Research Article

Extensive remodeling of a cyanobacterial photosynthetic apparatus in far-red light

See allHide authors and affiliations

Science  12 Sep 2014:
Vol. 345, Issue 6202, pp. 1312-1317
DOI: 10.1126/science.1256963

Making sure leftover light gets in

Bacteria that harvest sunlight through photosynthesis are everywhere and are vital for absorbing carbon dioxide. Like plants in a jungle, these cells have to compete for access to light. Gan et al. found that one species of such cyanobacteria, Leptolyngbya, survives on the scraps of leftover light that other photosynthetic cells can't use. When the light is crowded out by competing organisms, Leptolyngbya changes the pathways in its cells that make the light-harvesting molecules and remodels them into structures that can use redder light at longer wavelengths that is unused by other cells.

Science, this issue p. 1312


Cyanobacteria are unique among bacteria in performing oxygenic photosynthesis, often together with nitrogen fixation and, thus, are major primary producers in many ecosystems. The cyanobacterium, Leptolyngbya sp. strain JSC-1, exhibits an extensive photoacclimative response to growth in far-red light that includes the synthesis of chlorophylls d and f. During far-red acclimation, transcript levels increase more than twofold for ~900 genes and decrease by more than half for ~2000 genes. Core subunits of photosystem I, photosystem II, and phycobilisomes are replaced by proteins encoded in a 21-gene cluster that includes a knotless red/far-red phytochrome and two response regulators. This acclimative response enhances light harvesting for wavelengths complementary to the growth light (λ = 700 to 750 nanometers) and enhances oxygen evolution in far-red light.

Cyanobacteria are unique among bacteria in performing oxygenic photosynthesis and are ecologically important primary producers (1). Marine cyanobacteria, mostly Prochlorococcus, Synechococcus, and Trichodesmium species, account for >25% of the net primary productivity in oceans, and terrestrial cyanobacteria also contribute substantially to global photosynthesis (2, 3). The estimated cyanobacterial biomass in terrestrial, endolithic, and freshwater lake ecosystems is ~40% of that in oceans, but this estimate would increase if polar and subarctic soils, topsoils in subhumid climates, and shallow marine and freshwater benthic environments were included (2). Because approximately equal solar irradiance reaches Earth’s surface in the wavelength ranges 600 to 700 nm and 700 to 800 nm (18.4% versus 14.9% of the total irradiance between 400 and 1100 nm), the capacity to use far-red light to perform oxygenic photosynthesis in terrestrial niches could thus have important consequences in natural and engineered systems.

Cyanobacteria mainly use three large, multisubunit complexes to harvest and convert light into stable, energy-rich compounds: photosystem (PS) I, PS II, and phycobilisomes (PBSs) (1, 47). Cyanobacteria have evolved many mechanisms to maximize their photosynthetic efficiency in response to the incident irradiation. Examples include adjusting the total cellular chlorophyll (Chl) content and the ratio of PS II to PS I (“intensity adaptation”) (8); state transitions that redistribute light energy transfer from PBSs to PS I and PS II (9, 10); and nonphotochemical quenching by orange carotenoid-binding protein (11). Although cyanobacteria were once thought to have simple pigmentation (Chl a, β-carotene, and phycobiliproteins), it is now recognized that some cyanobacteria synthesize Chls b, d, or f (12), many functionally distinct carotenoids (13), and spectroscopically diverse phycobiliproteins (7). Limitation for iron and other nutrients causes changes in light-harvesting proteins (1416).

One of the best-characterized acclimative responses in cyanobacteria, complementary chromatic acclimation (CCA; formerly “adaptation”), was discovered more than 100 years ago. Gaidukov (17, 18) observed that Oscillatoria sancta is reddish brown when grown in green light but blue-green when grown in red light, and he correctly surmised that the color differences were due to altered pigment synthesis (fig. S1). CCA results from compositional remodeling of the peripheral rods of PBSs (19, 20) and occurs through transcriptional and posttranscriptional regulation of specific phycobiliprotein genes. Genes encoding phycoerythrin and its associated bilin reductases, bilin lyases, and linker proteins are expressed in cells grown in green light, whereas those for “inducible” phycocyanin and associated proteins are expressed in cells grown in red light (1922). Central regulatory elements controlling CCA include two response regulators (RcaF and RcaC) and a phytochrome-related sensor histidine kinase, RcaE (19, 20, 23). Because PBSs are primarily, but not exclusively, associated with energy transfer to PS II (9, 10), CCA enhances the overall rate of photosynthesis when the incident irradiation is complementary to cell coloration.

Using a systems-biology approach—including genome sequencing, comparative genomics, transcription profiling (RNA-seq), biochemical and spectroscopic analyses, and proteomics—we show that Leptolyngbya sp. strain JSC-1 (hereafter, JSC-1) (24) remodels its photosynthetic apparatus in response to far-red illumination. These changes are probably controlled by a red/far-red–responsive phytochrome. JSC-1 specifically synthesizes both Chl d and Chl f, in addition to Chl a, when cells are grown with light wavelengths ≥700 nm. Halomicronema hongdechloris, a cyanobacterium isolated from a stromatolite from Shark Bay, Australia, also synthesizes Chl f when cells are grown in far-red light (25, 26). However, JSC-1 additionally undergoes an extensive acclimative response, in which the expression of ≥40% of the genome changes more than twofold after a shift from white light to far-red light. The resulting changes in gene expression lead to the replacement of most of the core proteins of PS I and PS II and to structural remodeling of PBS core substructures. We have named this global acclimation response “far-red light photoacclimation” (FaRLiP), and we show here that FaRLiP substantially improves photosynthetic performance in far-red light.

The JSC-1 genome has a 21-gene cluster that encodes paralogs of most of the core subunits of PS I, PS II, and PBSs (fig. S2). Similar clusters occur in 12 other cyanobacteria, including members of all five sections of the taxon Cyanobacteria (27) (fig. S3). All but one of these clusters contain six genes encoding subunits of PS I: psaA2, psaB2, psaL2, psaI2, psaF2, and psaJ2 (in the remaining strain, psaF2 and psaJ2 are adjacent and probably cotranscribed with psaB3). The psaA2 and psaB2 genes are considerably diverged from the psaA1 (~77% identity, 85% similarity), psaB1 (85% identity, 92% similarity), and psaB3 (82% identity, 90% similarity) genes, respectively. The other psa genes in this cluster (psaL2, psaI2, psaF2, and psaJ2) are similarly distantly related to paralogous psa genes (psaL1, psaL3, psaI1, psaF1, and psaJ1) encoded elsewhere in the genome (see table S1 for a list of genes for subunits of PS I, PS II, PBSs, and related proteins). Transcriptional analyses (tables S1 and S2) showed that the psaA2, psaB2, psaL2, psaI2, psaF2, and psaJ2 genes are not expressed in cells grown under standard growth conditions for many cyanobacteria (28). Note that this locus also includes genes for a knotless phytochrome (rfpA), a DNA-binding response regulator (rfpB) with two CheY receiver domains and a winged-helix DNA binding domain, and a cheY-like gene (rfpC), which form an apparent operon upstream from psbA4 (fig. S2) (rfp stands for regulator of far-red photoacclimation). Phytochromes are widespread red/far-red photoreceptors (29), but orthologs of rfpA only occur in 12 other cyanobacteria, within photosynthesis gene clusters similar to that in JSC-1 (fig. S3). On the basis of phylogenetic analyses and conserved domain architecture, these photoreceptors form a distinctive subfamily among knotless phytochromes (Fig. 1). From these observations and biochemical studies described below, we hypothesize that the RfpA photoreceptor controls gene expression from this cluster.

Fig. 1 RfpA forms a distinctive lineage of knotless phytochromes.

The GAF-PHY domains of selected knotless phytochromes, including the products of 10 rfpA genes (fig. S3), were aligned with Synechocystis sp. PCC 6803 Cph1 and two closely related, knotted phytochromes from Leptolyngbya sp. JSC-1 as an outgroup (fig. S4). The resulting alignment was used to produce a maximum-likelihood (ML) phylogenetic tree; open and closed circles at nodes show the approximate statistical support for the branching pattern as indicated. The colors of the sequence names reflect seven domain architectures for knotless phytochromes (lower left), with phytochrome and related photosensory GAF domains highlighted in red and blue, respectively. RfpA orthologs have a conserved domain architecture (type IV) and form a distinct subfamily of knotless phytochromes, providing two independent lines of evidence that RfpA orthologs form a distinct lineage associated with FaRLiP gene clusters (fig. S3).

Expression of rfpA or its GAF domain in an engineered, phycocyanobilin-producing strain of Escherichia coli (28, 30) yielded red/far-red reversible proteins with nearly identical photochemical properties (figs. S5 and S6A). Like the knotless phytochrome NpR4776 from Nostoc punctiforme, recombinant RfpA-GAF holoprotein photoconverts between red-absorbing (Pr) and far-red–absorbing (Pfr) species under respective far-red or red illumination (fig. S6A). The Pr form of cyanobacterial phytochromes is initially synthesized in vivo (31, 32). RfpA is converted to the Pfr form under a broad range of light wavelengths, but only far-red light (λ ≥ 700 nm) specifically regenerates the Pr form (fig. S6B). Because genetic tools are not available to produce an rfpA mutant in JSC-1, the postulated role of RfpA in sensing far-red light and controlling expression of the 21-gene photosynthesis gene cluster cannot be verified by reverse genetics. However, RfpA specifically senses far-red light, exhibits higher transcript abundance in cells grown in far-red light, and is uniquely colocalized with genes that are only expressed in far-red light (figs. S2, S3, S5, and S6 and table S1). The JSC-1 genome contains other phytochromes and related photosensors, but transcript levels only increase in far-red light for the knotted phytochrome CYJSC1_DRAFT_40400 (Fig. 1 and table S2). Calothrix sp. PCC 7507 and Synechococcus sp. PCC 7335 have RfpA orthologs (fig. S3) but lack Cph1 orthologs. Thus, the distribution and photochemical properties of RfpA strongly support our proposal that RfpA controls the expression of the 21-gene cluster.

To examine the photobiology more closely, JSC-1 cells were grown under six light conditions (28): (i) white fluorescent light (WL), (ii) green-filtered fluorescent light (GL), (iii) red-filtered fluorescent light (RL), (iv) 645-nm or (v) 710-nm light provided by light-emitting diodes (LEDs), and (vi) far-red light (FR) produced from filtered tungsten light (λ > 690 nm) (see fig. S7). The absorption spectra of cells grown in 710-nm light and FR (Fig. 2) showed that they had gained absorption at ~700 to 750 nm (λmax ~706 nm) that was not present in cells grown in WL, GL, or 645-nm light. Low-temperature fluorescence emission spectra at 77 K (fig. S8) of cells grown under the first three listed conditions had emission maxima at 683 nm and 695 from PS II and a strong emission maximum at 725 nm from PS I. These spectra are typical of those for cells synthesizing Chl a and having a relatively high PS I:PS II ratio (24). JSC-1 cells grown in 710-nm light or FR have weak fluorescence emission at 683, 695, and 717 nm and strong emission at 745 nm, in which they resemble cells synthesizing Chl f (25, 26).

Fig. 2 JSC-1 cells have enhanced absorption at 700 to 750 nm when grown in far-red light.

(A) Appearance of cells grown in WL, GL, 645-nm light, and 710-nm light. (B) Absorption spectra of strain JSC-1 cells grown in WL (black line), GL (green line), 645-nm light (red line), FR (solid blue line), and 710-nm light (dotted blue line).

Pigments extracted from these cells were subjected to reversed-phase high-performance liquid chromatography. Only Chl a (figs. S9 and S10, peak 3) and carotenoids were observed in pigment extracts from cells grown in WL or 645-nm light. Cells grown in FR or 710-nm light still produced Chl a as the major Chl, but they also synthesized a pigment with an absorption spectrum identical to that of Chl f (figs. S9 and S10, peak 2) (33, 34). This pigment was confirmed to be Chl f by converting it to the corresponding pheophytin by removing the Mg2+ ion (figs. S11 and S12) and subsequent mass spectrometry (MS) and tandem MS (MS-MS) analyses (figs. S13 and S14). Another Chl, which was slightly more hydrophilic than Chl f (figs. S9 and S10, peak 1), was also detected. The absorption properties (figs. S10 and S12) of this Chl, as well as MS and MS-MS analyses (figs. S15 and S16) of the corresponding pheophytin, showed that it was Chl d (33, 34). Therefore, JSC-1 synthesizes three Chls: Chl a, Chl d, and Chl f.

Preliminary reverse transcription polymerase chain reaction experiments confirmed that genes of the psa2 operon (fig. S2) are transcribed in cells grown in FR (table S2). Cells were therefore grown in WL, transferred to FR for 24 hours, and transcription profiling (RNA-seq) was performed (28). Transcript levels for the photosynthesis-related genes in the 21-gene cluster (fig. S2) increased from 3-fold to 278-fold in cells grown in FR (Fig. 3 and table S1), and transcript levels increased at least twofold for ~900 genes (table S2). Transcript levels decreased by more than 50% for ~2000 genes in FR (table S2), including most of the paralogous genes encoding core subunits of photosynthetic complexes (Fig. 3 and table S1). The transcriptional changes during acclimation to FR are surprisingly extensive and exceed those for heterocyst (1036 genes) or hormogonia (1762 genes) differentiation in Nostoc punctiforme (35).

Fig. 3 Transcription of the 21-gene cluster in JSC-1 in far-red light is part of an extensive acclimative response.

Scatter plot showing relative transcript levels for JSC-1 cells 24 hours after a shift from WL to FR. The circled stars indicate the relative transcript levels for the genes in the 21-gene cluster shown in fig. S2. The magenta, pale green, and aqua triangles indicate relative transcript levels for paralogous genes encoding subunits of PS I, PS II, and PBSs, respectively, that are preferentially expressed in WL. The parallel lines indicate a twofold increase or decrease by half in transcript abundance.

The transcription profiling data indicated that most core polypeptides of PS I, PS II, and PBSs should be replaced by products of the 21-gene cluster when cells are grown in FR. To verify that this was the case, we analyzed the proteins of photosynthetic complexes isolated from JSC-1 cells grown in WL, FR, 645-nm light, and 710-nm light. Fractions enriched in PS I and/or PS II were isolated on sucrose gradients after solubilization of thylakoid membranes with n-dodecyl-β-d-maltoside (Fig. 4A). For JSC-1 cells grown in WL and 645-nm light, two green fractions were observed: (i) an upper green fraction containing PS I monomers and PS II dimers, and (ii) a lower green fraction containing PS I trimers (Fig. 4A). Gradients prepared with solubilized membranes from JSC-1 cells grown in FR and 710-nm light differed dramatically. No PS I trimers were observed, and only a single fraction containing PS I monomers and PS II dimers was observed (Fig. 4A). The complexes derived from cells grown in WL or 645-nm light had absorption (Fig. 4B) and 77K fluorescence emission maxima (Fig. 4C) expected for PS I (725 nm) and PS II (685 and 695 nm) complexes containing Chl a. However, complexes isolated from cells grown in FR and 710-nm light had additional absorption features at ~720 nm (Fig. 4B) and had fluorescence emission maxima at 745 nm (Fig. 4C). These spectra demonstrate that both PS I and PS II complexes of cells grown in FR and 710-nm light contain Chl f.

Fig. 4 PS I and PS II complexes are remodeled under far-red light.

(A) Sucrose density gradient centrifugation for isolation of Chl-containing complexes from thylakoid membranes of JSC-1 cells grown in WL, 645-nm light, 710-nm light, and FR after solubilization with n-dodecyl-β-d-maltoside. (B) Absorption spectra of gradient fractions containing PS I trimers (WL, solid black line; 645-nm light, solid red line) and a mixture of PS I monomers and PS II from WL (dashed black line); 645-nm light (dashed red line); FR (solid blue line); and 710-nm light (dashed blue line). (C). Low-temperature fluorescence emission spectra for the same fractions as in (B).

Proteins associated with isolated photosynthetic complexes were analyzed by trypsin digestion and mass spectrometry of the resulting peptides (table S1). This analysis demonstrated that the core polypeptides comprising complexes from cells grown in FR and 710-nm light were the products of those genes whose transcript levels strongly increased in cells grown under FR (e.g., psaA2, psaB2, psaL2, and psaF2 and psbB2, psbC2, psbD3, psbA3, psbA4, and psbH2) (table S1). In contrast, the core proteins in complexes from cells grown in WL or 645-nm light contained products of paralogous genes with lowered transcript levels under FR (e.g., psaA1, psaB1, psbB1, and apcE1) (Fig. 3 and table S1). Thus, the proteins composing PS I and PS II differed in their responses to the growth light conditions.

We isolated PBSs from cells grown in 645-nm and 710-nm light to assess whether absorption differences accompanied polypeptide replacement in FR (Fig. 5A). PBSs isolated from cells grown in 710-nm light have an absorbance feature at 708 nm that is correlated with a low-temperature fluorescence emission peak at 723 nm, but PBSs isolated from cells grown in 645-nm light lack these features (Fig. 5B). The 708-nm absorption band is red-shifted nearly 40 nm compared with allophycocyanin-B (λmax = 670 nm), previously the most red-shifted maximal absorption for characterized native phycobiliproteins (36). After denaturation of PBS proteins in 8.0 M urea at pH 3 (fig. S17), only phycocyanobilin and phycoerythrobilin were detected (37), which suggested that the long-wavelength absorption arises from phycocyanobilin and not from a new bilin chromophore. Even PBS ultrastructure changes in cells grown in FR. The ApcE1 core linker phycobiliprotein expressed in WL and 645-nm light has four linker-repeat (REP) domains and thus is predicted to assemble PBSs with “pentacylindrical” core substructures (7, 38) (Fig. 6A, top). ApcE1 has a highly conserved cysteine residue at position 197 that ligates a phycocyanobilin chromophore via a thioether linkage (7). ApcE2, the core linker phycobiliprotein synthesized in FR, has only two REP domains and, thus, is predicted to produce PBSs with only two core cylinders (7, 38) (Fig. 6A, bottom). ApcE2 has no cysteine residues in its N-terminal phycobiliprotein domain; thus, ApcE2 should bind phycocyanobilin noncovalently. Noncovalently bound phycocyanobilin would have one additional conjugated double bond, which should shift the absorption and fluorescence emission of ApcE2 to a longer wavelength, as shown by site-directed mutagenesis of apcE in Synechococcus sp. PCC 7002 (39). Such changes could extensively modify the energy transfer pathway(s) and efficiency in the cores of PBSs in cells grown in FR. These observations raise the possibility that other core PBS subunits (ApcA2, ApcB2, ApcD3, and ApcD4) could also bind phycocyanobilin noncovalently to increase absorption of FR light (λ ≥ 700 nm).

Fig. 5 Phycobilisomes isolated from cells grown in far-red light have enhanced absorption from 700 to 750 nm.

(A) Sucrose gradients showing phycobilisomes isolated from cells grown in 645-nm light and 710-nm light. (B) Absorption spectra (solid lines) and fluorescence emission spectra (dashed lines) for PBS fractions from cells grown in 645-nm light (red lines) and 710-nm light (blue lines).

Fig. 6 Diagrams summarizing the remodeling of components of PBSs, PS I, and PS II, during FaRLiP.

(A) Remodeling of PBS cores in FR. Except ApcF and ApcC (see table S1), all allophycocyanin-related components (shown in aqua) composing the PBS core substructure are replaced in FR. ApcE1 has four REP (linker) domains and should assemble a pentacylindrical substructure in cells grown in WL, GL, or RL (top). ApcE2 has only two REP (linker) domains and should assemble a bicylindrical core substructure in FR (7, 38) (bottom). PBSs assembled in FR contain some phycoerythrin (table S1; shown as red disks; phycocyanin is shown in blue). This may occur because FR cannot efficiently photoconvert the CCA photoreceptor into its green-absorbing (Pg) form (23). Remodeling of PS I (B) and PS II (C) illustrated using the x-ray structures of PS I [Protein Data Bank (PDB) = 1JB0] (47) and PS II (PDB = 3BZ2) (48) from Thermosynechococcus elongatus. The view of PS I is from the lumenal side, so PsaC, PsaD, and PsaE are not visible. The view of PS II is from the cytoplasm-facing side. Each subunit is shown in a different color (top), and subunits replaced in FR are shown in red (bottom). Subunit Ycf12 of PS II is not encoded in the draft genome of JSC-1.

JSC-1 thus remodels its photosynthetic apparatus during FaRLiP by synthesizing Chl d and f, replacing the core subunits of PS I and PS II, and modifying the proteins of the PBS core as summarized in Fig. 6. Analogously to CCA, these changes produce photosynthetic complexes with absorption that is complementary to the incident irradiation between 700 and 750 nm. Additionally, JSC-1 alters relative transcript levels for >40% of the JSC-1 genome, which leads to extensive modification of cellular metabolism (table S2). The transcription changes and global replacement of core components of PS I, PS II, and PBSs during FaRLiP are distinct from the incorporation of PsbA variants in PS II complexes of Synechocystis sp. PC 6803 and other cyanobacteria in response to high light intensity or anoxic conditions (40) and from reported transcription changes for psbD3 and psbE2 of Acaryochloris marina cells grown in FR or WL at very low irradiance (41).

There is no overlap between the structural remodeling of photosynthetic complexes that occurs during CCA and FaRLiP, and it is noteworthy that JSC-1 performs both acclimation responses (see fig. S1) (24). We assume that PS I, PS II, and PBS subunits produced during FaRLiP have adapted through evolution to perform photosynthesis more efficiently when some Chl a molecules are replaced by Chl f (and Chl d) in cells growing in FR. Indeed, JSC-1 cells that have acclimated to 710-nm light have 40% greater oxygen evolution with far-red actinic light than cells acclimated to 645-nm light, although the two types of cells have identical light saturation behavior when the actinic light is WL (fig. S18). This enhanced photosynthetic performance in FR would be ecologically important whenever cells grow in light that is strongly filtered by Chl a absorbance—for example, in mats, stromatolites, cyanobacterial blooms, or the shade of plants. FaRLiP should also benefit organisms living in sandy soils, because far-red light penetrates deeper than visible wavelengths (42). The 730 nm:650 nm ratio shifts from ~0.84 at the soil surface to 2.8 at a depth of 6 mm (43), and this could be further enhanced by Chl a absorption filtering. Thus, FaRLiP could have a substantial impact on cyanobacterial photosynthesis in soil crusts (two organisms in fig. S3 are soil isolates).

Our results show that it is possible for cyanobacteria to retain paralogous copies of genes for functionally specialized photochemical reaction centers. This observation has important implications for the evolution of type 1 and type 2 reaction centers during the evolution of photosynthesis, as well as for the extension of Chl biosynthetic pathways (44). If type 1 reaction centers evolved first, a likely mechanism for the origin of type 2 reaction centers was gene duplication and functional divergence (4446). As shown in this study, functionally distinctive and divergent PS I and PS II reaction centers are formed in strain JSC-1 during growth in FR. Finally, our findings could have important implications for introducing the capacity to utilize FR into plants.

Supplementary Materials

Materials and Methods

Figs. S1 to S18

Tables S1 and S2

References (4960)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: This study was funded by grant MCB-1021725 from the NSF to D.A.B. The genome sequence of Leptolyngbya sp. strain JSC-1 was determined under the auspices of the U.S. Department of Energy's Office of Science, Biological, and Environmental Research Program, and by the University of California, Lawrence Berkeley National Laboratory under contract no. DE-AC02-05CH11231, Lawrence Livermore National Laboratory under contract no. DE-AC52-07NA27344, and Los Alamos National Laboratory under contract no. DE-AC02-06NA25396. Spectroscopic characterization of RfpA and NpR4776-PCM was funded by a grant from the Chemical Sciences, Geosciences, and Biosciences Division, Office of Basic Energy Sciences, Office of Science, U.S. Department of Energy (DOE DE-FG02-09ER16117 to J.C.L.) The authors thank J. R. Miller for assistance in performing mass analyses on Chls d and f; B. Stanley and A. Stanley for advice and technical assistance in performing the tryptic peptide mass fingerprinting; and C. Praul and C. Price for assistance in performing RNA-seq profiling.The authors thank J. Lecomte, W. Schluchter, J. Golbeck, and A. Glazer for reading the manuscript and helpful suggestions. This whole-genome shotgun project for Leptolyngbya sp. strain JSC-1 (alternative name, Marsacia ferruginose; IMG taxon ID 2022827000; GOLD ID = Gi02032) has been deposited at DDBJ/EMBL/GenBank under the accession JMKF00000000; the version described in this paper is version JMKF01000000. RNA-seq data were deposited in the NCBI Sequence Read Archive (SRA) under accession number SRP041154.
View Abstract

Stay Connected to Science

Navigate This Article