2023 Volume 71 Issue 2 Pages 154-164
Rhodopsins are transmembrane proteins with retinal chromophores that are involved in photo-energy conversion and photo-signal transduction in diverse organisms. In this study, we newly identified and characterized a rhodopsin from a thermophilic bacterium, Bellilinea sp. Recombinant Escherichia coli cells expressing the rhodopsin showed light-induced alkalization of the medium only in the presence of sodium ions (Na+), and the alkalization signal was enhanced by addition of a protonophore, indicating an outward Na+ pump function across the cellular membrane. Thus, we named the protein Bellilinea Na+-pumping rhodopsin, BeNaR. Of note, its Na+-pumping activity is significantly greater than that of the known Na+-pumping rhodopsin, KR2. We further characterized its photochemical properties as follows: (i) Visible spectroscopy and HPLC revealed that BeNaR has an absorption maximum at 524 nm with predominantly (>96%) the all-trans retinal conformer. (ii) Time-dependent thermal denaturation experiments revealed that BeNaR showed high thermal stability. (iii) The time-resolved flash-photolysis in the nanosecond to millisecond time domains revealed the presence of four kinetically distinctive photointermediates, K, L, M and O. (iv) Mutational analysis revealed that Asp101, which acts as a counterion, and Asp230 around the retinal were essential for the Na+-pumping activity. From the results, we propose a model for the outward Na+-pumping mechanism of BeNaR. The efficient Na+-pumping activity of BeNaR and its high stability make it a useful model both for ion transporters and optogenetics tools.
All organisms need energy to live on Earth, where sunlight has fueled them for 3.5 billion years. Indeed, light is one of the most important energy sources and signals for organisms. To capture light, organisms have produced and developed various photoreceptive proteins with their cognate chromophore molecules through evolution. Among the photoreceptive proteins, microbial rhodopsin consists of a seven-transmembrane domain that binds an aldehyde derivative of vitamin A, i.e., retinal, as a chromophore, which can capture visible light ranging from 400 to 600 nm1–3) (Fig. 1A). In the apoprotein called opsin, an all-trans retinal molecule is covalently bound to a conserved lysine residue through a protonated retinal Schiff base (PRSB) linkage stabilized by a negatively charged carboxylate as a counterion1–3) (Fig. 1A). It is well-known that microbial rhodopsins play central roles in photoreception in a variety of microorganisms through photo-energy conversion and photo-signal transduction.1–3) Light absorption by the rhodopsin causes isomerization of the retinal chromophore from the all-trans form to the 13-cis form, and the stored energy of the excited retinal chromophore induces sequential conformational changes of the protein moiety.1–3) This sequential photoreaction, called “photocycle,” is approximately completed within milliseconds. During the photocycle, rhodopsins form kinetically distinctive photointermediates, such as the K, L, M, N and O intermediates, and show cognate protein function.
(A) Structure of the outward Na+-pumping rhodopsin KR2 (PDB: 6TK6). The all-trans retinal, conserved Lys residue (Lys255 in KR2 and Lys234 in BeNaR) and the functionally important residues for KR2 (His30, Ser70, Arg109, Asn112, Asp116, Gln123, Arg243, Asp251 and Ser254) were mapped on the structure with those of BeNaR (His19, Ser59, Arg94, Asn97, Asp101, Gln108, Arg222, Asp230 and Ser233). The triad Asn-Asp-Gln (i.e., Asn112-Asp116-Gln123 for KR2 and Asn97-Asp101-Gln108 for BeNaR) are called the NDQ motif. (B) Rooted phylogenetic tree of the known NaRs (KR2; GenBank ID: AB738960, DoNaR; JN827400, GLR; EHQ02967, NM-R2; KJ019877, NyNaR; KJ019875, NdNaR; AGC76155, FdNaR; AHN13814, IaNaR; EOZ93469, DeNaR; ESQ10031, JsNaR; WP_055664831, SrNaR; WP_052341415, PoNaR; WP_051881578, MiNaR; WP_018784639 and LaNaR; ERT09063) and BeNaR (OP508346). An outward H+ pumping rhodopsin, Halobacterium salinarum bacteriorhodopsin (HsBR; AAA72504), is also mapped for comparison. The scale bar indicates the number of substitutions per site. The numbers (e.g., 31 and 35) represent the bootstrap probabilities.
Ion transport is one of the major protein functions for microbial rhodopsins, and the ion transporting rhodopsins are divided into several types such as outward H+ pumps,4,5) inward H+ pumps,6,7) inward Cl− pumps,8) outward Na+ pumps,9) cation channels10) and anion channels.11) The generation and dissipation of ion concentration gradients formed by ion-transporting rhodopsins are utilized for various physiological responses in microorganisms. For instance, the proton gradient produced by outward H+ pumps is utilized as an energy source to produce the molecular currency, ATP.12) It should be noted that although all ion transporting rhodopsins consist of a similar seven-transmembrane architecture, they can selectively transport different ions, and their substrate ion species can be altered by adding a few amino acid substitutions.13) These characteristics make ion-transporting rhodopsins not only a model for photoactive proteins but also a model for ion transporters. In fact, the ion transport mechanisms of microbial rhodopsins have been extensively analyzed by physicochemical techniques, such as spectroscopy and crystallography.1) In addition to their basic importance, rhodopsins also have attracted plenty of attention from neurobiologists as tools for optogenetics, which is a method for controlling cellular activity by light.14,15) When the ion transporting rhodopsins are ectopically expressed in cell membranes using genetic methods, the ion concentration across the membranes can be regulated by light since the rhodopsins transport ions only during light irradiation. This optical regulation technique, called optogenetics, allows one to depolarize or hyperpolarize the electrical potential of the cell membrane even in millisecond scale. Therefore, neurobiologists use optogenetics to non-invasively manipulate neural activity.14,15) To expand the molecular basis of optogenetics, new rhodopsins have been progressively characterized by genomic and bioinformatic research.3)
An outward Na+-pumping rhodopsin (NaR) was first identified from the marine flavobacterium, Krokinobacter eikastus, in 2013 and named as KR2.9) Since then, a variety of NaRs (e.g., IaNaR, MiNaR, GLR and PoNaR) has been discovered in various bacteria including Indibacter alkaliphilus, Micromonospora sp. CNB 394, Gillisia limnaea DSM 15749 and Parvularcula oceani6,16–18) (Fig. 1B and Supplementary Fig. S1). Most of the known NaRs have an NDQ motif composed of asparagine (Asn), aspartic acid (Asp) and glutamine (Gln) residues (i.e., Asn112, Asp116 and Gln123 in KR2) in the third transmembrane helix, which is important for transporting Na+ in NaRs (Fig. 1A and Supplementary Fig. S1). Of note, all the known NaRs having the NDQ motif are obtained only from the bacteria living in high salt conditions (i.e., NaCl = 100–500 mM) such as marine water bodies and salt lakes (Fig. 1B). This is reasonable because the substrate for NaRs is a sodium ion (Na+) and an environmentally high concentration of Na+ has an advantage substrate uptake. On the other hand, metagenomic analysis of DNA from a hot spring microbial mat collected in 2017 from Mushroom Spring in Yellowstone National Park, WY, U.S.A., led to the identification of a gene encoding a putative rhodopsin in Bellilinea sp., a thermophilic member of the phylum Chloroflexota.19) It should be noted that this bacterium lives in low salt condition (NaCl = approx. 5 mM) and at high temperature (approx. 60 °C), yet the putative rhodopsin conserves all residues known to be important for transporting Na+ in KR2, such as His30, Ser70, Arg109, Asn112, Asp116, Gln123, Arg243, Asp251 and Ser254 in KR2 (Fig. 1A and Supplementary Fig. S1). Therefore, we named the rhodopsin Bellilinea Na+-pumping rhodopsin, BeNaR. In the cognate host organism, the putative substrate Na+ concentration for BeNaR is much lower than that for other NaRs, and the surrounding temperature for BeNaR is much higher than that for the other NaRs (e.g., approx. 15–20 °C for KR2, approx. 30 °C for IaNaR and approx. 1 °C for GLR),9,20,21) suggesting that BeNaR might show relatively high Na+-pumping activity and high thermal stability as adaptations to the environment.
In this study, we characterized the molecular function and photochemical properties of BeNaR by the biochemical and spectroscopic analyses of recombinant proteins expressed in Escherichia coli. By comparison with the typical NaR, KR2, we discuss the characteristic features of BeNaR and propose a model for its function.
The E. coli strain DH5α was used as a host for DNA manipulation. A gene for BeNaR (GenBank ID: OP508346) was chemically synthesized by the company Eurofins Genomics (Tokyo, Japan) with codon optimization for E. coli. During the optimization, the NdeI and XhoI restriction enzyme sites were added to each terminus and the codon-optimized gene was inserted into pET21a vector. Consequently, the plasmid encodes a hexahistidine tag (LEHHHHHH) at the C-terminus for protein purification. Mutant genes were constructed by In-Fusion Cloning Kit (TaKaRa Bio, Inc., Japan) as described previously.22) All constructed plasmids were analyzed using an automated DNA sequencer to confirm the expected nucleotide sequences. The plasmids for KR2, AR3 and TR were prepared as described previously.23) For phylogenetic analysis, amino acid sequences of microbial rhodopsins including BeNaR and other NaRs were aligned using ClustalW, and the evolutionary distances were estimated by using the JTT matrix-based method.24) The Maximum Likelihood tree was constructed using bootstrap values based on 1000 replications; evolutionary analyses were conducted using MEGA X software (https://www.megasoftware.net/).24)
Recombinant Protein Expression and the Ion Transport MeasurementsThe E. coli strain BL21 (DE3) was used as a host for protein expression as described previously.25) In short, the E. coli cells having the expression plasmid were grown in LB medium to an OD660 of 1.0–1.7 and the protein production was induced by adding 1 mM Isopropyl-β-D-thiogalactopyranoside (IPTG) for BeNaR, KR2 and TR or 0.1% (w/v) L(+)-arabinose for AR3 (Wako Pure Chemical Corporation, Japan). Simultaneously, 10 µM all-trans retinal (Sigma-Aldrich, U.S.A.) was added to the medium. After induction for 3 h at 37 °C, the cells were collected by centrifugation (2460 × g for 5 min).
For the ion transport measurement, the E. coli cells were washed three times in a solution containing 500 mM NaCl, 167 mM Na2SO4, 500 mM KCl or 5 mM NaCl/495 mM KCl for removal of culture medium and were resuspended in the same solution to adjust the cell density to OD660 = 7. Then, 7 mL of cell suspension was placed in the dark until the pH became stable and the temperature was kept at 25 °C using a thermostat. The suspension was illuminated for 3 min using a 300 W xenon lamp (MAX-303, Asahi Spectra, Japan) through a band-pass filter (520 nm; MX0520, Asahi Spectra) or a glass cut-off filter (Y-44, AGC Techno Glass, Japan). The light intensity was set at 7 mW/cm2 or 100 mW/cm2 using an optical power meter (#3664, Hioki, Japan) with an optical sensor (#9742, Hioki). The light-induced pH changes of the cell suspensions were measured with a Horiba F-72 pH meter. Measurements were repeated under the same conditions after the addition of the protonophore carbonylcyanide-m-chlorophenylhydrazone (CCCP, final concentration, 30 µM) and further addition of tetraphenylphosphonium (TPP+, final concentration, 20 mM). For NaRs, the initial slope amplitudes of the pH changes in the presence of CCCP from 0 to 10 s after illumination, which correspond to Na+ movement with a subsequent secondary H+ movement across the membrane, were defined as Na+-pumping activity.9)
Protein Purification, UV-Vis Spectroscopy and HPLC AnalysisFor the protein purification, the E. coli cells were disrupted by sonication (UD-211, TOMY Seiko Co., Ltd., Japan; Output 7, Duty 50 (1 pulse/0.5 s)) in a buffer containing 50 mM Tris–HCl (pH 8.0) and 1 M NaCl, and the suspension was incubated in an ice-cold waterbath until the liquid became clear after about 40 min. After the removal of debris by centrifugation (3000 × g for 10 min), the crude membranes were collected by ultracentrifugation (97300 × g for 60 min) and solubilized with 1.5% (w/v) n-dodecyl-β D/maltoside (DDM, Dojindo, Japan). The solubilized fraction was purified by Ni2+ affinity column chromatography with the linear gradient of imidazole as described previously.25) The purified protein was concentrated by centrifugation using an Amicon Ultra filter (30000 MW cutoff; Merk Millipore, U.S.A.). The sample medium was replaced by the appropriate buffer solution by ultrafiltration more than 3 times.
UV-visible spectra were recorded using a UV-2450 spectrophotometer (Shimadzu, Japan). For the measurement of the spectra, the purified samples were suspended in buffer A [50 mM Tris–HCl (pH 8.0), 0.05% (w/v) DDM and 1 M NaCl], buffer B [50 mM MOPS–NaOH (pH 7.0) and 0.05% (w/v) DDM] or buffer C [50 mM MOPS–NaOH (pH 7.0), 0.05% (w/v) DDM and 333 mM Na2SO4]. For the measurement of time-dependent thermal denaturation, the protein concentration of each sample was adjusted to 2 µM and the temperature was kept at 69 °C on a block heater. The residual protein activities after incubation were estimated from the absorbance at 524 nm for BeNaR, 524 nm for KR2, 558 nm for AR3 and 529 nm for TR, respectively. The logarithm of the amount of residual active protein, F, was plotted against time, t (min), and fitted by a single exponential decay function as follows (Eq. 1),
 | (1) |
where k is the denaturation rate constant and f0 is the initial amount of the active protein (approx. 0.2).
The retinal oxime isomer compositions of the purified sample in the buffer A were analyzed by normal-phase HPLC. The HPLC system consists of an LC-20AT chromatograph equipped with an SPD-20A photodetector (Shimadzu) and a YMC-Pack SIL 6.0 × 150 mm column (YMC, Japan) as described previously.25) Extraction of retinal oxime from the sample was carried out with hexane after denaturation in methanol and 500 mM hydroxylamine (Wako Pure Chemical Corporation) at room temperature (approx. 25 °C). The HPLC column was previously equilibrated with solvent containing 12% (v/v) ethyl acetate and 0.12% (v/v) ethanol in hexane (Wako Pure Chemical Corporation). The flow rate was 1.0 mL/min. The molar compositions of retinal oxime isomers were calculated from the areas of peaks monitored at 360 nm with molecular extinction coefficients of retinal isomers.25) Assignment of each peak was carried out by reference to the peaks of retinal oximes as described previously.25)
Quantification of Na+-Pumping ActivityThe E. coli cells used for measuring the light-induced pH change of BeNaR and KR2 were centrifuged (2460 × g for 10 min) and suspended in a buffer containing 1 M NaCl and 50 mM Tris–HCl (pH 8.0). After being washed and resuspended three times in the same solution, the cell suspensions were disrupted by sonication as described above. The membrane fraction was carefully collected by ultracentrifugation (97300 × g for 30 min) and only the red pellet was resuspended in the same buffer. The final volume of such suspensions was adjusted to 4 mL. UV-visible spectra were then recorded using a UV-2450 spectrophotometer with an ISR-2200 integrating sphere (Shimadzu) (Fig. 3C and Supplementary Fig. S3). The contribution of the background light scattering was eliminated by using the POWER function of reciprocal wavelength (λ) such as α+ (β/λ)γ 26) (Fig. 3C and Supplementary Fig. S3). Furthermore, the spectra were fitted with four log-normal equations that are composed of three or four bands (p1, p2, p3 and p4) of the pigments (p1; the main band, p2; the unknown band, p3; an additional band for the UV absorption of the protein, p4; the denatured band) (Fig. 3C and Supplementary Fig. S3). The log-normal equation is as follows (Eq. 2);
 | (2) |
where λmax is wavelength (nm); ω is half-bandwidth (cm−1); ρ is a parameter of skewness; A(λ) is absorbance; A is amplitude as previously described.25) From the absorbance of the main band (p1) at the wavelength of λmax and its molecular extinction coefficient (ε = 64000 M−1cm−1 for BeNaR and 56000 M−1cm−1 for KR2) (Supplementary Fig. S4A), the total amounts of photoactive proteins were estimated (Supplementary Fig. S4B). Finally, the initial slope amplitudes of the light-induced pH changes (see “Recombinant protein expression and the ion transport measurements” section) were normalized against the amounts of photoactive proteins to estimate the Na+-pumping activities.
The Time-Resolved Transient Absorption SpectroscopyFor the time-resolved absorption spectroscopy, the samples were adjusted to about 10 µM in the buffer A and placed in quartz cells. We employed two laser systems to measure the absorption changes at a wide range of time domains (i.e., from nanosecond to second). For detecting the fast component (100 ns–10 μs), we used second-harmonic light from a Nd3+: YAG laser (λ = 532 nm, pulse width 5 ns, Minilite II Continuum, U.S.A.) for a pump pulse as described previously.27) The energy of the pump pulse was 3 mJ/pulse. The pump beam was focused onto the sample with a lens. A Xe lamp (L9289-01, Hamamatsu Photonics, Japan) was used as a probe source equipped with a >300 nm band-pass filter (LB-145 filter) which enabled us to measure the absorption change after excitation over a range from 405 to 650 nm, and it was focused onto the sample from the opposite side of the pump beam. After being transmitted through the solution, the probe light was collected with a lens and was introduced to a polychromator (HR 320, Jobin Yvon, France). The dispersed probe light was projected onto an ICCD camera (PI-MAX-512-HQ, Princeton Instruments, U.S.A.) and therefore the transmitted light intensity could be monitored at various wavelengths. Finally, we measured the transient absorption spectra at 0.1, 0.3, 0.6, 1, 3, 6 and 10 μs after the excitation by changing the delay between the pump pulse and the ICCD gate.
For detecting the slow component (10 μs−1 s), we employed a homemade computer-controlled flash-photolysis system equipped with a Nd3+:YAG laser (Surelite I-10, Continuum, U.S.A.) as an actinic light source as described previously.26) The wavelength of the actinic pulse was tuned at 525 nm (4 ns) by an optical parametric oscillator (Surelite OPO Plus, Continuum). The pulse intensity was adjusted to 2.0 mJ/pulse. The time-dependent absorption changes were measured from 370 to 730 nm at 5 nm intervals. Fifty to 200 temporal traces were averaged at each selected wavelength to improve the signal-to-noise ratio.
To analyze the data, we employed an irreversible sequential model as reported previously.26) In short, the data obtained at all wavelengths were simultaneously fitted with a multi-exponential function. The appropriate number of exponents was determined from the reductions in the standard deviation of the weighted residuals. The experiments were performed by use of purified samples and the data taken before the flash of light were adopted as a baseline.
To investigate whether BeNaR encodes a functional protein, we designed a codon-optimized gene for BeNaR, had the gene synthesized, and then expressed it in E. coli cells to produce the recombinant protein. As seen, pink-colored E. coli cells were obtained for the cells having the expression plasmid of BeNaR, whereas the unchanged color of E. coli itself was observed in cells that contain the vector alone (Fig. 2A). Thus, BeNaR was successfully expressed in E. coli cells as a holoprotein.
(A) Photographs of E. coli cells expressing BeNaR and harboring the pET22b vector. (B) Light-induced pH changes of E. coli cells expressing wild-type BeNaR in solutions containing 500 mM NaCl, 167 mM Na2SO4 and 500 mM KCl in the absence (gray broken lines) or presence (black solid lines) of 30 µM CCCP and in the presence of 30 µM CCCP and 20 mM TPP+ (black broken lines). E. coli cells harboring pET22b vector in a solution containing 500 mM NaCl are used as a control. The cell suspension was illuminated with light (520 ± 10 nm for 3 min with intensity of approx. 7 mW/cm2). (C) The absorption spectra of the purified BeNaR of light-adapted, dark-adapted states and in no salt condition (i.e., buffer B) in the dark. (D) HPLC patterns of the retinal isomers of BeNaR in light-adapted and dark-adapted states. For light-adaptation, the sample was illuminated at 520 ± 10 nm for 3 min. The retinal isomers were extracted as retinal oximes. Ts, Ta, 13s and 13a denote all-trans/15-syn, all-trans/15-anti, 13-cis/15-syn and 13-cis/15-anti retinal isomers, respectively. The molar composition of each retinal isomer was calculated from the areas of the peaks in HPLC patterns using the coefficients of retinal isomers as described in Experimental.
As described, functionally essential residues for the known NaRs such as the NDQ motif (i.e., Asn97, Asp101 and Gln108) were completely conserved in BeNaR (Fig. 1A and Supplementary Fig. S1). To investigate whether BeNaR works as a Na+ pump, we measured the light-induced pH change of suspensions of cells expressing BeNaR (Fig. 2B), because increase in pH has been reported to correspond to outward Na+ movement with a subsequent secondary inward H+ movement across the membrane.9) In addition, NaRs did not show the outward Na+ transport activity when Na+ was completely removed in the extracellular solution,9,16,18) which suggests that the intracellular ion concentration is in equilibrium with the extracellular one, presumably due to the contribution of intrinsic ion transporters expressed in E. coli cells such as an Na+/H+ antiporter (Nha) and K+ transporters (Trk, Kdp, and Kef). In the solution containing 500 mM NaCl, the cells expressing BeNaR showed a light-induced pH increase, and the amplitude was strongly enhanced by adding the protonophore CCCP, but the pH increase was almost completely abolished by adding the ionophore TPP+ (Fig. 2B). As a control, the cells without BeNaR (pET22b vector) showed no light-induced pH change with and without CCCP. These results indicate that BeNaR transported Na+ outwardly or Cl− inwardly with a subsequent secondary inward H+ movement across the membrane according to a transmembrane electrical potential. To distinguish between these two possibilities, we performed similar experiments in the solution containing 167 mM Na2SO4 and 500 mM KCl. As seen, in Na2SO4 solution, the cells expressing BeNaR showed the similar light-induced pH increase, which was strongly enhanced by adding CCCP like in the solution containing NaCl (Fig. 2B). On the other hand, in KCl solution, the cells expressing BeNaR showed a light-induced pH decrease, which was abolished by adding CCCP (Fig. 2B). This behavior indicates that BeNaR transported H+ outwardly in the absence of Na+. From these results, we concluded that BeNaR is a light-driven outward Na+-pumping rhodopsin in the natural environment containing Na+. Of note, the amplitude of the pH change for BeNaR (maximum amplitude is approx. 0.15) is relatively higher than those of the other ion transporting rhodopsins we reported so far (e.g., Archaerhodopsin-3 (AR3) = 5.0 × 10−2,28) Rubricoccus marinus xenorhodopsin (RmXeR) = 0.1,29) Chlamydomonas reinhardtii cation channelrhodopsin (CrCCR) = 2.0 × 10−2 30) and Guillardia theta anion channelrhodopsin-2 (GtACR2) = 1.3 × 10−3.31) This means BeNaR has a relatively large Na+-pumping activity. Furthermore, we measured light-induced pH changes in low Na+ concentration (i.e., NaCl = 5 mM), where the ionic strength was kept constant with 495 mM KCl. As seen in Supplementary Fig. S2, the acidification and alkalization of the external media were observed for the E. coli cells without and with CCCP, respectively, suggesting that BeNaR works both as an outward H+ pump (without CCCP) and as an outward Na+ pump (with CCCP) under the host organism’s physiological environment (Na+ = approx. 5 mM).
To investigate the photochemical properties of BeNaR, we purified BeNaR by Ni2+ affinity column chromatography and carried out spectroscopic measurements (Fig. 2C). As seen, the absorbance maximum of BeNaR was located at 524 nm in the dark, which is similar to those of other NaRs, such as KR2 (523.5 nm) and IaNaR (525 nm).9,32) The absorbance spectrum was not significantly changed by light adaptation and by removal of substrate Na+, suggesting no retinal configuration changes and no Na+ binding to the protein around the chromophore retinal. These features were also observed other NaRs.9,16) We then performed HPLC experiments to determine the retinal configuration. Archaeal ion pumping rhodopsins, such as bacteriorhodopsin and halorhodopsin, possess both all-trans (30–50%) and 13-cis retinal (50–70%) in the dark-adapted state, and the ratio of all-trans retinal is increased to approximately 100% upon light-adaptation.33,34) On the other hand, the bacterial ion pumping rhodopsins, such as proteorhodopsin, xanthodrhodopsin and other NaRs, possess all-trans retinal predominantly in both the dark- and light-adapted states.9,35,36) As shown in Fig. 2D, the isomeric state of retinal is predominantly all-trans (>96%), with a small proportion (<4%) of 13-cis in both the dark- and light-adapted states, indicating that the retinal composition of BeNaR reflects that of bacterial ion pumping rhodopsins.
Quantitative Na+-Pumping ActivityAs described above, the light-induced pH change for BeNaR is relatively higher than those of the other ion transporting rhodopsins, suggesting high Na+-pumping activity of BeNaR as expected. However, the signal amplitude is also affected by the amount of the protein production. Then, we quantitatively estimated the Na+-pumping activity of BeNaR and compared it with that of KR2 (Fig. 3, Supplementary Figs. S3 and S4). The estimation has been performed with the same strategy as described previously.26) Firstly, we measured the light-induced pH changes in 500 mM NaCl solution with CCCP, and the initial slope was calculated from 0 to 10 s after light irradiation at different light intensities (Figs. 3A, 3B). To check the light intensity, these initial slope amplitudes were plotted against the excitation light intensity (Fig. 3B). Because a linear relationship was observed at light intensities below 175 mW/cm2 for both BeNaR and KR2, we obtained the initial slope at 100 mW/cm2 as an apparent Na+-pumping activity. Secondly, we estimated the amount of protein production by the spectroscopic method (see the text of Experimental for details). Figure 3C showed the absorbance spectrum of the membrane fraction BeNaR sample identical to that used for measuring the light-induced pH change as shown in Fig. 3A. The spectrum was fitted by the power functions of reciprocal wavelength (λ) such that α+(β/λ)γ, which reflects the contribution of background light scattering of the sample (curve 2 in Fig. 3C and Supplelmentary Fig. S3). The residual spectrum was then fitted by four log-normal equations that are composed of four vibronic bands p1–p4 with small residuals (curve 4 in Fig. 3C and Supplementary Fig. S3). The absorbance peaks (525 nm for BeNaR and 529 nm for KR2, respectively) of the main band p1 were almost identical to the peaks of the absorbance spectrum of the purified BeNaR and KR2 (524 and 524 nm, respectively) (Fig. 2C), implying that p1 was from the holoprotein of rhodopsins. The other bands, p2, p3 and p4 were estimated to be from some pigments contained in the membranes (p2 and p3) and UV absorbance of the aromatic residues such as tryptophan (Trp) and tyrosine (Tyr) (p4). Then, using the absorbance of the band p1 at the wavelength of absorption maximum (λmax) and the molar extinction coefficient of the purified BeNaR and KR2 (64000 and 56000 M−1cm−1, respectively) (Supplementary Fig. S4A), we calculated the molar concentration of BeNaR and KR2 in the sample as 2.2 and 1.8 µM, respectively (Supplementary Fig. S4B). Finally, the initial slope amplitudes of BeNaR and KR2 at 100 mW/cm2 were normalized by the amounts of photoactive proteins to evaluate the Na+-pumping activities (Fig. 3D). As a result, the Na+-pumping activity for BeNaR was 2.5 times higher than that for KR2, indicating that BeNaR was a highly efficient Na+-pumping rhodopsin.
(A) Light-induced pH changes of E. coli cells expressing BeNaR and KR2 in solutions containing 500 mM NaCl in presence of 30 µM CCCP. The cell suspension was illuminated (>420 nm with intensity of approx. 100 mW/cm2). E. coli cells harboring pET22b vector in a solution containing 500 mM NaCl are used as a control. The initial slope amplitudes from 0 to 10 s after illumination was obtained. (B) Relationship between the initial slope amplitudes and light intensities. The initial slope amplitudes were plotted against varying intensities of light. The data below 175 mW/cm2 were fitted well by a linear regression (R2 = 0.99). (C) Estimation of the total amounts of photoactive BeNaR proteins in E. coli cells used for the ion transport assay. The contribution of background light scattering (curve 2) was subtracted from the spectra of E. coli membrane fragments used for the ion transport assay (curve 1) to obtain the spectrum consisting of the absorption of photoactive proteins and contaminated proteins (curve 3). Curve 3 was fitted by four log-normal equations (p1: pink curve, p2: orange curve, p3: green curve, p4: blue curve). Curve 4 is the residual spectrum between curve 3 and the fitting curve. From the absorbance of the main band (p1), the total amounts of photoactive proteins were estimated. (D) Quantitative evaluation of the Na+-pumping activities. The Na+-pumping activities of BeNaR and KR2 were estimated from the initial slope of light-induced pH changes with CCCP at 100 mW/cm2 and then normalized by the expression level of photoactive proteins (Supplementary Fig. S4B). Each measurement was performed three times and the data were averaged with their standard deviations (n = 3). Data are presented as mean ± standard error of the mean (S.E.M.) (n = 3). * p < 0.01 Dunnett’s test versus BeNaR.
As described in the introduction, ion transporting rhodopsins have attracted much attention not only as models for ion transporters but also as optogenetic tools. However, it is known that the chromophore retinal is easily bleached upon irradiation above 40 °C, and therefore many microbial rhodopsins show low thermal stability at high temperature,23) which potentially disturbs their comprehensive physicochemical analysis and their suitability for optogenetics. From these points of view, high thermal stability has advantages for analyzing and utilizing microbial rhodopsins. Because BeNaR was discovered in the genome of a thermophilic bacterium that normally lives at 60 °C,19) it is expected to have high thermal stability. To check the thermal stability, we incubated the purified BeNaR at 69 °C (Fig. 4A). As seen, BeNaR held its absorbance at λmax (i.e., 524 nm), and more than 50% of the protein retained the color after heating for 300 min (Fig. 4A). For comparison, we performed similar experiments for Na+ pump KR2, a typical outward H+ pumping rhodopsin AR3, and thermally stable outward H+ pumping rhodopsin thermophilic rhodopsin (TR) (Supplementary Fig. S5). Figure 4B shows the time-dependent decrease in the absorbance at λmax of BeNaR, KR2, AR3 and TR at 69 °C. AR3 lost its absorbance at λmax (i.e., 558 nm) within 10 min (blue circle in Fig. 4B), while KR2 lost its absorption (approx. 82%) at λmax (i.e., 524 nm) after the 300 min incubation (pink circle in Fig. 4B). On the other hand, BeNaR and TR held the absorbance at λmax (524 nm for BeNaR and 529 nm for TR) after the 300 min incubation and more than 56% of BeNaR and 83% of TR retained their colors (black circle for BeNaR and gray circle for TR). The denaturation rate (time) constants were estimated using a single exponential function as described in Experimental (solid lines in Fig. 4B), and resulted in 2.4 × 10−3 [min−1] (τ1/2 = 6.9 h) for BeNaR, 6.0 × 10−3 [min−1] (τ1/2 = 2.8 h) for KR2, 7.4 × 10−1 [min−1] (τ1/2 = 2.2 × 10−2 h) for AR3 and 8.3 × 10−4 [min−1] (τ1/2 = 20.0 h) for TR (Fig. 4C). These results indicate that BeNaR is more stable than KR2 by 2.5-times and AR3 by 309-times at 69 °C (Fig. 4C), whereas the stability of BeNaR is lower than TR by 2.9-times at 69 °C. In this study, we employed a single exponential function for comparison. However, the curve for BeNaR (black line) does not seem to be the best fit, suggesting BeNaR has multiple denaturation processes.
(A) Time-dependent decreases in the absorbance at 524 nm for BeNaR. The purified sample was suspended in the buffer A. The protein concentration of sample was adjusted to 2 µM. The temperature was kept at 69 °C using a thermostat. (B) Denaturation kinetics at 69 °C obtained from time-dependent decreases in visible absorbance of BeNaR (black circles), KR2 (pink circles), TR (gray circles) and AR3 (blue circles) in the buffer A. Their raw data of KR2, TR and AR3 were shown in Supplementary Fig. S5. The solid lines represent fitting curves of a single exponential function. (C) Comparison of the denaturation rate constants of BeNaR, KR2, TR and AR3. All data are presented as mean ± S.E.M. (n = 3). * p < 0.01 Dunnett’s test versus BeNaR.
Recently, we successfully produced more thermally stable mutants of TR and Rubrobacter xylanophilus rhodopsin (RxR) based on structural and computational analyses.37) We expect that the further structural and computational analyses of BeNaR should lead to the production of more thermally stable BeNaR mutants.
Photocycle of BeNaRUpon light absorption, microbial rhodopsins exhibit a photocycle during which the cognate biological function is induced.1–3) Therefore, investigation of the photocycle is essential for understanding the protein function. It is known that each photointermediate has a characteristic spectral sensitivity so that it is possible to detect the formation and decay of intermediates by observing the transient changes in absorbance at specific wavelengths.1–3) For instance, KR2 sequentially forms K intermediate at 590 nm, L intermediate at 472 nm, M intermediate at 420 nm and O intermediate at 605 nm from nanosecond to millisecond time domains.9,38) To investigate the photocycle of BeNaR, we employed time-resolved transient absorption spectroscopy for the purified BeNaR.
Figure 5A shows the flash-induced difference absorbance spectrum from 100 ns to 1000 ms after irradiation. As seen, we observed a negative peak at around 530 nm corresponding to the bleaching of the original state and positive peaks at around 420 and 600 nm, due to the formation of some photointermediates. From the analogy with NaRs, we obtained absorbance changes of BeNaR at 530, 420, 470, and 600 nm which could be tentatively assignable to the absorbance changes of original state, M, L and K/O intermediates, respectively (Fig. 5B). To express the photoreaction more precisely, we performed global fitting and found that the experimental results were well simulated by the sum of four exponential terms (solid lines in Fig. 5B), indicating the existence of four photochemically defined states (designated as P1–P4) during the photocycle of BeNaR. According to the sequential model, the absorbance spectra of the P1–P4 states were calculated as shown in Fig. 5C, and the decay time constants of P1–P4 (i.e., τ1–τ4) were determined as 11, 120 µs, 1.3, and 19 ms, respectively, as shown in Fig. 5D. The P1 state showed a λmax at around 600 nm, indicating that it arises from the K intermediate. The P2 state was produced by the conversion from the P1 state with a time constant of 11 µs, and showed a broad absorption spectrum from 400 to 700 nm, which seems not to follow Gaussian distribution, indicating the existence of several photointermediates. Judging from the time region and the locations of absorption bands, we tentatively assigned the P2 state as mixture of two intermediates, K (600 nm) and L (470 nm). During the transition from the P2 state to the P3 state with a time constant of 120 µs, a spectral blue-shift was observed. The P3 state showed a broad absorption spectrum from 400 to 550 nm, which was tentatively assigned to mixture of two intermediates, L (470 nm) and M (420 nm). Then, the P3 state was converted to the P4 state with a time constant of 1.3 ms. The main absorption of the P4 state appeared at 600 nm, which was assigned to O intermediate. Finally, the P4 state was converted to the original P0 state with a time constant of 19 ms, which is the rate limiting step of the photoreaction, and thus the photocycle is finished. The estimated photocycle scheme of BeNaR is summarized in Fig. 5D.
(A) Flash-induced difference absorption spectra of the purified BeNaR in the buffer A at 25 °C. (B) Absorbance changes at 530, 420, 470, and 600 nm correspond to the depression and recovery of the original state (green circles), the formation and decay of M intermediate (light blue circles), the formation and decay of L intermediate (purple circles), the formation and decay of K and O intermediates (red circles), respectively. The data were fitted by four log-normal equations (solid black lines). (C) The calculated absorption spectra of four kinetically-defined intermediate states, P1, P2, P3, P4, and an original state P0 (i.e., absorption spectrum of the purified BeNaR in the dark). (D) Photocycle model of BeNaR.
It is well-known that charged amino acid residues such as Asp, glutamic acid (Glu) and lysine (Lys) play essential roles for ion transporters.1–3) For KR2, it has been reported that Asp116 acts as the counterion of PRSB and receives a proton (H+) from PRSB upon the formation of the M intermediate after illumination.9) Because of the lack of Na+-pumping activity by the substitution of Asp116 (e.g., D116A, D116N, D116E and D116T) in KR2, the H+ transfer from PRSB to Asp116 is thought to be essential for the formation of the Na+ transport pathway during photocycle.9,13) In addition, it has been reported that Asp251 forms the hydrogen-bonding network with Asn112 to form Na+ binding site at O intermediate.39) In fact, the substitution of Asp251 (e.g., D251A, D251N and D251E) abolished the Na+-pumping activity, indicating the importance of Asp251 in Na+ pump functions. Thus, Asp116 and Asp251 are essential residues for Na+-pumping ability in KR2. Because the corresponding two carboxylates, Asp101 and Asp230 (Asp116 and Asp251 in KR2) are also conserved in BeNaR (Fig. 1A and Supplementary Fig. S1), we decided to investigate their roles in BeNaR by mutational experiments. It is known that the neutralization of the counterion of the microbial rhodopsins causes a large spectral red-shift about 30–40 nm.1,9) The chromophore retinal is surrounded by seven-transmembrane α-helices and absorbs light at different wavelengths due to differences in the electronic energy gap between its ground- and excited states, leading to changes in the λmax of the rhodopsins. Therefore, a spectral red-shift by the neutralization of the counterion is conceivable as the excited state charge density is shifted against the charge of the counterion, leading to a change in the dipole moment.1) Therefore, we prepared and measured the absorbance spectra of D101N and D230N mutants of BeNaR in Na2SO4 solution (Fig. 6A), because Cl− can compensate for the spectral shift by binding to BeNaR as a surrogate counterion in the disappearance of the negative charge of the counterion residue. As seen, the λmax of D230N was located at 524 nm, which was almost identical to that of wild-type BeNaR (524 nm), whereas that of D101N was largely shifted to 570 nm (Fig. 6A). These results clearly indicated that Asp101 of BeNaR was the counterion and that Asp230 was not.
(A) Absorption spectra of the purified wild-type BeNaR, D101N and D230N mutants in the buffer C. (B) Light-induced pH changes of E. coli cells expressing D101N and D230N in solutions containing 167 mM Na2SO4 in the absence (gray broken lines) or presence (black solid lines) of 30 µM CCCP. The cell suspension was illuminated with light (>420 nm with intensity of approx. 100 mW/cm2). (inset) Photographs of E. coli cells expressing BeNaR D101N and D230N after washing in a solution containing 167 mM Na2SO4.
To investigate the functional roles of the two carboxylates, we measured the light-dependent pH changes of D101N and D230N mutants (Fig. 6B). The bright purple-colored and pink-colored E. coli cells were observed for the D101N and D230N mutants, respectively. Judging from the depth of the color, the expression level of the mutants is roughly comparable to that of the wild-type (Fig. 2A). As seen, the cells expressing D101N mutant in Na2SO4 solution showed almost no light-induced alkalization in contrast with the wild-type BeNaR. These results indicated that Asp101 is an essential residue for the Na+-pumping ability of BeNaR as the counterion and H+ acceptor of PRSB as well as KR2. In addition, the cells expressing D230N mutant in Na2SO4 solution showed almost no light-induced alkalization as well as the D101N mutant, indicating that Asp230 is also an essential residue for Na+-pumping ability of BeNaR as well as KR2. This suggests the existence of the hydrogen-bonding network, consisting of Asn97 (Asn112 in KR2), Asp230 (Asp 251 in KR2) that forms Na+ binding site at O intermediate in BeNaR like KR2.
In this study, we newly identified BeNaR which works as an outward Na+-pumping rhodopsin. Flash-photolysis experiments revealed that BeNaR absorbs green light and sequentially forms four kinetically distinctive intermediates, K, L, M, and O, and then returns to the original state. Results suggest that BeNaR outwardly transports Na+ according to conformational changes of the protein moiety during the photocycle. From the results of the photochemical and mutational analysis of BeNaR and other related findings for KR2, we discuss and propose a hypothetical model for the outward Na+-pumping mechanism of BeNaR (Fig. 7) as follows; (i) In the dark (unphotolyzed) state, BeNaR possesses all-trans retinal predominantly and the PRSB is stabilized by the deprotonated counterion residue, Asp101 to absorb green light (i.e., 524 nm). Based on the crystal structure of KR2, we assume that Asp230 forms the hydrogen-bonding network with a water molecule and Asn97.39) The similar absorbance spectra of BeNaR in the presence and absence of the substrate Na+ (Fig. 2C) suggests that the substrate Na+ does not bind to the unphotolyzed BeNaR like KR2.39–41) On the other hand, it should be noted that the negative charge of Asp102 locating at the extracellular side in KR2 forms a binding site to Na+. However, this residue is replaced and neutralized by a glycine (Gly) residue for BeNaR (i.e., Gly87) and therefore it is assumed that the Na+ binding site is absent in BeNaR and suggesting that the binding site is not indispensable for Na+-pumping ability. In fact, the replacement of Asp102 in KR2 did not change the Na+-pumping ability.41) (ii) After illumination of the original state, the retinal chromophore is isomerized from the all-trans to the 13-cis upon formation of the K intermediate in a similar way to other microbial rhodopsins including KR2. From the stimulated Raman spectroscopic analysis for KR2, we assumed that the K intermediate contained a distorted chromophore structure in BeNaR.42) (iii) During the relaxation process of its chromophore distortion, deprotonation of the Schiff base and concomitant protonation of the counterion (i.e., Asp101) were caused upon formation of the M intermediate through the L intermediate. The loss of the ion pair between Asp101 and Lys234 induces the flipping motion of Asp101 and the water molecule around Asp230 shifts away, which increases in the cavity around the Schiff base as proposed from the crystal structure of the intermediates of KR2.41) Because the Na+-pumping ability was abolished by D101N mutation in BeNaR (Fig. 6B), the loss of the ionic interaction probably play an essential role for the Na+ transport. (iv) Upon the formation of O intermediate from the M intermediate, the Schiff base is re-protonated by proton transfer from Asp101. The structural changes lead to the uptake of Na+ from the intracellular side to the transient Na+ binding site consisting of Asn97 and Asp230, which is supported by the crystal structure and the experimental results that the formation of O intermediate is accelerated at a high Na+ concentration for KR2.9,39) (v) Thermal relaxation occurs to the retinal and it re-isomerizes from 13-cis to all-trans upon the decay of the O intermediate into the original state. From the analogy with KR2, Na+ is transported from the transient binding site to the extracellular side. To confirm the hypothesis, we will perform further spectroscopic and structural analysis in the future.
The substrate Na+ is transported from the intracellular side to the extracellular one during the photocycle through several steps as following; (i) The substrate Na+ does not bind to the unphotolyzed BeNaR. Asp101 works as a counterion of the PRSB in the original state. Asp230 forms the hydrogen-bonding network with a water molecule and Asn97. (ii) BeNaR absorbs green light (i.e., 524 nm) and its absorption triggers the formation and decay of four kinetically distinctive intermediates, K, L, M and O. The retinal chromophore is isomerized from the all-trans to the 13-cis upon formation of the K intermediate. (iii) Deprotonation of the Schiff base and concomitant protonation of Asp101 were caused upon formation of the M intermediate through the L intermediate. The loss of the ion pair between Asp101 and Lys234 induces the flipping motion of Asp101 and the water molecule around Asp230 shifted away, which increases in the cavity around the Schiff base for the uptake of Na+. (iv) Upon the formation of O intermediate from the M intermediate, the Schiff base is re-protonated by the proton transfer from Asp101. The structural changes lead to the uptake of Na+ from the intracellular side to the transient binding site consisting of Asn97 and Asp230. (v) Thermal relaxation occurs to the retinal and it is re-isomerized from the 13-cis to the all-trans, returning to the original state.
The quantitative ion transport measurements revealed that the Na+-pumping activity BeNaR was 2.5-times greater than that of the typical Na+ pump KR2. KR2 has been used as a neural silencing tool by its light-induced outward Na+-pumping activity.41,43) Therefore, BeNaR is expected to induce larger photocurrents and could be used as a new optogenetics tool. We focused on the molecular origin of the difference in the Na+-pumping activity between BeNaR and KR2. As a hint, identity and similarity of the amino acid sequences between them are 41 and 59%, respectively (Supplementary Fig. S1). The sequence difference (i.e., amino acid substitution, insertion and deletion) should be involved in the activity difference. Because the substrate Na+ is transported inside the protein moiety through the retinal chromophore, the amino acid residues around the retinal may be a main determinant of the difference in Na+-pumping activity between BeNaR and KR2. We compared the amino acid residues located within 6 Å from the conjugated polyene chain of the retinal or any methyl group of the polyene chain in the crystal structure of KR2 (PDB : 6TK6) and found that four residues (i.e., Leu137, Gly157, Thr160, and Tyr163 in BeNaR) differ between them. These four residues should be good candidates as determinants regulating the Na+-pumping efficiency. The Na+-pumping efficiency is thought to be proportional to the photocycle turnover rate since NaRs transport the substrate Na+ during the photocycle. The photocycle turnover rate was estimated to be 53 ms−1 in BeNaR, which is larger than that of KR2 (approx. 20 ms−1).9) The difference in the photocycle rate might be one of factors to explain the difference in the Na+-pumping activity between BeNaR and KR2. To explore the molecular origin of the difference in the Na+-pumping activity, we will perform further functional and spectroscopic analysis of the BeNaR mutants against the above four residues and KR2 in the future.
In addition to the high Na+-pumping activity, BeNaR showed higher thermal stability as a unique feature. Instead of KR2, BeNaR can be a suitable model as a Na+-pumping rhodopsin for comprehensive physicochemical analysis, particularly methods that require high protein stability, such as X-ray crystallography, X-ray free-electron laser analysis and NMR spectroscopy. However, the variety of amino acid sequence differences between BeNaR and KR2 (Supplementary Fig. S1) makes it difficult to focus on the region which is essential for the high stability of BeNaR, because all parts can participate in the protein stability in terms of both increase in entropy (e.g., hydrophobic interaction) and decrease in enthalpy (e.g., hydrogen bonding).23,44) In fact, even N- and C-terminal regions having high flexibility are involved in the interaction with not only membrane-embedded region but also with phospholipids in the membrane, leading to the alteration of the protein stability.23,44) So far, we measured the denaturation kinetics of various microbial rhodopsins such as Gloeobacter violaceous rhodopsin (GR), AR3, Haloquadratum walsbyi bacteriorhodopsin (HwBR) and TR at varying temperatures, and determined their activation energies, Ea, using Arrhenius plots.23) As a result, the estimated Ea values fell within the range of 9–13 kcal/mol, in spite of extremely large stability differences among them (2.5–306-fold). We concluded that the rhodopsins having high thermal stability such as TR and RxR have smaller values for the Arrhenius pre-exponential factor “A” compared with the rhodopsins having low thermal stability. To investigate whether BeNaR has the small pre-exponential factor as well as TR and RxR, further investigation will be needed.
The efficient Na+-pumping activity and high thermal stability of BeNaR make it a useful tool for optogenetics. We discuss the use of BeNaR for optogenetics in mammalian cells. It is known that the expression level and membrane localization of NaRs including KR2 are generally low in mammalian neurons.43,45) For the application of BeNaR as an optogenetics tool, it will be necessary to investigate and improve the expression level and membrane localization of BeNaR in neurons by the attachments of endoplasmic reticulum and the Golgi export sequences with the sequence of BeNaR as reported in KR2.43) In the future we will check the applicability of BeNaR as an optogenetics tool by introducing the optimized BeNaR sequences with the above attachments into mammalian cultured neurons.
In this study, we successfully identified and characterized a rhodopsin from a thermophilic Bellilinea sp., a member of the phylum Chloroflexota. Indeed, some spectroscopic properties of the protein named BeNaR such as absorbance maximum, retinal configuration and photocycle reaction have been clarified. From the results, we propose a hypothetical model for the outward Na+-pumping mechanism of BeNaR. Of note, BeNaR showed highly efficient Na+-pumping activity and high thermal stability. The characteristics of BeNaR make it a useful model both for ion transporters and for use as an optogenetic tool.
We wish to thank Drs. Takashi Tsukamoto and Takashi Kikukawa for fruitful experimental supports, and Drs. Keiichi Inoue and Hideki Kandori for a kind gift of the KR2 gene. This work was financially supported by JSPS KAKENHI Grant Numbers JP19K16090 and JP21K15054 to KK, JP18H02411, JP19H04727, JP19H05396, JP20K21482, JP21H0040413, and JP21H0244613 to YS, and JST CREST (JPMJCR1656) and AMED (20dm0207060h0004) to YS. This study was partly funded by the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences of the Department of Energy through Grant DE-FG02-94ER20137 to DAB. DMW and DAB additionally acknowledge support from the NASA Exobiology program (NX09AM87G). This work was also partly supported by the U.S. Department of Energy (DOE), Office of Biological and Environmental Research (BER), as part of BER’s Genomic Science Program 395 (GSP). This contribution originates from the GSP Foundational Scientific Focus Area (FSFA) at the Pacific Northwest National Laboratory (PNNL) under a subcontract to DAB. The materials used in this study were collected under permit#YELL-SCI-0129 held by DMW and administered under the authority of Yellowstone National Park. The authors especially thank Christie Hendrix and Stacey Gunther for their advice and assistance.
The authors declare no conflict of interest.
This article contains supplementary materials.
