Light-driven Proton Pumps as a Potential Regulator for Carbon Fixation in Marine Diatoms

Diatoms are a major phytoplankton group responsible for approximately 20% of carbon fixation on Earth. They perform photosynthesis using light-harvesting chlorophylls located in plastids, an organelle obtained through eukaryote-eukaryote endosymbiosis. Microbial rhodopsin, a photoreceptor distinct from chlorophyll-based photosystems, was recently identified in some diatoms. However, the physiological function of diatom rhodopsin remains unclear. Heterologous expression techniques were herein used to investigate the protein function and subcellular localization of diatom rhodopsin. We demonstrated that diatom rhodopsin acts as a light-driven proton pump and localizes primarily to the outermost membrane of four membrane-bound complex plastids. Using model simulations, we also examined the effects of pH changes inside the plastid due to rhodopsin-mediated proton transport on photosynthesis. The results obtained suggested the involvement of rhodopsin-mediated local pH changes in a photosynthetic CO2-concentrating mechanism in rhodopsin-possessing diatoms.

Diatoms are unicellular, photosynthetic algae found throughout aquatic environments and are responsible for up to 20% of annual net global carbon fixation (Nelson et al., 1995;Field et al., 1998). Since their contribution to primary production in the ocean is significant, their light utilization mechanisms are essential to correctly understand marine ecosystems. Diatoms contain chlorophylls a and c and carotenoids, such as fucoxanthin, as photosynthetic pigments in the plastids acquired by eukaryote-eukaryote endosymbiosis (Keeling, 2004). Some diatoms have recently been shown to contain microbial rhodopsin (henceforth rhodopsins), a light-harvesting antenna distinct from the chlorophyllcontaining antenna for photosynthesis (Marchetti et al., 2015). Although rhodopsin-mediated light-harvesting may support the survival of these diatoms in marine environments, the physiological role of rhodopsin in diatom cells remains unclear.
Microbial rhodopsins are a large family of seven transmembrane photoreceptor proteins (Spudich et al., 2000). Rhodopsin has an all-trans retinal as the light-absorbing chromophore, and its protein function is triggered by the light-induced isomerization of the retinal. The first microbial rhodopsin, the light-driven proton pump bacteriorhodopsin (BR), was discovered in halophilic archaea (Oesterhelt and Stoeckenius, 1971). Although rhodopsin was initially considered to only occur in halophilic archaea inhabiting hypersaline environments, subsequent studies showed that the rhodopsin gene is widely distributed in all three domains of life (Beja et al., 2000;Sineshchekov et al., 2002). Rhodopsins are classified based on their functions into light-driven ion pumps, light-activated signal transducers, and light-gated ion channels. The two former functional types of rhodopsin have so far been identified in prokaryotes (Ernst et al., 2014). Rhodopsins in prokaryotes, regardless of function, localize to the cell membrane in which they operate. For example, proton-pumping rhodopsins export protons from the cytosol across the cell membrane to convert light energy into a proton motive force (PMF) (Yoshizawa et al., 2012). The PMF induced by rhodopsin ion transport is utilized by various physiological functions, such as ATP synthesis, substrate uptake, and flagellar movement.
Rhodopsins functioning as light-driven ion pumps and light-gated ion channels have been reported in eukaryotic microorganisms (Ernst et al., 2014;Kikuchi et al., 2021). A light-gated ion channel called channelrhodopsin, which localizes to the plasma membrane over the eyespot within the chloroplast of green algae, has been extensively exam-ined for its role in phototaxis (Nagel et al., 2002). The other type of rhodopsin in eukaryotes, light-driven ion-pumping rhodopsins has been detected in a number of organisms belonging to both photoautotrophic and heterotrophic protists (Slamovits et al., 2011;Marchetti et al., 2015). Since the intracellular membrane structure of eukaryotic cells is more complex than that of prokaryotes, containing various organelles, even light-driven ion-pumping rhodopsins may have distinct physiological roles depending on their subcellular localization (Slamovits et al., 2011). However, due to the difficulties associated with identifying the exact localization of rhodopsins in eukaryotic cells, their subcellular localization remains unknown.
In the present study, to clarify the physiological function of rhodopsin in a marine pennate diatom, we investigated the phylogeny, protein function, spectroscopic characteristics, and subcellular localization of rhodopsin from a member of the genus Pseudo-nitzschia. Heterologous expression techniques were used to analyze protein functions and spectroscopic features. The expression of rhodopsin fused with a green fluorescent protein, eGFP revealed its subcellular localization in a model diatom (Phaeodactylum tricornutum). Furthermore, a model-based analysis was performed to evaluate the impact of the potential roles of rhodopsin in cellular biology.

Rhodopsin sequences and phylogenetic analysis
The rhodopsin sequence of the diatom Pseudo-nitzschia granii was previously reported (Marchetti et al., 2015). All other rhodopsin sequences used in the phylogenetic analysis were collected from the National Center for Biotechnology Information. Detailed information on the strains used in this analysis is given in Extended Data Fig. 1. Sequences were aligned using MAFFT version 7.453 with the options '-genafpair' and '-maxiterate 1000' (Katoh and Standley, 2013). The phylogenetic tree was inferred using RAxML (v.8.2.12) with the 'PROTGAMMALGF' model using 1,000 rapid bootstrap searches (Stamatakis, 2014). Model selection was performed with the ProteinModelSelection.pl script in the RAxML package.

Gene preparation, protein expression, and ion transport measurements of Escherichia coli cells
In the present study, a functional analysis of the rhodopsin possessed by the diatom P. granii (named PngR, accession no. AJA37445.1) was performed using a heterologous expression system. The full-length cDNA for PngR, the codons of which were optimized for E. coli, were chemically synthesized by Eurofins Genomics and inserted into the NdeI-XhoI site of the pET21a(+) vector as previously described (Hasegawa et al., 2020). A hexahistidine-tag was fused at the C terminus of PngR, which was utilized for the purification of the expressed protein. The heterologous protein expression method is the same as that previously reported (Inoue et al., 2018). E. coli BL21(DE3) cells har-boring the cognate plasmid were grown at 37°C in LB medium supplemented with ampicillin (final concentration of 50 μg mL -1 ). Protein expression was induced at an optical density at 600 nm of 0.7-1.2 with 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) and 10 μM all-trans retinal, after which cells were incubated at 37°C for 3 h. The proton transport activity of PngR was measured as light-induced pH changes in suspensions of E. coli cells as previously described (Inoue et al., 2018). Briefly, cells expressing PngR were washed more than three times in 150 mM NaCl and then resuspended in the same solution for measurements. Each cell suspension was placed in the dark for several min and then illuminated using a 300 W Xenon lamp (ca. 30 mW cm -2 , MAX-303; Asahi Spectra) through a >460 nm long-pass filter (Y48; HOYA) for 3 min. Measurements were repeated under the same conditions after the addition of the protonophore carbonyl cyanide mchlorophenylhydrazone (CCCP) (final concentration=10 μM). Light-induced pH changes were monitored using a Horiba F-72 pH meter. All measurements were conducted at 25°C using a thermostat (Eyela NCB-1200; Tokyo Rikakikai).
Purification of PngR from E. coli cells and spectroscopic measurements of the purified protein E. coli cells expressing PngR were disrupted by sonication for 30 min in ice-cold buffer containing 50 mM Tris-HCl (pH 7.0) and 300 mM NaCl. The crude membrane fraction was collected by ultracentrifugation (130,000×g at 4°C for 60 min) and solubilized with 1.0% (w/v) n-dodecyl-β-D-maltoside (DDM; DOJINDO Laboratories). The solubilized fraction was purified by Ni 2+ affinity column chromatography with a linear gradient of imidazole as previously described (Kojima et al., 2020b). The purified protein was concentrated by centrifugation using an Amicon Ultra filter (30,000 M w cut-off; Millipore). The sample medium was then replaced with Buffer A (50 mM Tris-HCl, pH 7.0, 1 M NaCl, and 0.05% [w/v] DDM) by ultrafiltration 3 times.
The absorption spectra of purified proteins were recorded using a UV-2450 spectrophotometer (Shimadzu) at room temperature in Buffer A. The retinal composition in PngR was analyzed by highperformance liquid chromatography (HPLC) as previously described (Kojima et al., 2020a). Regarding dark adaptation, samples were kept under dark conditions at 4°C for more than 72 h, whereas those for light adaptation were illuminated for 3 min at 520±10 nm, with light power being adjusted to approximately 10 mW cm -2 . The molar compositions of the retinal isomers were calculated from the areas of the peaks in HPLC patterns monitored at 360 nm using the extinction coefficients of retinal oxime isomers as previously described (Kojima et al., 2020a). In pH titration experiments, samples were suspended in Buffer A. The pH values of the samples were adjusted to the desired acidic values by the addition of HCl, after which absorption spectra were measured at each pH value. All measurements were conducted at room temperature (approximately 25°C) under room light. After these measurements, the reversibility of spectral changes was examined to confirm that the sample was not denatured during measurements. Absorption changes at specific wavelengths were plotted against pH values and plots were fit to the Henderson-Hasselbalch equation assuming a single pK a value as previously described (Inoue et al., 2018).
The transient time-resolved absorption spectra of the purified proteins from 380 to 700 nm at 5-nm intervals were obtained using a homemade computer-controlled flash photolysis system equipped with an Nd: YAG laser as an actinic light source. Using an optical parametric oscillator, the wavelength of the actinic pulse was tuned at 510 nm for PngR. Pulse intensity was adjusted to 2 mJ per pulse. All data were averaged to improve the signal-to-noise ratio (n=30). All measurements were conducted at 25°C. In these experiments, samples were suspended in Buffer A. After measurements, the reproducibility of data was checked to confirm that the sample was not denatured during measurements. To investigate proton uptake and release during the photocycle, we used the pH indicator pyra-nine (final concentration=100 μM; Tokyo Chemical Industry), which has been extensively used to monitor light-induced pH changes in various rhodopsins. pH changes in the bulk environment were measured as the absorption changes of pyranine at 450 nm. The absorption changes of pyranine were obtained by subtracting the absorption changes of samples without pyranine from those of samples with pyranine. Experiments using pyranine were performed in an unbuffered solution containing 1 M NaCl and 0.05% (w/v) DDM (pH 7.0) to enhance signals. The results of 1,000 traces were averaged to improve the signal-to-noise ratio.

Subcellular localization of PngR in the model diatom
The PngR:eGFP recombinant gene, coding the full length of PngR C-terminally tagged with eGFP, was cloned into the expression vector for the model diatom P. tricornutum, pPha-NR (Stork et al., 2012), by CloneEZ (GenScript) according to the manufacturer's instructions. The plasmid was electroporated into cells of P. tricornutum UTEX642 with the NEPA21 Super Electroporator (NEPAGENE), and transformed cells were selected with a Zeocinbased antibiotic treatment as previously described (Miyahara et al., 2013;Dorrell et al., 2019). Selected clones were observed under an Olympus BX51 fluorescent microscope (Olympus) equipped with an Olympus DP72 CCD color camera (Olympus). The nucleus stained with DAPI and chlorophyll autofluorescence from the plastid were observed with a 420-nm filter at 330 to 385 nm excitation. GFP fluorescence was detected with a 510-to 550-nm filter at 470 to 495 nm excitation.

Quantitative model of C concentrations in a diatom: Cell Flux Model of C Concentration (CFM-CC)
Membrane transport model. We combined the membrane transport of CO 2 and C fixation. Parameter definitions, units, and values are provided in Supporting Information Table S1 and S2, respectively. The key model equation is the balance of the concentration of CO 2 in the cytosol, [CO 2 ] p : where t is time, D is the diffusion coefficient, and [CO 2 ] m is the concentration of CO 2 in the inner side of the outermost membrane of the plastid (hereafter "the middle space"). The first term represents the diffusion of CO 2 from the middle space to the cytosol, while the second term V Cfix represents the C fixation rate following Michaelis-Menten kinetics (Berg et al., 2010;Hopkinson, 2014): where V max is the maximum CO 2 fixation rate and K is the half saturation constant.
Solving this equation for CO 2 p 2 leads to: Note that the other solution for the negative route is unrealistic because it may lead to the overall negative value of [CO 2 ] p . Once we obtain [CO 2 ] p , we may then calculate the rate of C fixation V Cfix with [eq. 2].

Rhodopsin sequences and phylogenetic analysis
We performed a phylogenetic analysis using the rhodopsin (named PngR, accession no. AJA37445.1) of the diatom P. granii and microbial rhodopsin sequences reported to date (Marchetti et al., 2015). This phylogenetic tree revealed that PngR is not included in the proteorhodopsin (PR) clade commonly found in oceanic organisms, but belongs to the Xanthorhodopsin (XR)-like rhodopsin (XLR) clade, which is presumed to have an outward proton transporting function ( Fig. 1 and Extended Data Fig. 1). A comparison of the motif sequences necessary for ion transport showed that the amino acids in the putative proton donor and acceptor sites of XR and PR were conserved in PngR, suggesting that PngR functions as an outward proton pump (Extended Data Fig. 2). Furthermore, the homology search for rhodopsin sequences in the XLR clade from Marine Microbial Eukaryote Transcriptome Sequencing Project (MMETSP) revealed that not only diatoms (Ochrophyta, Stramenopiles), but also dinoflagellates (Dinophyceae, Alveolata) and haptophytes have rhodopsin genes in the same XLR clade (Supporting Information Table S3). These results indicate that rhodopsins of the XLR clade are widely distributed among the major phytoplankton groups, which are important primary producers in the ocean.

Function and spectroscopic features of diatom rhodopsin
To characterize the function of PngR, we heterologously expressed the synthesized rhodopsin gene in E. coli cells. A light-induced decrease in pH was observed in the suspension of E. coli cells expressing PngR, and this reduction was almost completely abolished in the presence of the protono-phore carbonyl cyanide m-chlorophenylhydrazone (CCCP) (Fig. 2A). The pH changes observed clearly showed that PngR exported protons from the cytoplasmic side across the cell membrane.
We then examined the spectroscopic characteristics of PngR using the recombinant protein purified from E. coli. The absorption maximum of PngR was located at 511 nm (Fig. 2B), which was markedly shorter than those of XR (565 nm) and GR (Gloeobacter rhodopsin 541 nm) in the XLR clade (Balashov et al., 2005). It is important to note that while P. granii is a marine species, XR and GR are both distributed in terrestrial organisms. Therefore, the present results are consistent with the shorter wavelength of the absorption maximum of rhodopsin in marine environments than in the terrestrial environment (Man et al., 2003), indicating that PngR is well adapted to light conditions in the ocean, particularly the open ocean.
We then examined the retinal configuration in PngR by HPLC. In light-and dark-adapted samples, the isomeric state of retinal was predominantly all-trans (Extended Data Fig. 3), which was similar to the isomeric state of retinal in prokaryotic GR in the XLR clade, but different from that in BR (Miranda et al., 2009). Since the pK a value of the proton acceptor residue (Asp85 in BR) is an indicator of the efficiency of proton transport by rhodopsin, we estimated the pK a values of the putative proton acceptor in PngR (Asp91) by a pH titration experiment (Extended Data Fig. 4). This experiment estimated that the pKa of this residue acceptor was approximately 5.0, indicating that the proton acceptor of PngR works well in marine and intracellular environments. Furthermore, the photochemical reactions that proceed behind the ion-transportation mechanism of PngR were examined by a flash photolysis analysis (Extended Data Fig.  5). All photocycles required for ion transportation in PngR were completed in approximately 300 ms, suggesting that the cycle was sufficiently fast to pump protons in a physiologically significant time scale. The results and a discussion of the flash photolysis analysis are described in the supplementary information.

Subcellular localization of PngR in a model diatom
The PngR sequence bears neither an apparent N-terminal extension nor a detectable N-terminal signal peptide, and, thus, in silico analyses are unable to predict the subcellular localization of PngR. To identify the subcellular localization of PngR, a C-terminal eGFP-fusion PngR was expressed in the model diatom P. tricornutum, which may be transformed by electroporation and is often used in a heterologous expression analysis (Nakajima et al., 2013;Dorrell et al., 2019). The transformed P. tricornutum cell was examined under differential interface contrast and epifluorescent microscopes (Fig. 3A). We observed the fluorescence of GFP, DAPI, and chlorophylls to establish the localization of recombinant PngR:eGFP, the nucleus, and chloroplast, respectively, in multiple cells (Extended Data Fig. 6 and 7). The fluorescence signal of the PngR:eGFP transformant appeared to localize at the periphery of chlorophyll fluorescence and DAPI signals, corresponding to the outermost plastid membrane, called the chloroplast endoplasmic reticulum membrane (CERM), which is physically connected to the nuclear membrane (Fig. 3A). A few cells also exhibited GFP signals within vacuolar membranes in addition to CERM (Extended Data Fig. 7). The insertion of the complete sequence of the PngR:eGFP gene in transformant DNA was confirmed by PCR followed by Sanger sequencing.
Based on the results of the heterologous expression experiment and microscopic observations, we concluded that PngRs primarily localized to the outermost membrane of the plastid. However, fluorescence signals were also observed to a lesser extent in the vacuolar membrane, suggesting the involvement of other factors, such as cell growth conditions, in their localization. These results imply that light-driven proton transport by PngR acidifies or alkalizes the inner region of CERM (Fig. 3B). Therefore, the physiological role of pH changes in this region in diatoms warrants further study. The electrochemical gradient formed by rhodopsin may be a driving force for various secondary transport processes. Alternatively, based on the primary purpose of plastids, local pH changes may be related to photosynthesis. The pH in this region is considered to be important for the transport of inorganic carbon (Ci) to ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) (Gee and Niyogi, 2017). This is because in the carbonate system, pH affects the proportion of carbonate species (CO 2 , HCO 3 -, and CO 3 2-) in water. Under weakly alkaline conditions in the ocean, the majority of dissolved inorganic carbon (DIC) is generally present in the form of HCO 3 -, with only approximately 1% being present in the form of CO 2 . However, RuBisCO localized in the stroma only reacts with Ci in the form of CO 2 , not HCO 3 -. The RuBisCO enzyme in diatoms exhibits low affinity even for CO 2 (Km of 25~68 μM, while CO 2 aq in the ocean is approximately 10 μM at 25°C) and, thus, requires concentrated CO 2 for efficient fixation at the site of RuBisCO. In other words, the ocean is always a CO 2 -limited environment for most phytoplankton (Riebesell et al., 1993). Consequently, due to the membrane impermeability of HCO 3 -, phytoplankton have developed a number of CO 2concentrating mechanisms (CCM) to efficiently transport Ci to the site of RuBisCO by placing HCO 3 -transporters in appropriate membranes and carbonic anhydrase (CA) in these compartments, the latter of which catalyzes the rapid interconversion between HCO 3 -and CO 2 . However, since difficulties are associated with directly examining pH changes and the forms of Ci of the small compartment in eukaryotic microbial organelles, a model simulation is a powerful alternative approach (Hopkinson et al., 2011). In the present study, we used a model simulation to investigate whether rhodopsin-mediated pH changes in this region were involved in CCM.

A quantitative model of carbon concentrations in diatoms: CFM-CC
Our subcellular localization analysis suggested that proton transport by rhodopsin acidified or alkalized the inner side of the outermost membrane of the plastid (the middle space). To quantitatively examine the effects of pH in the middle space on C fixation, we developed a simple quantitative model of carbonate chemistry combined with membrane transport and C fixation (CFM-CC: Cell Flux Model of C Concentration) (Fig. 4 upper panel). A comprehensive model of the concentration of CO 2 within diatoms was developed (Hopkinson et al., 2011;Hopkinson, 2014). CFM-CC uses a conceptually similar structure to this model, focusing on more specific membrane layers, designed to test the effects of pH changes in the middle space.
Our model results showed that the concentrations of CO 2 in the middle space ([CO 2 ] m ) were strongly dependent on pH (pH m ), suggesting that proton pumping by rhodopsin affected C fixation (Fig. 4 bottom panel). The calculation of C chemistry in the middle space revealed that a decrease in The C fixation rate relative to that with pH in the middle space of 7.59, the only mean value we found for intracellular pH in a diatom (Burns and Beardall, 1987 [Burns and Beardall, 1987]). We noted that the potential leaking of CO 2 into the cytosol may change DIC in the middle space, but used a constant DIC value because this effect has not been experimentally demonstrated and is difficult to quantify due to unknown factors (e.g., the balance of DIC uptake and CO 2 leaking). At the reference point (we used pH m =7.59 [Burns and Beardall, 1987]), [CO 2 ] m was 17 μmol L -1 , but increased to 64, 410, and 870 μmol L -1 for pH m values of 7, 6, and 5, respectively (Fig. 4A). Due to increased [CO 2 ] m , the concentration of CO 2 in the plastid ([CO 2 ]p) also increased with lower pH m ; however, the level of this increase was dependent on V max /D (the ratio of the maximum C fixation rate to the diffusion constant) (Fig. 4B). When V max /D was small, the diffusion of CO 2 from the middle space to the plastid dominated the change, resulting in [CO 2 ] p similar to [CO 2 ] m . In contrast, when V max /D was large, CO 2 uptake dominated, and the effect of [CO 2 ] m on [CO 2 ] p was small.
The rate of C fixation increased with lower pH m because increased [CO 2 ] m accelerated the transport of CO 2 into the plastid (Fig. 4C). However, the magnitude of this increase depended on V max /D. The model showed that the effects of pH m on C fixation were greater when V max /D was large because the rate of C fixation was changed more by [CO 2 ] m [eq. 6], which is directly affected by pH m . However, when V max /D was small, the C fixation rate was changed more by C uptake kinetics [eq. 4], which were saturated at a relatively low [CO 2 ] p (K value of 44 μmol L -1 [Jensen et al., 2020]). Based on the possible range of V max /D, C fixation showed 2.1-to 3.8-, 3.2-to 24.2-, and 3.4-to 51.2-fold increases when pH m decreased from 7.59 to 7, 6, 5 respectively. These results suggested that pH m markedly affected C uptake at any V max /D as well as also the benefit for cells to have high V max relative to the diffusivity of CO 2 across the membrane. This was most likely the case because the D value was shown to be reduced when there were multiple membranes (Eichner et al., 2019;Inomura et al., 2019). Therefore, this simple yet elegant system with rhodopsin to manipulate pH m provides a powerful mechanism in C concentrations and, thus, adjusts the C fixation rate to given physiological conditions in some rhodopsin-containing diatoms, enabling them to be more successful primary producers in the ocean.
In our model simulation, we examined the effects of rhodopsin-mediated pH changes in the middle space on CCM efficiency. The results obtained suggested that C fixation was enhanced when the pH of the middle space was acidified by a light-driven proton pump. CCM based on CO 2 diffusion (termed the pump-leak type) has been proposed as a possible mechanism by placing CAs in appropriate locations. For example, Nannochloropsis oceanica (Ochrophyta), possessing the same four membrane-bound complex plastids as those found in diatoms, is considered to generate CO 2 by placing CA in the middle space (Gee and Niyogi, 2017). Furthermore, the centric diatom Chaetoceros gracilis is considered to generate CO 2 by placing CAs outside the cell and allowing CO 2 to flow into the cell (Tsuji et al., 2021). In contrast to the CA-based model, the acidification-based model was formerly proposed to facilitate the CO 2 fixation of RuBisCO in the thylakoid lumen of plastids; HCO 3 -is converted into CO 2 through acidification by photosynthetic proton pumping into the thylakoid lumen (Raven, 1997). Our rhodopsin-mediated Ci transform model, CFM-CC proposes that pH changes in the middle space by protonpumping rhodopsin also plays the role of a CO 2 regulator. This proposed mechanism may be useful in most parts of the ocean where CO 2 chronically limits photosynthesis, but may be even more valuable in specific environments. For example, since CA, which plays a central role in CCM, requires cobalt or zinc ions as the reaction center, and photosynthetic proton-pumping systems need iron, rhodopsinderived acidic pools may be useful for Ci uptake in oceans where these metal ions are depleted (such as the HNLC region of the North Pacific Ocean) (Moore et al., 2013). In the HNLC region of the North Pacific, where P. granii rhodopsin-containing cells were initially identified (Marchetti et al., 2012), primary production may be limited by iron and affected by other trace metals (Saito et al., 2008). In other words, our proposed mechanism appears to be particularly effective in the ocean where trace metals involved in CCM are depleted.
In the present study, we clarified the function and subcellular localization of PngR in a photosynthetic diatom. The results obtained suggest that proton transport by rhodopsin changes pH inside the outermost membrane of the plastid (CERM). A quantitative simulation indicated that the creation of an acidic pool by light provides positive feedback on C fixation efficiency, while alkalization of the middle space may restrict C fixation. If PngR acidifies the middle space, diatom rhodopsin may contribute to CCM (Extended Data Fig. 8). Future analyses of cultured rhodopsin-bearing microbial eukaryotes will corroborate the present results and promote further research on the mechanisms by which rhodopsin-mediated proton transport promotes their growth in the ocean.