Abstract
Microbial rhodopsins are prevalent in many cyanobacterial groups as a light-energy-harvesting system in addition to the photosynthetic system. It has been suggested that this dual system allows efficient capture of sunlight energy using complementary ranges of absorption wavelengths. However, the diversity of cyanobacterial rhodopsins, particularly in accumulated metagenomic data, remains underexplored. Here, we used a metagenomic mining approach, which led to the identification of a novel rhodopsin clade unique to cyanobacteria, cyanorhodopsin-II (CyR-II). CyR-IIs function as light-driven outward H+ pumps. CyR-IIs, together with previously identified cyanorhodopsins (CyRs) and cyanobacterial halorhodopsins (CyHRs), constitute cyanobacterial ion-pumping rhodopsins (CyipRs), a phylogenetically distinct family of rhodopsins. The CyR-II clade is further divided into two subclades, YCyR-II and GCyR-II, based on their specific absorption wavelength. YCyR-II absorbed yellow light (λmax = 570 nm), whereas GCyR-II absorbed green light (λmax = 550 nm). X-ray crystallography and mutational analysis revealed that the difference in absorption wavelengths is attributable to slight changes in the side chain structure near the retinal chromophore. The evolutionary trajectory of cyanobacterial rhodopsins suggests that the function and light-absorbing range of these rhodopsins have been adapted to a wide range of habitats with variable light and environmental conditions. Collectively, these findings shed light on the importance of rhodopsins in the evolution and environmental adaptation of cyanobacteria.
Introduction
Almost all biological processes on Earth are driven by the energy supplied by sunlight. Chlorophyll-based photosystems convert sunlight energy into chemical energy through photosynthetic reactions, and this energy flows into ecosystems in the form of organic compounds. Cyanobacteria, one of the most abundant photosynthetic microorganisms utilizing chlorophyll-based photosystems, inhabit almost any light-exposed environment [1, 2]. This boundless environmental adaptation is supported by light harvesting through their highly diverse chlorophyll-based photosystems. For example, cyanobacterial lineages have differentiated absorptive chlorophyll pigments present in photosystem I and II reaction centers [3–6]. In addition, various lineages employ a light-harvesting antenna known as the phycobilisome. The phycobilisome is efficiently used by various lineages to capture available wavelength in the environment, such as marine Synechococcus strains that use phycobilisome to capture blue-to-green light available in their habitats [7, 8].
Microbial rhodopsins, another light-utilizing system, are distributed in broad bacterial taxa, including cyanobacteria [9–13]. Rhodopsin is a photoreceptor membrane protein that binds retinal as the sole chromophore, a simpler light-utilizing system than the chlorophyll-based photosystem, which consists of many proteins and typically multiple chromophores. All rhodopsins are activated by visible light and return to their original state through a photocycle that forms various intermediates (called photo-intermediates) [14]. During the photocycle, they exhibit their cognate protein functions such as ion transport [15]. In contrast, absorption wavelengths of the retinal chromophore are not homogeneous. For example, bacteriorhodopsins (BRs) of halophilic archaea, light-driven outward H+ pumping rhodopsins, absorb yellow light (λmax = 570 nm) [16]. Proteorhodopsins (PRs) of marine bacteria are light-driven outward H+ pumping rhodopsins that are divided into two types of absorption maxima: 490 nm (blue-absorbing PR, BPR) [17] and 500–525 nm (green-absorbing PR, GPR) [18]. The former is relatively abundant in blue-light enriched deep/pelagic waters and the latter green-light-enriched surface/coastal waters [19, 20]. The difference in absorption maxima is thought to be the result of adaptation to the light conditions of their habitats.
Two cyanobacteria-specific rhodopsin clades, namely cyanobacterial halorhodopsin (CyHR) and cyanorhodopsin (CyR), have been identified as light-driven inward Cl− and outward H+ pumping rhodopsin clades, respectively [21]. The discovery of these rhodopsin clades unique to cyanobacteria suggests that the rhodopsin-mediated and chlorophyll-based photosystems have long coexisted in cyanobacterial cells. These photosystems have complementary absorption wavelength ranges, allowing for harmonized capture of sunlight energy in cyanobacterial cells [21]. However, due to the limited rhodopsin explorations focusing on cultured strains [9, 11–13, 21], the distribution of rhodopsin genes within cyanobacterial lineages and the phylogenetic diversity of cyanobacterial rhodopsins remain underexplored, considering the accumulated metagenomic data such as numerous metagenome-assembled genomes (MAGs) [22].
In this study, we explored rhodopsin sequences in both cyanobacterial genomes and metagenomic assemblies. Phylogenetic analysis of the detected rhodopsin sequences led to the identification of three cyanobacteria-specific rhodopsin clades, including a new clade Cyanorhodopsin-II (CyR-II). CyR-II was characterized with the predominance of uncultured cyanobacteria. Spectroscopic, X-ray structural, and mutational analyses demonstrate that the rhodopsins in the clade found in this study have red-shifted absorption maxima due to slight changes in the side-chain structure.
Materials and methods
Rhodopsin genes of cyanobacteria were identified as follows. Cyanobacterial genome assemblies and their coding sequences were downloaded from the National Center for Biotechnology Information (NCBI) and the US Department of Energy Joint Genome Institute (JGI) Integrated Microbial Genomes (IMG) database. In addition, the OceanDNA MAG catalog [22] and metagenomic assemblies derived from metagenomes published in previous studies [23–26] were also used to explore the rhodopsin genes. Genes encoding rhodopsins were explored using two hidden Markov models (Supporting Data S1 and S2) [27] constructed from large-scale sequence alignments of known rhodopsins and their homologs. The phylogenetic trees of rhodopsins and cyanobacterial genomes were reconstructed using IQ-TREE (v. 1.6.12) [28] by 1000 ultrafast bootstrap searches [29]. Habitat and morphological information was collected manually: habitat information is based on “isolation source” of genome assembly in NCBI or the previous report [30] and morphological classifications are based on the previous report [31]. Data files of sequence alignments and phylogenies of rhodopsin and cyanobacterial genomes are available (Supporting Data S3–S8).
The codon-optimized DNA fragments to Escherichia coli encoding the P7104R (Nodosilinea nodulosa PCC 7104 rhodopsin, Accession No. WP_017301364.1), CBR35R (Chroococcidiopsidaceae cyanobacterium CP_BM_RX_35, MBV8882851.1), and MAG18R (uncultured cyanobacterium SRR6869040_bin.18, SRR6869040_N0001714_12) genes were chemically synthesized and inserted into the pET21a (+) plasmid vector by Eurofins Genomics (Japan). The ion transport activities of the CyR-IIs were examined by monitoring the light-induced pH changes in suspensions of CyR-IIs-expression E. coli cells. Ion transport measurements were repeated under the same conditions after the addition of the proton-selective ionophore, 30 μM carbonyl cyanide m-chlorophenylhydrazone (CCCP; Sigma-Aldrich, USA).
For spectroscopic analyses, the His-tagged CyR-II proteins (P7104R, CBR35R, and MAG18R) were expressed in E. coli C41 (DE3) (Lucigen, USA). Cells were solubilized with 1.0% (w/v) n-dodecyl-β-d-maltoside (DDM, Dojindo Lab., Japan). The solubilized-fraction was purified by Ni2+ affinity column chromatography. All UV–Vis spectra were measured using a UV-2450 spectrophotometer (Shimadzu, Japan).
The retinal isomer composition was determined by using high-performance liquid chromatography as previously described [21, 32]. The pH of the purified P7104R was adjusted to the desired values (1.11–11.17) to ensure that no denaturation occurred. The absorption spectrum (250–750 nm) was measured for samples under the various pH conditions. The pKa value or values were then estimated by fitting the data to the Henderson–Hasselbalch equation with two (pKa1 and pKa2) or one pKa.
For the flash-photolysis experiment, crude membranes of P7104R were obtained by disruption by sonication on ice-cold water and ultracentrifugation. Time-resolved absorption spectra from 370 to 720 nm at 5-nm intervals in purified P7104R and from 390 to 720 nm at 10-nm intervals in E. coli membrane expressing P7104R were measured by using a computer-controlled flash-photolysis system. The wavelength of the actinic light pulse was tuned to 565 nm. The data at 565 and 570 nm in purified P7104R were excluded from the analysis to account for the scattering of the light pulses. To monitor proton uptake and release during the photocycle, we repeated the flash-photolysis experiment using purified P7104R under the same conditions after the addition of the pH indicator pyranine (final concentration = 100 μM; Tokyo Chemical Industry Co., Ltd., Japan), which is a pH-sensitive fluorochrome.
The P7104R protein used for crystallization was synthesized by using an E. coli cell-free protein synthesis system according to previously reported protocols [33–36] and crystallized by the in meso method. Diffraction data for P7104R were collected at the BL32XU beamline of the SPring-8 synchrotron by using the multiple small-wedge scheme implemented in the ZOO system [37, 38]. Data processing was performed using the KAMO program [39], and the structure was solved by molecular replacement using the Phaser program [40] in the Phenix suite [41]. The search model was the BR [42], and the structure was manually rebuilt with the Coot program [43].
Results
Collection of cyanobacterial rhodopsins and their phylogeny
We explored rhodopsin genes through a homology search against cyanobacterial genomes in databases and cyanobacteria-affiliated contigs assembled from publicly available metagenomic data. A total of 102 rhodopsin genes were identified, 96 of which were from genomes and 6 from metagenomic contigs. Phylogenetic analysis was performed using the 102 cyanobacterial rhodopsins in addition to noncyanobacterial rhodopsins. The 91 rhodopsins were classified into five previously reported clades, including those derived from cyanobacteria: xanthorhodopsin-like rhodopsin (XLR, 3 genes), Na+ pumping rhodopsin (NaR, 6 genes), xenorhodopsin (XeR, 35 genes), CyHR (31 genes), and CyR (16 genes). The remaining 11 rhodopsins constitute a new rhodopsin clade with high bootstrap support (100%) (Figs 1A and S1). We name the clade CyR-II, which contains 10 metagenome-derived rhodopsins (four from MAGs and six from metagenomic contigs), and only one rhodopsin derived from an isolated strain (Nodosilinea nodulosa PCC 7104). The CyR-II clade was divided into two subclades, YCyR-II and GCyR-II, with high bootstrap support (99% for both). These two subclades have characteristic environmental distributions: YCyR-II mainly from sediment and soil of terrestrial environments, and GCyR-II from bacterial mats and biofilms of marine environments (Fig. 1B). The three clades (CyHR, CyR, and CyR-II) collectively constitute a phylogenetically distinct family of cyanobacteria-specific rhodopsins, which is designated as “cyanobacterial ion-pumping rhodopsins” (CyipRs) in this study.
![Phylogeny of cyanobacterial rhodopsins and distribution among cyanobacterial lineages: (A) A maximum likelihood tree of amino acid sequences of 102 cyanobacterial and 69 non-cyanobacterial rhodopsins; rhodopsin clades are indicated as follows: NaR, ClR (Cl− pumping rhodopsin), XLR, PR, SR (sensory rhodopsin I and sensory rhodopsin II), BR, HR (halorhodopsin), DTG-motif rhodopsin, XeR, CyHR, CyR, and an uncharacterized cyanobacteria-specific clade (CyR-II); CyipRs are phylogenetically distinct cyanobacteria-specific rhodopsins consisting of CyR-II, CyR, and CyHR; the green branches indicate cyanobacterial rhodopsins, and the black branches indicate others; bootstrap probabilities (≥50%) are indicated by colored circles; the scale bar represents the number of substitutions per site; (B) An enlarged view of the novel cyanobacteria-specific clade CyR-II with habitat and source type information; the clade is divided into two subclades: YCyR-II (yellow-absorbing CyR-II) and GCyR-II (green-absorbing CyR-II); the circles in front of the sequence names were colored based on cyanobacterial lineages; names starting with [CON] and [MAG] indicate that the sequence is from the metagenomic contig and MAG, respectively; (C) distribution of rhodopsins in cyanobacterial lineages; the phylogenomic tree was constructed by maximum likelihood estimation based on conserved amino acid sites of 120 ubiquitous single-copy genes of bacteria and collapsed by lineages; “genomes” indicates the number of genomes per lineage, and “rhodopsin-harboring genomes” indicates the number of genomes containing rhodopsin; the numbers within the dashed squares indicate the fraction of genomes that contain each clade of rhodopsins in comparison with the total number of genomes in a lineage.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ismej/18/1/10.1093_ismejo_wrae175/2/m_wrae175f1.jpeg?Expires=1747873798&Signature=gNukJ-V1RX3V5ytdlIBzebFTyZpqjUzec7JApudzVhi4VVS115OZGr0OtrUCNiwj1fJYhCRca-zk1cThtJh80fNPU2BQnw6y9r4yzMDnN4IVlc8DFPihmX7aN2ED~p0JrPEvtK~AN32F13FFc2U40PxlSCgZLWSDpzw-Z1Q2lEwBPAriADnw-oDBzsRge3gD7oP3NsbS~YvycPVzU6BEQrNqTIjtqt25R859J020wTQKmcRdNrm22RauqMWG8ovXi2Z0zWHhPllao2LiOvB2frZED33hPop-G8HIb7U95bBMn8t8G~cPp14fsSMFK6NgwOMDlUw-f7QAzLB0rTT3cg__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Phylogeny of cyanobacterial rhodopsins and distribution among cyanobacterial lineages: (A) A maximum likelihood tree of amino acid sequences of 102 cyanobacterial and 69 non-cyanobacterial rhodopsins; rhodopsin clades are indicated as follows: NaR, ClR (Cl− pumping rhodopsin), XLR, PR, SR (sensory rhodopsin I and sensory rhodopsin II), BR, HR (halorhodopsin), DTG-motif rhodopsin, XeR, CyHR, CyR, and an uncharacterized cyanobacteria-specific clade (CyR-II); CyipRs are phylogenetically distinct cyanobacteria-specific rhodopsins consisting of CyR-II, CyR, and CyHR; the green branches indicate cyanobacterial rhodopsins, and the black branches indicate others; bootstrap probabilities (≥50%) are indicated by colored circles; the scale bar represents the number of substitutions per site; (B) An enlarged view of the novel cyanobacteria-specific clade CyR-II with habitat and source type information; the clade is divided into two subclades: YCyR-II (yellow-absorbing CyR-II) and GCyR-II (green-absorbing CyR-II); the circles in front of the sequence names were colored based on cyanobacterial lineages; names starting with [CON] and [MAG] indicate that the sequence is from the metagenomic contig and MAG, respectively; (C) distribution of rhodopsins in cyanobacterial lineages; the phylogenomic tree was constructed by maximum likelihood estimation based on conserved amino acid sites of 120 ubiquitous single-copy genes of bacteria and collapsed by lineages; “genomes” indicates the number of genomes per lineage, and “rhodopsin-harboring genomes” indicates the number of genomes containing rhodopsin; the numbers within the dashed squares indicate the fraction of genomes that contain each clade of rhodopsins in comparison with the total number of genomes in a lineage.
To investigate the evolutionary process of rhodopsin acquisition and deletion in cyanobacterial genomes, we performed a phylogenomic reconstruction of cyanobacteria. A phylogenomic tree based on 120 single-copy marker genes confirmed that cyanobacteria are divided into seven lineages (A–G) that are referred to as subclades in previous studies [21, 44], with the slight difference that the lineage of Leptolyngbya ohadii IS1 is shifted from C to D (Figs 1C and S2). Lineage C was further divided into three lineages with high bootstrap support (C-I–III). Gloeobacter violaceus PCC 7421T, thought to be the earliest branching cyanobacterium containing an XLR [45], does not belong to any lineage. No rhodopsin is detected in genomes of lineages A, F, and G, whereas genomes of lineages B, C-I, and D contains various clades of rhodopsins. Genomes containing CyR-II belong to lineages B (two genomes) and C-II (three genomes).
Sequence and functional characterization of CyR-IIs
Specific amino acid residues (motif sequences) crucial for ion transport activity were compared after all CyR-IIs with known rhodopsins (Fig. 2A pink background and Fig. S3 for representative protein sequences). CyR-IIs have two motif types in the third helix (helix C), Asp-Thr-Ser (DTS motif, seven genes, P7104R and MAG18R in Fig. 2A) and Asp-Thr-Asp (DTD motif, four genes, CBR35R in Fig. 2A). These correspond to the Asp85, Thr89, and Asp96 (DTD motif) of BR. This indicates that CyR-IIs possess a putative proton acceptor, Asp residue (corresponding to Asp85BR). However, the amino acid residue of the putative proton donor, Asp or Glu residue (Asp96BR), differed between the two motif types. Rhodopsin binds to retinal chromophore by forming a Schiff base linkage through the Lys residue (corresponding to Lys216BR) in the seventh helix (helix G), which is stabilized by a negatively charged counterions (corresponding to Asp85BR and Asp212BR). Because the Lys residue (such as Lys215P7104R, italic bold in Fig. 2A) in the helix G and two putative counterions (such as Asp85P7104R and Asp 211P7104R) were conserved in all sequences other than SRS3241718_N0233988_1 (see Figs 1B and S1); almost all CyR-IIs probably can make a stable Schiff base linkage between the rhodopsin protein and the retinal chromophore and can transport protons.
Characterization of P7104R. (A) Alignment of critical amino acids of light-driven outward H+ pumping rhodopsins and cyanobacterial rhodopsins. Rhodopsin names and clades (also see Fig. 1A) are indicated as follows: P7104R (Nodosilinea nodulosa PCC 7104 rhodopsin, CyR-II), CBR35R (Chroococcidiopsidaceae cyanobacterium CP_BM_RX_35 rhodopsin, CyR-II), MAG18R (uncultured cyanobacterium SRR6869040_bin.18 rhodopsin, CyR-II), N2098R (Calothrix sp. NIES-2098 rhodopsin, CyR), N4075R (Tolypothrix sp. NIES-4075 rhodopsin, CyR), MrHR (Mastigocladopsis repens halorhodopsin, CyHR), SyHR (Synechocystis halorhodopsin, CyHR), ASR (Anabaena sensory rhodopsin, XeR), BR (bacteriorhodopsin, BR), PR (proteorhodopsin, PR), and GR (Gloeobacter rhodopsin, XLR). The “No.” indicates the number of amino acids counted from the first amino acid (“start codon,” also see Fig. S3). The known function of displayed amino acids are as follows: primary proton acceptor (Asp85BR), proton donor (Asp96BR), proton release group (Glu194BR and Glu204BR), counterion (Asp212BR), and Schiff base (Lys216BR). Two carboxylates, Asp (D) and Glu (E), are shown in blue and green, respectively, and Schiff base Lys (K) is shown in red. Specific amino acid residues (motif sequences) crucial for ion transport activity were shown by pink background (No. 85, 89, and 96 in P7104R). The Lys residue (Lys216BR), which forms a Schiff base linkage with the retinal, was highlighted by yellow background. (B) Light-induced changes of the pH of suspensions of E. coli expressing P7104R, CBR35R, and MAG18R. The changes in pH in the absence (solid line) and presence (broken line) of the proton-selective ionophore, 30 μM CCCP, are shown. The measurements were performed under the dark conditions (gray shading) with illumination at 520 ± 10 nm for 3 min (white shading, from 180 sec to 300 sec).
To validate the prediction that CyR-IIs function as H+ pumps, three rhodopsin genes (P7104R, CBR35R, and MAG18R genes) were used for heterologous expression analysis. A light-induced pH decrease was observed in the suspension of the E. coli cells expressing each CyR-II (Fig. 2B, solid line). These pH changes were almost completely abolished in the presence of the protonophore CCCP (Fig. 2B, broken line), which collapses the proton motive force across the membrane. These results suggested that CyR-IIs export protons from the cytoplasmic side across the cytoplasmic membrane.
Spectroscopic characterization of P7104R
Spectroscopic differences between CyR and CyR-II were also investigated. The absorption maxima of dark-adapted purified P7104R, CBR35R, and MAG18R were located at 570, 565.5, and 546.5 nm, respectively (Fig. 3A). The present rhodopsin phylogeny revealed that the first and second belong to the sediment and soil group, and the third belongs to the marine bacterial mat and biofilm group. The two groups of CyR-II were therefore characterized by utilization of yellow (around 570 nm) or green lights (550 nm), respectively. The former can utilize longer wavelength light than the latter and CyRs. We thus named the sediment and soil group Yellow-absorbing CyR-II (YCyR-II) and the bacterial mats and biofilms group Green-absorbing CyR-II (GCyR-II).
Absorption spectra of CyR-IIs and structure of P7104R. (A) UV–Vis spectra of three CyR-IIs, P7104R (purple line), CBR35R (green line), and MAG18R (red line). The location of amino acids around the retinal chromophore in P7104R (B) and N2098R (C, CyR of Calothrix sp. NIES-2098, PDB code 6LM0). UV–Vis spectra of P7104R mutants (D), rearranged in putative critical amino acids contributing to color-tuning.
We then performed detailed spectroscopic analyses using P7104R because it has the longest absorption maximum among the three CyR-IIs. Although the absorption maximum of light- and dark-adapted BR have been known to differ [16, 46], the absorption characteristic of P7104R did not vary between experiments with (light-adapted) and without (dark-adapted) illumination (Table 1 and Fig. S4A). The retinal configuration of P7104R was estimated that predominantly all-trans isomeric state of retinal (75.8% in dark-adapted and 88.0% in light-adapted) based on the areas of peaks in the HPLC patterns using absorption coefficients of retinal isomers (Fig. S4B). This differed from that of the haloarchaeal BR but resemble that of other bacterial rhodopsins, including cyanobacterial CyR (N2098R) and GR (Table 1).
Photochemical properties of light-driven outward proton pump rhodopsins.
| Opsin type | λmax (nm) | Retinal configuration | M-decay rate constant | pKa value | Refs |
|---|---|---|---|---|---|
| P7104R | 570 | All-trans (predominant) | ~0.0009 ms−1 | <1.1 and ~6.6 (Asp85) ~10.8 (Lys215) | This study |
| N2098R | 550a | All-trans (predominant)a | ~0.016 ms−1a | <2.0 (Asp74)a ~10.7 (Lys204)a | [21]a |
| GR | 544b | All-trans (predominant)c,d | ~2.3 ms−1c, ~1 ms−1e | ~5.9 (Asp121)b ~9.12 (Lys257)b | [47]b, [10]c, [48]d, [49]e |
| BR | 570f | All-trans (~50%)g 13-cis (~50%)g | ~0.25 ms−1h | ~2.6 (Asp85)i ~13.3 (Lys216)j | [16]f, [46]g, [50]h, [51]i, [52]j |
| PR (GPR) | 520k | All-trans (predominant)l | ~4 ms−1l | ~7.9 (Asp97)m ~11.3 (Lys227)n | [18]k, [53]l, [54]m, [55]n |
| PR (BPR) | 490o | N.D. | N.D. | ~6.2 and ~7.9 (Asp97)p N.D. (Lys227) | [17]o, [56]p |
| Opsin type | λmax (nm) | Retinal configuration | M-decay rate constant | pKa value | Refs |
|---|---|---|---|---|---|
| P7104R | 570 | All-trans (predominant) | ~0.0009 ms−1 | <1.1 and ~6.6 (Asp85) ~10.8 (Lys215) | This study |
| N2098R | 550a | All-trans (predominant)a | ~0.016 ms−1a | <2.0 (Asp74)a ~10.7 (Lys204)a | [21]a |
| GR | 544b | All-trans (predominant)c,d | ~2.3 ms−1c, ~1 ms−1e | ~5.9 (Asp121)b ~9.12 (Lys257)b | [47]b, [10]c, [48]d, [49]e |
| BR | 570f | All-trans (~50%)g 13-cis (~50%)g | ~0.25 ms−1h | ~2.6 (Asp85)i ~13.3 (Lys216)j | [16]f, [46]g, [50]h, [51]i, [52]j |
| PR (GPR) | 520k | All-trans (predominant)l | ~4 ms−1l | ~7.9 (Asp97)m ~11.3 (Lys227)n | [18]k, [53]l, [54]m, [55]n |
| PR (BPR) | 490o | N.D. | N.D. | ~6.2 and ~7.9 (Asp97)p N.D. (Lys227) | [17]o, [56]p |
The data presented, except for P7104R, refer previous studies corresponding to a–p.
Photochemical properties of light-driven outward proton pump rhodopsins.
| Opsin type | λmax (nm) | Retinal configuration | M-decay rate constant | pKa value | Refs |
|---|---|---|---|---|---|
| P7104R | 570 | All-trans (predominant) | ~0.0009 ms−1 | <1.1 and ~6.6 (Asp85) ~10.8 (Lys215) | This study |
| N2098R | 550a | All-trans (predominant)a | ~0.016 ms−1a | <2.0 (Asp74)a ~10.7 (Lys204)a | [21]a |
| GR | 544b | All-trans (predominant)c,d | ~2.3 ms−1c, ~1 ms−1e | ~5.9 (Asp121)b ~9.12 (Lys257)b | [47]b, [10]c, [48]d, [49]e |
| BR | 570f | All-trans (~50%)g 13-cis (~50%)g | ~0.25 ms−1h | ~2.6 (Asp85)i ~13.3 (Lys216)j | [16]f, [46]g, [50]h, [51]i, [52]j |
| PR (GPR) | 520k | All-trans (predominant)l | ~4 ms−1l | ~7.9 (Asp97)m ~11.3 (Lys227)n | [18]k, [53]l, [54]m, [55]n |
| PR (BPR) | 490o | N.D. | N.D. | ~6.2 and ~7.9 (Asp97)p N.D. (Lys227) | [17]o, [56]p |
| Opsin type | λmax (nm) | Retinal configuration | M-decay rate constant | pKa value | Refs |
|---|---|---|---|---|---|
| P7104R | 570 | All-trans (predominant) | ~0.0009 ms−1 | <1.1 and ~6.6 (Asp85) ~10.8 (Lys215) | This study |
| N2098R | 550a | All-trans (predominant)a | ~0.016 ms−1a | <2.0 (Asp74)a ~10.7 (Lys204)a | [21]a |
| GR | 544b | All-trans (predominant)c,d | ~2.3 ms−1c, ~1 ms−1e | ~5.9 (Asp121)b ~9.12 (Lys257)b | [47]b, [10]c, [48]d, [49]e |
| BR | 570f | All-trans (~50%)g 13-cis (~50%)g | ~0.25 ms−1h | ~2.6 (Asp85)i ~13.3 (Lys216)j | [16]f, [46]g, [50]h, [51]i, [52]j |
| PR (GPR) | 520k | All-trans (predominant)l | ~4 ms−1l | ~7.9 (Asp97)m ~11.3 (Lys227)n | [18]k, [53]l, [54]m, [55]n |
| PR (BPR) | 490o | N.D. | N.D. | ~6.2 and ~7.9 (Asp97)p N.D. (Lys227) | [17]o, [56]p |
The data presented, except for P7104R, refer previous studies corresponding to a–p.
The photochemical reactions of microbial rhodopsin are triggered by absorbing visible light and isomerization of the retinal chromophore from all-trans to the 13-cis form. Subsequent to the retinal isomerization, the rhodopsin protein forms various photo-intermediates (e.g. K, L, M, N, and O intermediates in BR) [14] and returns to its original state within a picosecond-to-second timeframe. During the photocycle, the microbial rhodopsins show their cognate protein functions such as ion transport [15]. Among the microbial rhodopsins, the H+ pumping rhodopsins vectorially transfer a proton from the intracellular to the extracellular side during a single photocycle step composed of various sequentially formed photo-intermediates [15]. Therefore, the lifetime of the intermediate is directly related to the H+ transport activity. The flash-photolysis analyses using the purified P7104R were thus performed to reveal its ion-transport mechanism. The absorption at around 570 nm, which was emerged from the absorption maximum of P7104R, was bleached but recovered within 10 000 ms (Fig. S4C). In addition, two positive peaks were observed around 415 and 675 nm: these peaks appeared within 0.1 and 0.01 ms, respectively, and they bleached within 10 000 ms. Based on the time and location of the absorption maxima, we postulated that the changes of absorbance at 415 and 675 nm were attributable to the K- and M-intermediates, respectively (Fig. S4D). The photocycle of P7104R was thus completed by about 8000 ms after completing the following reactions: the K- and M-intermediates were formed within a microsecond time frame; then, when the K-intermediate was bleached, the M-intermediate continued to be sustained; subsequently, the M-intermediate was bleached, and the rhodopsin protein returned to its original state (570 nm). This photocycle was also observed in the E. coli membrane expressing P7104R (Fig. S4E and F). To elucidate the timing of proton uptake and release during the photocycle of P7104R, the absorption changes in the presence or absence of pyranine, a pH-sensitive dye, were monitored using the purified P7104R. A decrease in the pyranine signal was observed within 10 000 ms (Fig. S4D, gray line), indicating that the solvent was acidified. Because the solvent alkalization occurred within a time interval shorter than the detection limit, no increase in the signal was observed. Furthermore, by fitting a single exponential equation to the signal, the M-decay rate constant at pH 7.0 was estimated to be 0.0009 ms−1. This result indicated that the M-intermediate of P7104R was much more long-lived than the M-intermediates of the other light-driven outward H+ pumping rhodopsins, including N2098R (Table 1).
During the H+ transport in rhodopsins, the protons (H+) are transferred through some charged residues such as Asp, Glu, and Lys inside the proteins. These charged residues hence are known to be important for H+ transport. In other words, the different properties of these charged residues also imply different modes of H+ transport. We performed pH titration experiments to estimate the pKa values of essential charged residues in P7104R (i.e. Asp85P7104R and Lys215P7104R) and compared with other rhodopsins. The pH titration data of Asp85P7104R were fitted to the Henderson–Hasselbalch equation assuming two pKa values. The pKa values were estimated to be <1.0 and 6.6 (Table 1, Fig. S4G–I). These pKa values are similar to those of BPR and GPR but not to those of other cyanobacterial rhodopsins. The pKa of Lys215P7104R was estimated to be 10.8 based on fitting titration data to the Henderson–Hasselbalch equation with a single pKa (Table 1, Fig. S4J–L). And that pKa of Lys215P7104R was equivalent to the pKa of other rhodopsins. These results suggest that CyRs and CyR-IIs function as H+ pumps, but that the specific mechanism of H+ transport has been modified during the evolution of rhodopsins.
Structure and red-tuning mechanism of P7104R
We elucidated which structural changes in the process of rhodopsin evolution cause the differences in spectroscopic properties, especially the absorption maximum. The crystal structure of cell-free-synthesized P7104R was solved at 2.07 Å to clarify the red-shifted absorption mechanism of P7104R (Table S1). The P7104R monomer was composed of seven transmembrane helices (helices A–G) as well as other rhodopsins (Fig. S5A and B). The overall structure of P7104R was more like that of N2098R (PDB ID code 6LM0, root-mean-square deviation [RMSD] 0.766 Å) [21] than BR (PDB ID code 1C3W, RMSD 1.44 Å) [42], but the structure of the β-sheet distributed outside the cell membrane was more similar to that of BR (Fig. S5A and B). A detailed comparison between the structure of P7104R and N2098R or BR (Fig. S5) showed that a pentagonal cluster near the retinal of P7104R, which is known to be important for proton transport, consisted of three water molecules adjacent to the Lys215P7104R, Asp85P7104R, and Asp211P7104R (corresponding to Lys216BR, Asp85BR, and Asp212BR). This was also like that of N2098R (Fig. S5C and D). The structure around the proton exit was also similar in arrangement, although Glu (Glu182N2098R and Glu194BR) was replaced by Lys193P7104R (Fig. S5E and F). The cyclohexane ring comprising the retinal chromophore of P7104R was more slanted than those of N2098R and BR (Figs 3B and C, S5C, D, and G–I). A comparison of the amino acids around the retinal showed that the Ile171N2098R was rearranged to a Trp182P7104R containing a larger functional group, and the distorted structure of the retinal was canceled by Trp182P7104R, through Phe181P7104R and Glu144P7104R (Fig. S3, orange bold). These three residues were also completely conserved in yellow-absorbing CBR35R but not in green-absorbing MAG18R (Fig. S3, orange bold). These results suggest that Trp182P7104R affects the structure of the retinal and contributes to making the absorption maximum longer in P7104R and CBR35R.
To validate this hypothesis, we performed mutational analyses of P7104R. The F181W, W182I, and F181W/W182I mutants of P7104R absorbed at wavelengths of 565.0, 565.5, and 557.5 nm, respectively, which were blue-shifted compared with the wild type of P7104R (λmax = 570 nm) (Fig. 3D). These results provide support for the hypothesis and indicate that Phe181P7104R and Trp182P7104R play a role in the red-shift of the absorption maximum in YCyR-IIs. In contrast, these results also suggest that a large red-shift has occurred in conjunction with alterations of other amino acid residues, not solely Phe181P7104R and Trp182P7104R.
Discussions
A survey of cyanobacterial rhodopsins in assembled genomes and metagenomic contigs was performed. Phylogenetic analysis of the identified rhodopsins led to the discovery of a new cyanobacteria-specific rhodopsin clade, CyR-II, containing predominantly metagenome-derived sequences (Fig. 1A). CyR-II is the third rhodopsin clade unique to cyanobacteria and, together with CyHRs and CyRs, forms CyipRs, phylogenetically distinct cyanobacteria-specific rhodopsins. Phylogenomic analysis revealed that CyipRs, which evolved within cyanobacteria, are widely distributed among different cyanobacterial lineages (Figs 1C and S2). The inconsistent topology between the phylogeny of cyanobacterial species and CyipRs suggests that CyipRs spread within cyanobacterial lineages through repeated horizontal gene transfer (HGT) events. As an example of CyR-IIs, Chroococcidiopsidaceae MAGs (lineage B) and Nodosilinea nodulosa PCC 7104 (lineage C-II) are phylogenomically distant and both possess YCyR-IIs, suggesting that YCyR-IIs were transferred via HGT (Figs 1B and S2). In addition, CyRs of Phormidesmis priestleyi BC 1401 (lineage D) and CyHRs of two MAGs (lineage C-II) appear to have been acquired by HGTs (Fig. S1). Other than CyipRs, HGTs are also important for rhodopsin radiation in cyanobacterial lineages. In XeRs, multiple HGTs within cyanobacteria should have occurred due to the mixed cyanobacterial lineages within the XeR clade (Figs S1 and S2). Cyanobacterial NaRs and XLRs appear to have been acquired from noncyanobacterial lineages (Figs S1 and S2) [21]. Almost all cyanobacteria are known to possess the diox1 gene for the biosynthesis of retinal, a rhodopsin chromophore [21], suggesting that cyanobacteria only need to acquire a rhodopsin gene for rhodopsin-mediated photosynthesis to function. Rhodopsin-mediated photosystem is a simpler light-utilizing system and requires roughly half of the maintenance energy per molecule of chromophore than the chlorophyll-based photosystem [57]. Thus, the evidence of repeated HGT events of rhodopsin genes in cyanobacterial lineages proposes that rhodopsin acquisition plays an important role in cyanobacterial survival.
Cyanobacteria show different rates of rhodopsin possession according to their morphology, which was previously defined by five sections (Table S2) [31]. In particular, the morphology section I (unicellular forms) shows a low rate of rhodopsin possession compared with sections II–V (Table S2). Strains in sections II–V are multicellular or filamentous, which causes cell aggregation during growth, resulting in competition for light energy among cells due to self-shading [58]. In addition, filamentous organisms (sections III–V) often form bacterial mats and are exposed to light competition among phototrophs within the mat [59]. These factors suggest that light utilization by rhodopsin may help cyanobacteria to survive in an environment where they compete for light with other cells. Indeed, high rhodopsin possession rate lineages such as the C-II, C-III, and D lineages are dominated by filamentous organisms (section III) (Fig. S2). In contrast, rhodopsin-mediated photosynthesis may not be important for the survival of sparsely distributed cyanobacterial cells (section I). For example, rhodopsin genes have not been detected in cyanobacteria of the C-I lineage (Fig. S2), such as Prochlorococcus and marine Synechococcus, which are free-living unicellular oceanic organisms, whereas most marine bacteria living in the photic zone are known to possess rhodopsins for light energy harvesting [19].
Heterologous expression analyses showed that the CyR-IIs function as light-driven outward H+ pumps (Fig. 2B). H+ pumping rhodopsins in cyanobacteria have already been found (e.g. XLRs and CyRs) [21,47], whereas PRs, the most diverse and abundant H+ pumping rhodopsins in the ocean, have never been found in any cyanobacteria. This suggests that PR is a light-driven H+ pump specialized for marine heterotrophs living in oligotrophic environments [19]. The representative CyR-II, P7104R, does not exhibit the dark-adapted and light-adapted phenotypes (Fig. S4A). Furthermore, the retinal configuration of P7104R is predominantly all-trans retinal (Fig. S4B, Table 1). It is known that BR forms a highly dense structure on the native membrane (called the purple membrane, BR:lipids = 75:25) [60], whereas the amount of other H+ pumping rhodopsins, including CyR, in the native membrane is significantly less [18]. If BR has only all-trans retinal, it induces a large pH change on the cytoplasmic side, resulting in unfavorable effects on the archaeon itself. In contrast, the pH change induced by other H+ pumping rhodopsins is much less. Thus, the presence of dark–light adaptation in BR and the absence of it in other H+ pumping rhodopsins seem to be reasonable. Cyanobacteria possess complex photoreceptor systems [61,62]. All of the photoreceptor proteins that have been studied in detail, such as phytochrome, phycocyanin, and phycoerythrin, work to increase the efficiency of photosynthesis [63,64]. In contrast, the light energy received by the retinal of CyR or CyR-II would be consumed as energy to transport H+ out of the membrane, independent of the photosynthetic system. It is generally considered that the electrochemical gradient generated by H+-transporting rhodopsin is used for ATP synthesis, flagellar rotation, and so on [65,66]. In order to elucidate the physiological functions of CyR and CyR-II, it is necessary to clarify their subcellular localization in future studies.
Both CyR and CyR-II function as light-driven outward H+ pumps, but the physical and spectroscopic characteristics of the rhodopsin proteins are slightly different (Figs 2B and 3A, and Table 1). These differences may have arisen during the process of cyanobacterial adaptation to their habitat. P7104R, which is the representative CyR-II, has a high pKa value of proton acceptor (Asp74P7104R, <1.1 and ~6.6) and a lack of proton-releasing group amino acids (corresponding to Glu194BR and Glu204BR) (Figs 2A, S3, and S5E and F); these properties of P7104R were more similar to PR than CyR (Table 1). These features may reflect that Nodosilinea nodulosa PCC 7104 harboring P7104R is isolated from the ocean (Figs 1B and S2). The high pKa value in PR (Asp97GPR, ~7.9; Asp97BPR, ~6.2 and ~7.9) [54,56], presumably suitable for the weak alkaline marine environments, is stabilized via a hydrogen bond between Asp97PR (corresponding to Asp79BPR) and His75PR (His57BPR) [67]. Through our structural analysis, the high pKa value in P7104R may also be stabilized via a hydrogen bond between Ser96P7104R and His42P7104R (Fig. S5E and F). In contrast to CyR, the proton donor (corresponding to Asp96BR) in some CyR-II is replaced by Ser (Ser96P7104R) (Figs 2A, S3, and S5J and K). This suggests that a proton is supplied directly to the Schiff base Lys215P7104R from the environment, rather than from a proton donor [32]. The observed changes in the amino acid residues of the proton donor are indicative of a process of adaptation to the environment, whereby rhodopsin not only underwent modification to adapt to cyanobacteria, but also underwent further modification to adapt to each environment.
The CyR-II clade found in this study is divided into two groups: GCyR-IIs with an absorption maximum at 550 nm and YCyR-IIs at 570 nm (Figs 1B and 3A). Mutational analyses showed that the red-shifted absorption maximum of P7104R is attributed to the presence of a larger functional group amino acids (Phe181P7104R and Trp182P7104R) around the retinal (Fig. 3D), which are residues that differ from those reported in PR [68]. Such red shifts in the absorption wavelength of photoreceptors have also been reported in chlorophyll-based photosystems, such as chlorophyll d [1] and f [69,70]. The BR (λmax = 570 nm) possessed by halophilic archaea is also known to utilize relatively long-wavelength region light [16,60]. This absorption range is suitable for hypersaline lakes because carotenoids produced by hyper-halotolerant organisms in the upper layer, such as the unicellular green alga Dunaliella salina [71], absorb blue-to-green light. In other words, the light-absorbing range of rhodopsin as well as chlorophyll is highly dependent on the light conditions of their habitat [2]. The YCyR-II genes were found mainly in cyanobacteria inhabiting certain terrestrial environments, such as sediment and soil (Fig. 1B), suggesting that red-shifted rhodopsin may harvest light and function more efficiently in their habitats. In addition, the presence of cyanobacteria with both CyR and CyR-II (two Chroococcidiopsidaceae MAGs, see Figs S1 and S2) may also be advantageous because their absorption maximum does not overlap, suggesting flexible adaptation to different light conditions.
Conclusions
Our survey of cyanobacterial rhodopsins in assembled genomes and metagenomic contigs and phylogenetic analysis revealed a new cyanobacteria-specific rhodopsin clade: CyR-II. CyR-IIs function as light-driven outward H+ pumps and are subdivided into two subclades: YCyR-II (λmax = 570 nm) and GCyR-II (λmax = 550 nm). In addition, the present rhodopsin phylogeny revealed that YCyR-II was mainly identified from sediment and soil, and GCyR-II from marine bacterial mat and biofilm. This study introduces a new key player, “rhodopsin,” into the history of cyanobacterial radiation to diverse environments, which was previously thought to be driven solely by chlorophyll diversification. Cyanobacteria may have expanded into a wide range of environments not only by diversifying the photosynthetic photoreceptor pigments, but also by acquiring rhodopsins and adjusting the absorption wavelength of the rhodopsins to adapt to the light conditions of their habitat.
Acknowledgements
We thank Drs. K. Hirata, N. Sakai, and Y. Kawano at the BL32XU of SPring-8 (Hyogo, Japan) for their technical assistance with data collection and data processing. We also thank K. Hanada for expression plasmid preparation, and K. Katsura for cell-free expression reagents preparation. We also thank K. Kogure and R. Narikawa for discussions useful for this study.
Author contributions
M.H.-T., M.S., Y.S., and S.Y. designed the research; M.H.-T., T.H., K.K., Y. Nishimura, M.K., Y. Nakajima, Y.I.-K., and T.K.-S. performed the research; M.H.-T., T.H., K.K., and S.Y. analyzed the data; and M.H.-T., T.H., Y. Nishimura, K.K., T.K.-S., M.S., Y.S., and S.Y. wrote the paper.
Conflicts of interest
The authors declare no competing interests.
Funding
This work was supported by MEXT/JSPS KAKENHI, grant numbers 19J21582 and 22KJ3197 (to M.H.-T.); 17K07324 and 20H05450 (to T.H.); 21H02446 (to Y.S.); and 15K14601 and 22H00557 (to S.Y.). This research was partially supported by Research Support Project for Life Science and Drug Discovery (Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)) from AMED under grant number JP23ama121001 (support numbers 1455 and 1463) and JP23ama121013 to M.S. and MEXT Advancement of Technologies for Utilizing Big Data of Marine Life JPMXD1521474594 to S.Y. This work is also supported by JST, PRESTO, grant number JPMJPR23G9 (to M.H.-T.) and ACT-X, grant number JPMJAX21BK (to Y.N.), Japan. Computation time was partly provided by the Super Computer System, Institute for Chemical Research, Kyoto University.
Data availability
The coordinates and structure factor for P7104R have been deposited in the Protein Data Bank, www.wwpdb.org (PDB ID code 8H79). The sequences of cyanobacterial metagenome-assembled genomes and contigs used in this study and Supporting Data S1–S8 are available at Figshare (https://doi.org/10.6084/m9.figshare.26232335).
