Channelrhodopsin-1: A Light-Gated Proton Channel in Green Algae

See allHide authors and affiliations

Science  28 Jun 2002:
Vol. 296, Issue 5577, pp. 2395-2398
DOI: 10.1126/science.1072068


Phototaxis and photophobic responses of green algae are mediated by rhodopsins with microbial-type chromophores. We report a complementary DNA sequence in the green alga Chlamydomonas reinhardtiithat encodes a microbial opsin-related protein, which we term Channelopsin-1. The hydrophobic core region of the protein shows homology to the light-activated proton pump bacteriorhodopsin. Expression of Channelopsin-1, or only the hydrophobic core, inXenopus laevis oocytes in the presence of all-trans retinal produces a light-gated conductance that shows characteristics of a channel selectively permeable for protons. We suggest that Channelrhodopsins are involved in phototaxis of green algae.

The earliest detectable rhodopsin-mediated responses in phototactic green algae are electrical currents within the eye spot. These currents are carried by Ca2+ and H+ (1, 2). InChlamydomonas reinhardtii or Haematococcus pluvialis, photoreceptor currents occur with a delay of <30 μs after flash stimulation (3, 4), suggesting an intimate link between photoreceptor and electrical conductance. The two dominant retinal-binding proteins in the eye spot, Cop1 and Cop2, are not the photoreceptors that mediate photomovement responses (5). Searching a C. reinhardtii EST (expressed sequence tag) database (6) revealed overlapping cDNA sequences that encode a 76.4-kD opsin-related protein. We have named this protein Channelopsin-1 (Chop1). The core region (amino acid residues 76 to 309 out of 712 total) comprises seven hypothetical transmembrane segments with sequence similarity (15 to 20%) to the archaeal sensory rhodopsins (SRs), the ion transporters bacteriorhodopsin (BR) and halorhodopsin (HR), and rhodopsin fromNeurospora (nop1) (7). Although the overall sequence homology is low, several amino acids are conserved that define the retinal-binding site and the H+-transporting network in BR (8, 9) (Fig. 1 and Supporting Online Material). The consensus motif LDxxxKxxW (10) suggests that K296 is the retinal-binding amino acid. Thirteen of the 22 amino acids that are in contact with the retinal in BR are identical (9 amino acids) or conservatively exchanged (4 amino acids) in Chop1. The conserved residues are close to the protonated Schiff-base site in the archaeal rhodopsins.

Figure 1

Comparison of the truncated Chop1 (GenBank accession no. AF385748) sequence (10), Chop1-346 (encoding amino acids 1 to 346), with bacterioopsin (Bop) from Halobacterium salinarum. Amino acids that are known from the bacteriorhodopsin (BR) structure to interact directly with the retinal (8, 9) are indicated with an asterisk (*). Amino acids that are identical in most microbial opsins are highlighted in green; those that are functionally homologous in microbial opsin sequences are in yellow; and those that are identical in Chop1, BR, and only some other microbial opsins are in blue. Amino acids that contribute to the H+-conducting network in BR (8, 9) are shown in red. Residues that are part of the transmembrane H+-network are in bold type. The key substitutions, D85 and D96 in BR to E162 and H173 in Chop1, are shown in red. Underlined amino acids indicate identified or hypothetical transmembrane helices. Amino acids labeled in blue are leader sequences. The numbering of bacterioopsin begins after the leader sequence for historical reasons. The hypothetical retinal-binding Lys is labeled with “#”. For further details, see supporting online material.

We expressed cRNA encoding Chop1 in oocytes of Xenopus laevis, in the presence of all-trans retinal, to study ion transport under voltage-clamp conditions, as demonstrated for BR and sensory rhodopsin II (11–13). Illumination by green light, but not red light, induced inward currents inChop1 RNA–injected oocytes at a membrane potential of –100 mV (Fig. 2A). Similar results were obtained with truncated Chop1 RNAs encoding amino acids 1 to 346 or 1 to 517. At an external pH (pHo) of 7.5, the inward current induced by green light reversed at a voltage near –15 mV (Fig. 2B) with clearly visible outward photocurrents at positive membrane potentials. The dependence of photocurrent direction on the applied potential suggests that the reconstituted Channelrhodopsin-1 (ChR1) mediates a light-induced passive ion conductance. A reversal potential of –15 mV is close to the Nernst potential for Cl [–20 mV, as deduced by expressing CFTR (cystic fibrosis transmembrane conductance regulator) Clchannels in oocytes (14)] or H+ [–12 mV, at an intracellular pH (pHi) of 7.3, as measured with microelectrodes by us and others (15)], but far from Nernst potentials for Na+, K+, or Ca2+.

Figure 2

Light and voltage dependence of photocurrents at pH = 7.5. (A) Photocurrents recorded during illumination of oocytes with green or red light (500 ± 25 nm or 700 ± 25 nm, respectively; 1022 photons m−2 s−1). Membrane potential (V) = –100 mV; the light pulse is indicated by a bar. Bath solution: 96 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 5 mM MOPS (pH = 7.5). Trace A: A noninjected oocyte, green light. Trace B: A Chop1 oocyte, green light. Trace C: The same Chop1 oocyte as in trace B, irradiated with red light. (B) Current responses of a Chop1 oocyte to voltage steps from –100 mV to +40 mV (in 20-mV steps; holding potentialV h = –40 mV), followed by green light pulses of 200-ms duration. Bath solution: 96 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 5 mM MOPS (pH = 7.5).

To investigate the ionic specificity of the light-induced permeability, we systematically changed the composition of the bath solution. Lowering the pHo to 6.0 increased inward photocurrents and shifted the reversal potential to >+40 mV (Fig. 3, A and B). Replacing Clby aspartate (pHo = 6) had no discernible effect on the photocurrent amplitude or its current-voltage (I-V) relation (Fig. 3, A and B), thus excluding Cl as the conducted ion. Similarly, Na+ and Ca2+ were excluded: Photocurrents were not changed either by replacing Na+ byN-methyl-d-glucamine (NMG) or by replacing Ca2+ with Mg2+ (Fig. 3, A and B). However, increasing the H+ concentration of the bath solution, [H+]o, to 10 μM (pHo = 5) enhanced the light-induced inward currents, and changing [H+]o to 100 μM (pHo = 4) further increased the photocurrents (Fig. 3, A to C). Figure 3C demonstrates the magnitude of light-induced H+ currents at pHo = 4 in an oocyte where light-induced H+ conductance exceeds the background conductance of the oocyte by more than fivefold. Lowering pHo shifts the reversal potential of the photocurrent to more positive values (Fig. 3B). Taken together, these results indicate that the photocurrent is carried by protons.

Figure 3

Dependence of photocurrent (I p) on ionic conditions, pHo, pHi, and voltage. (A) Results from one (out of five) characteristic Chop1 oocyte, plotted in the order of measurement (∼150-s interval); V = –100 mV, green light as inFig. 2A. Solutions are buffered with 5 mM MOPS (pH = 7.5), MES (pH = 6), or citrate (pH = 5 and 4). Concentration (in mM): bar 1: 100 NaCl, 2 CaCl2(pH = 7.5); bar 2: 100 NaCl, 2 CaCl2(pH = 6.0) (reference condition); bar 3: 100 Na-aspartate, 2 CaCl2 (pH = 6.0); bar 4: 100 NMG-Cl, 2 CaCl2 (pH = 6.0); bar 5: same as bar 2; bar 6: 100 NaCl, 2 EGTA, 2 MgCl2 (pH = 6.0); bar 7: 200 sorbitol, 5 EGTA (pH = 5.0); bar 8: 200 sorbitol, 5 EGTA (pH = 4.0). (B) Voltage dependence of photocurrents (I p) from (A). Concentrations as in (A). (C) Current responses to voltage jumps from −100 to +40 mV (in 20-mV steps,V h = −40 mV), followed by green light pulses of 200-ms duration. Bath solution: 200 mM sorbitol, 5 mM citrate (pH = 4.0) with NMG (same oocyte as in Fig. 2B). (D) Voltage dependence of photocurrent (I p) in different bath solutions, containing 100 μM undissociated butyric acid: (▪) 60 mM NaCl + 40 mM Na-butyrate (pH = 7.4); (▴) 84 mM NaCl + 16 mM Na-butyrate (pH = 7.0); (▾) 93.6 mM NaCl + 6.4 mM Na-butyrate (pH = 6.6); (⧫) 97.4 mM NaCl + 2.6 mM Na-butyrate (pH = 6.2) (n = 5). (E) (○) pHo dependence of reversal potentials, derived from (D). The lines show the theoretical relation for a proton-selective conductance (−58 mV/pH) at a constant pHiof 6.6 (dotted) or 6.8 (solid). (▪) pHo dependence of reversal potentials, derived from experiments (n = 6) with 5.4 mM undissociated acetic acid, as demonstrated in (F). The line shows the theoretical relation for a constant pHi of 5.5 and −58 mV/pH. (F) Current responses to voltage steps fromV = −100 mV to +40 mV, followed by green light pulses. Bath solution: 100 mM Na-acetate, 5 mM MES (pH = 6.0) with NaOH.

The reversal potential of the photocurrent should be shifted to more negative potentials by increasing the cytosolic [H+]. At pHo = 7.4, a concentration of 40 mM Na-butyrate, which corresponds to 100 μM undissociated, membrane-permeable butyric acid, has been reported to increase the cytosolic [H+] of an oocyte from 50 nM (pHi = 7.3) to 160 nM (pHi = 6.8) (15). We confirmed by pH measurements with ion-selective microelectrodes at pHo = 7.4 that 40 mM butyrate caused a 0.5-unit decrease in pHi. At pHo = 7.4, 100 μM butyric acid caused a shift of the photocurrent reversal potential from –5 ± 2 mV to –33 ± 2 mV (n = 5) (Fig. 3D). Photocurrents were measured at four pHo values—7.4, 7.0, 6.6, and 6.2—with the undissociated butyric acid concentration of 100 μM to maintain the pHi at ∼6.8 (15). The direction of photocurrent reversed at a different potential for each pHo (Fig. 3D). The dependence of reversal potential (V rev) on pHo is shown in Fig. 3E (open circles). The V rev of –33 ± 2 mV at pHo = 7.4 and pHi = 6.8 (15) is in close agreement with the H+ V rev of –35 mV at ΔpH = 0.6. This close fit shows that the light-activated currents are passive. In addition, the dependence of photocurrent V rev on pHo indicates a high selectivity for protons. A small change of pHi with changing pHo, as reported previously for oocytes (15), can explain the deviation from the ideal slope of –58 mV/pH unit. Indeed, a shift of pHifrom 6.8 to 6.6 (as shown by the two lines with ideal slope in Fig. 4C) would bring the observed dependence in accordance with theoretical expectation for a photoconductance that is selectively permeable for H+. When a high concentration (5.4 mM) of undissociated acetic acid (pHo = 5 to 6) was administered, large outward photocurrents at positive potentials (Fig. 3F) and an ideally pHo-dependentV rev with a slope of –58 mV/pH unit were observed (Fig. 3E, filled circles). The V revobserved in these experiments suggests that application of 5.4 mM acetic acid lowers the pHi to ∼5.5 (Fig. 3E). We confirmed this value using microelectrodes to measure the pHi of oocytes in the presence of 5.4 mM acetic acid. Thus, the observed dependence of V rev on pHo implies that the ChR1-mediated light-sensitive conductance is passive and highly selective for protons.

Figure 4

Dependence of photocurrent on temperature, wavelength, and amino acid at position 173. (A) Current at –100 mV (voltage steps from –40 mV) and pHo= 4 at 7°, 19°, and 30°C (same oocyte). Decay of photocurrent was fitted with a biexponential function at all three temperatures: 7°C: τf = 115 ms, τs = 810 ms; 19°C: τf = 35 ms, τs = 150 ms; 30°C: τf = 19 ms, τs = 49 ms. (B) Arrhenius plot of inverse time constants [(▾)k s, (▴) k f] and photocurrent [(○) I p] from four experiments. Lines indicate an activation energy of 22 kJ/mol forI p and 60 kJ/mol for k sand k f. (C) Wavelength dependence of the light-induced inward current at pHo = 5.5 and –40 mV. The photocurrents were corrected for equal photon flux. (D) Photocurrents of mutants Chop1-H173R, Chop1-H173D, and Chop1-H173Y at –80 mV, pHo = 5.

The dependence of the direction of proton transport on the electrochemical potential for H+ is compatible with either facilitated diffusion of a stoichiometric number of protons during each ChR1 photocycle or a light-gated proton channel in which an intermediate of the ChR1 photocycle allows permeation of many H+ ions. However, further analysis supported the model of a light-gated channel. The photocurrent decays with two time constants after illumination, similar to bacteriorhodopsin (16) (Fig. 4A). These time constants are strongly temperature sensitive, indicating an activation energy for the photocycle of ∼60 kJ/mol. The temperature sensitivity of the photocurrent amplitude is much smaller, ∼20 kJ/mol (Fig. 4, A and B). This suggests that intermediates of the photocycle serve as the proton-conducting channel. The maximal ChR1 photocurrent exceeds maximal BR photocurrents, providing an additional argument against a stoichiometric coupling of photocycle and proton transport. Large photocurrents of a strictly coupled system would require either a photocycle that is faster than that of BR, which is incompatible with the observed slow decay time constants, or a high expression level (∼100-fold greater than that of BR), which is unlikely. If we assume that the level of ChR1 expression in oocytes is similar to that of BR [∼2 × 109 molecules per oocyte (11)], then we can calculate the electrical current contribution of a single ChR1 molecule. At pHo = 4 and –100 mV, we measured a light-induced inward current of ∼1.4 μA, i.e., ∼4000 protons/s or 0.7 fA per molecule, which is much too small for single-channel resolution.

The wavelength dependence of the inward photocurrent was determined at pHo = 5.5 and –40 mV (Fig. 4C). The maximum wavelength of ∼500 nm closely resembles that observed in the action spectra for photoreceptor currents (1), phototaxis (17), and photoshock responses (18) of intact C. reinhardtii cells. The pHo-dependent photocurrent, I P2, recorded from intact C. reinhardtii cells (2) was found to dominate the stationary current in continuous light at low pHo(2). We speculate that I P2might be carried by ChR1.

During the BR photocycle, a proton is transferred from the Schiff base to Asp85 (corresponding to Glu162 in Chop1) and released to the surface by way of Arg82, Glu194, and Glu204. The corresponding residues in Chop1 are Arg161, Glu274, and Ser284. The key amino acid for the reprotonation of the retinal Schiff base in BR is Asp96 (8,19). The corresponding amino acid in Chop1 is His173(Fig. 1), which we exchanged with three different amino acids: Asp, Arg, or Tyr. The substitution His173 → Asp (H173D) resulted in a complete loss of light-gated conductance, whereas H173R or H173Y were still functional (Fig. 4D). These results indicate that His173 does not function as a proton donor of a deprotonated Schiff base. Therefore, we suggest that in ChR1 the retinal Schiff base is not deprotonated during the photocycle. This claim is corroborated by the observation that blue light, which is absorbed by deprotonated retinylidene, did not quench the stationary currents (20). It is conceivable that isomerization of retinal or a conformational change, tightly coupled to isomerization, gates the ChR1 proton channel.

In conclusion, ChR1 is an ion channel that opens in response to absorption of light, i.e., a combined photoreceptor and ion channel. It is not unlikely that such directly light-sensitive ion channels are widely distributed in other phototactic microalgae, as well as in gametes and zoospores of macroalgae, or even in fungi (e.g., nop1) (7). Moreover, the ability of ChR1 to mediate a large light-switched H+ conductance in oocytes holds promise for the use of ChR1 as a tool for measuring and/or manipulating electrical and proton gradients across cell membranes, simply by illumination.

  • * To whom correspondence should be addressed. E-mail: nagel{at}


View Abstract

Navigate This Article