A unique clade of light-driven proton-pumping rhodopsins evolved in the cyanobacterial lineage

Phylogenetic analysis and distribution of rhodopsin genes in cyanobacteria

To survey the distribution of rhodopsin genes in cyanobacteria, a sequence homology search was performed using 154 cyanobacterial genomes, including 126 genomes used in a large-scale comparative genomic study of cyanobacteria²¹ and 28 genomes known to possess rhodopsin genes obtained from a public database. Based on this search, 56 rhodopsin genes in 42 cyanobacterial genomes were identified (Table 1). Stratified by the habitat, the rhodopsin genes were found almost exclusively in freshwater cyanobacteria: 29 in freshwater, 9 in high salinity, 2 in marine, and 2 in NA (not available). There are, however, 9 genomes from a high salinity environment, 8 were from the same site and showed little genetic variation. Phylogenetic analysis of the amino acid sequences of the rhodopsins encoded by the 56 identified genes revealed that the proteins belonged to four known rhodopsin clades: XLR (3 genes), NaR (1 gene), XeR (15 genes), and CyHR (24 genes) and one novel clade (13 genes) (Table 1 and Fig. 1a and Supporting Information Fig. S1); the novel rhodopsin clade consisted entirely of rhodopsin genes from cyanobacterial genomes so we named the clade “cyanorhodopsin” (CyR) (Fig. 1b).

Cyanobacterial nomenclature is based largely on morphological features; therefore, species names often do not reflect their lineage. Therefore, to examine the rhodopsin gene distribution we reconstructed a phylogenomic tree of the 154 cyanobacteria genomes by using conserved phylogenetic marker proteins (120 ubiquitous single-copy proteins). Seven subclades (A–G) were assigned to the phylogenomic tree according to a previous study²¹, together with information on source environment, morphology, presence or absence of the retinal biosynthetic gene diox1 (carotenoid oxygenase), and genome size (Supporting Information Fig. S2). Almost all of the genomes were found to encode diox1 (152/154), indicating that microbial strains with rhodopsins generally also have the ability to produce retinal. In addition, we found that the rhodopsin-possessing strains were not evenly distributed across the subclades (Supporting Information Fig. S2, Table 1 and Supporting Information Table S1): strains in subclades A, C (mainly marine cyanobacteria with relatively smaller genomes), F, and G did not possess any rhodopsin genes; in contrast, almost all the strains in subclade D did possess rhodopsin genes and subclade B contained all functional types of cyanobacterial rhodopsin detected in this study (Supporting Information Fig. S2 and Table 1). No morphological bias in rhodopsin distribution was observed (Supporting Information Fig. S2).

Amino acid sequences and functions of the rhodopsins in the CyR clade

To examine the functions of the rhodopsins in the CyR clade, a motif sequence containing specific amino acid residues that are crucial for ion transport activity was examined. In BR, the motif corresponds to Asp85^BR, Thr89^BR, and Asp96^BR (DTD) in the third helix (helix C); the Asp85 and Asp96 residues work as proton acceptor and donor, respectively, and Thr89 forms a hydrogen bond with Asp85. Of the 13 rhodopsins in the CyR clade, the DTD motif was detected in 10, whereas the corresponding motif was Asp, Thr, Glu (DTE) in Calothrix sp. NIES-2098, Nostoc sp. RF31Y, and Nostoc sp. 106C (Supporting Information Fig. S3). All of the CyRs included Lys204^N2098R in the seventh helix (helix G), which is known to make a Schiff base linkage between the rhodopsin protein moiety and the retinal chromophore in other rhodopsins (Supporting Information Fig. S3). Also, an aspartic acid residue (Asp200^N2098R) and two glutamic acid residues (Glu182^N2098R and Glu192^N2098R), which are classified as a counterion stabilizing the Schiff base and a proton release group, respectively, were conserved in BR and all of the CyRs^22,23. Based on this analysis, CyRs were expected to function as light-driven H⁺ pumps.

Next, to further examine the ion-transporting activities of the CyRs, we heterologously expressed three synthesized rhodopsin genes—N2098R (BAY09002.1), B1401R (WP_074382570.1), and N4075R (GAX43141.1)—in Escherichia coli (see Fig. 1b). Rhodopsin-expressing E. coli cells showed more colors than control vector (pET21a) (Fig. 2a) and protein expressions were detected by western blots using anti-His-tag antibody (Fig. 2b). We examined the light-induced change in the pH of the cell suspension. A light-induced decrease in pH was observed in the suspensions of the cells expressing each of the rhodopsins (Fig. 2c, solid line), and this decrease was almost completely abolished in the presence of the protonophore carbonyl cyanide m-chlorophenylhydrazone (CCCP) (Fig. 2c, broken line). These results showed that N2098R and B1401R transport proton from the cytoplasmic side to the extracellular space. On the other hand, N4075R was thought to act as an outward proton pump, but its activity was smaller than that of the others. Therefore, the possibility of transporting other ions cannot be excluded.

Spectroscopic characterization of N2098R

To characterize the photochemical properties of the CyRs, we focused on N2098R, a well-expressed, stable rhodopsin. After adaption of purified N2098R to the light or dark condition, the absorption maxima of both adapted samples of N2098R was located at 550 nm (Fig. 3a), which was similar to that of GR but not to that of BR or PR (Table 2)^6,16,24.

Next, we examined the retinal configuration in N2098R by high-performance liquid chromatography. Both in light- and dark-adapted samples, the isomeric state of retinal was predominantly all-trans (Supporting Information Fig. S4 and Table 2), which was similar to the isomeric state of retinal in PR²⁵ and GR^15,26 but different from that in BR (Table 2)²⁷.

Charged residues (e.g., Asp85^BR and Lys216^BR in BR) are essential for proton transportation by rhodopsins²⁸. We therefore estimated the pK_a values of the charged residues in N2098R (i.e., Asp74^N2098R and Lys204^N2098R) by pH titration and fitted the data using the Henderson–Hasselbalch equation assuming a single pK_a; the pK_a values of Asp74^N2098R and Lys204^N2098R were estimated to be < 2.0 and 10.7, respectively (Supporting Information Fig. S5 and Table 2). The pK_a of Asp74^N2098R was much smaller than that reported for the equivalent amino acid in other microbial rhodopsins^16,29,30,31, whereas the pK_a of Lys204^N2098R was comparable with that reported for the equivalent amino acid in other microbial rhodopsins (Table 2)^16,32,33.

When microbial rhodopsins absorb visible light, the retinal chromophore is isomerized from the all-trans to the 13-cis form and the rhodopsin protein then forms various photointermediates before returning to its original state. These photochemical reactions, which occur within the picosecond-to-second time frame, are collectively referred to as the photocycle. In BR, several photointermediates (designated as intermediates K, L, M, N, and O) have been identified³⁴. To examine the ion-transportation mechanism of N2098R, the photochemical reactions it undergoes were examined by flash-photolysis analysis. The flash-induced difference spectra of purified N2098R showed bleaching and recovery to the original state within 977 ms at a wavelength of around 550 nm (Fig. 3b), which coincided with the absorption maxima of N2098R (Fig. 3a). Two large positive peaks were observed at around 620 and 405 nm; the absorbance at around 620 nm appeared within 0.01 ms and the absorbance at around 405 nm appeared within 2.04 ms (Fig. 3b). In addition, one small positive peak was observed at around 645 nm; the absorbance at around 645 nm appeared with in 81.1 ms (Fig. 3c).

Next, the time courses of the changes of absorbance at wavelengths of 405, 550, 620, and 645 nm were examined (Fig. 3d and e). From the time and location of the absorption maxima, the increases and decreases at around 620, 405, and 645 nm were attributed to the formation and decay of the K-, M-, and O-intermediates, respectively. That is, when the K-intermediate was bleached, the M-intermediate began to form; then when the M-intermediate decayed, the O-intermediate began to form. After then, O-intermediate was bleached and the rhodopsin protein returned to its original state (550 nm). The photocycle was complete by about 300 ms. The M-decay rate constant at pH 7.0 was estimated to be 0.016 ms⁻¹ by fitting a single exponential equation. This decay rate constant was much smaller than that reported for PR (4 ms⁻¹)²⁵, BR (0.25 ms⁻¹)³⁵, and GR (2.3 ms⁻¹ and 1 ms⁻¹)^15,36 indicating that the M-intermediate of N2098R was much longer-lived than that of these other rhodopsins (Table 2). When the same experiment was repeated under various pH conditions, the M-decay rate was markedly increased under acidic conditions, suggesting that proton concentration mediates the rate of decay of the M-intermediate (Supporting Information Fig. S6).

To clarify the timing of proton uptake and release during the photocycle, we monitored the changes in absorption of N2098R in the presence or absence of pyranine, a pH-sensitive dye³⁷. The pyranine signal in the difference spectrum decreased within 0.1 ms and then increased within 100 ms (Fig. 3d, gray line); these curves coincided well with the M-formation (K-decay) and M-decay (recovery of the original state). Because decreases and increases of the pyranine signal reflect the acidification and alkalization, respectively, of the bulk solution, this finding indicates that a proton is released from N2098R upon M-formation and is taken up from the bulk solution upon M-decay. The order of the timing is the same as that of BR³⁸, but different from that of PR²⁵.

Structure of N2098R and comparison with those of BR, GR, and ASR

The crystal structures of cell-free-synthesized N2098R and N4075R were determined at 2.65 Å and 1.9 Å resolution, respectively (Supporting Information Table S2). Since the two structures were almost identical, only the structure of N2098R is described (Fig. 4 and Supporting Information Fig. S7a–c). The N2098R protein crystallized into space group C222₁, with a N2098R trimer per asymmetric unit. The N2098R monomer is composed of seven transmembrane helices (helices A–G) and a long C-terminal helix (Fig. 4a). The long C-terminal helix is bent at the end of helix G, Gly215^N2098R, and covers the cytoplasmic surface of helix C–F (Fig. 4a, right panel). However, this bend of the C-terminal helix could be due to crystal packing, so it is unclear whether the C-terminal helix of N2098R is bent under physiological conditions.

The overall structure of N2098R is similar to that of BR (PDB code 1C3W)³⁹ (root-mean-square deviation [RMSD] 0.995 Å), except for the presence of the long C-terminal helix in N2098R (Supporting Information Fig. S7d and Supporting Information Table S3). The structure around the retinal in N2098R and BR is shown in Fig. 4b and Supporting Information S7e; in both rhodopsins, the retinal is unbent. A pentagonal cluster consisting of three water molecules adjacent to the Schiff base of Lys204, Asp74, and Asp200 plays an important role in proton transport (Fig. 4b, broken lines). This water cluster near the retinal is similar to that next to the Schiff base of BR (Lys216^BR, Asp85^BR, and Asp212^BR) (Supporting Information Fig. S7e). In addition, the location of the putative proton release group comprising Glu182 and Glu192 (corresponding to Glu194^BR and Glu204^BR in BR) is also very similar (Supporting Information Fig. S7f.).

The structure of N2098R was also compared with the cyanobacterial rhodopsin GR (PDB code 6NWD)⁴⁰ and ASR (PDB code 1XIO)⁴¹, which are included in the XLR and XeR rhodopsin clades, respectively. The overall structure of N2098R is similar to that of GR (RMSD 1.91 Å) and ASR (RMSD 1.40 Å) (Supporting Information Fig. S7g and j and Supporting Information Table S3), but the structural position of the water molecules and amino acids around the retinal and the putative exit region of the proton are quite different (Supporting Information Fig. S7h, I, k, and l). These structural differences also suggest that CyR including N2098R is a “unique” proton-pumping rhodopsin, unlike known cyanobacterial rhodopsin such as GR and ASR.

To examine the amino acid residues crucial for the proton-transport activity of N2098R, mutants at the putative proton acceptor Asp74^N2098R, proton donor Glu85^N2098R, or counterion Asp200^N2098R were examined. Replacement of these amino acids prevented light-mediated change of the pH of the cell suspensions despite expressed as well as wild type in E. coli cells, indicating that these residues are essential for proton transport (Supporting Information Fig. S8); this finding provides further evidence that N2098R functions as a light-driven outward proton pump. In addition, deletion-mutation analysis of the C-terminus (1–125) indicated that the helix is important for structural stabilization (Supporting Information Fig. S8).

Source: Ecology - nature.com