtii contain 712 and 737 amino acids in total (837 and 747 in Volvox carteri). Among them, the 7-TM rhodopsin domain is made up of only about 300 amino acids. Thus, 400-500 amino acids are located mainly on the cytoplasmic side. However, no signi cant homology was found among the long cytoplasmic extension of the characterized ChRs from these different organisms. Since the discovery of ChRs, the role of cytoplasmic domains remains barely understood. The initial reports of CrChR1 and CrChR2 described that truncation of the cytoplasmic C-terminal domain did not alter the ion channel function 9,10 . Thus, researchers investigated their molecular properties by using truncated proteins carrying only the 7-TM domain. However, an extensive study of the phosphoproteome of the eyespot fraction of C. reinhardtti showed three phosphorylation sites for CrChR1 and one for CrChrR2 33 . Thus, a kinase-dependent regulation of function was also proposed. Recently, it has been shown that phosphorylation of CrChR1 is connected to photomotility and calcium signaling of green alga 34 . Predicting the phosphorylation of KnRh3 suggested that several motifs are involved in phosphorylation (Fig. S7). It would be interesting to delineate the functional role of these predicted phosphorylation patterns of KnRh3 in a native system.
We demonstrated that the C-terminus region affects the light-gated ion channel function. The channel closure rate (off-kinetics) slowed down from 10 ms to 130 ms as the cytoplasmic domain was shortened (Fig. 2H). In other words, the C-terminus domain accelerates channel closure and the photocycle. Positively charged residues, R287 and R291, in particular, are crucial for the acceleration, because the double mutant R287A/R291A reduced the fast kinetics (Fig. 4F). Furthermore, light sensitivity was altered following systematic truncation of the C-terminus (Fig. 6A and B). Therefore, we propose that an electrostatic interaction exists between the residues of the C-terminus and the 7-TM domain of KnRh3 (Fig. 4I). Although we did not identify the counterparts of R287 and R291 in this study, the candidates could be negatively charged residues such as Asp and Glu in the TM domain, which should be investigated in the future. Modulation of photocycle kinetics by the cytoplasmic domain has never been proven in ion-transporting rhodopsin thus far. However, in the case of enzyme rhodopsin (HKR or 2C-cyclops), the full-length HKR has been shown to exhibit a normal photocycle which reverts to the initial dark state in 20 sec, whereas truncation of the enzyme domain at the C-terminus region resulted in production of a bi-stable photocycle in which rhodopsin possesses two dark states (400 nm and 380 nm), which exhibit photochromism 35,36 . Detailed functional characterization of KnRh3 with spectroscopy and a structural study together with mutation analysis would be needed to reveal the mechanism of the functional modulation by the cytoplasmic domain.
The action spectrum of KnRh3, assessed from a voltage clamp experiment, revealed its absorption maxima (λ max ) at 430 and 460 nm. To the best of our knowledge, this is one of the most blue-shifted λ max in naturally occurring ChRs. Govorunova et al. previously reported a blue-shifted ChR from Platymonias subcordiformus (PsChR) that displayed an action spectrum of λ max = 445 nm, and an absorption maximum of 437 nm after puri cation (in solution) 19 . They ascribed the blue-shift to amino acid residues G142 and A146 in 4-TM (T159 and G163 in CrChR2). In fact, replacement of these residues resulted in a blue shift when tested with C1C2 and iC + + 37,38 . KnRh3 conserves A146 (as A161) while G142 is replaced by A157. Thus, A161 would be one of the color determinants in KnRh3, which could be applied to optogenetics. Usually, for optogenetics tools, long-wavelength absorption is ideal for minimizing light scattering by biological tissues. No photocurrent was observed in KnRh3 after illumination with 550 nm light (Fig. 5). The blue-shifted spectrum of KnRh3 is ideal for combinatorial applications with long wavelength-absorbing channelrhodopsin or uorescent indicators to separate the stimulation or excitations.
Given its high homology to CrChR2, KnRh3 exhibits a non-selective cation channel, and permeating monovalent and divalent cations including H + , Na + , K + , Mg 2+ and Ca 2+ . The permeation ratio was similar to that of CrChR2. The cation channelrhodopsin from Cryptophyte algae (DTD channelrhodopsin) exhibited different ratio of cation permeation, in which monovalent metal cations were mainly transported with minimum H + and Ca 2+ permeation 39,40 . KnRh3 conducts Ca 2+ at a relatively high amount ( Fig. 7 and Fig. S6). There is the potential for engineering highly Ca 2+ permeable channels, similar to the calcium-permeable ChR2 variant (Catch) 22 . The fast off-kinetics of KnRh3 397 amino acids has great potential for optical neuron excitation with high frequency.