Thylakoid potassium channel is required for efficient photosynthesis in cyanobacteria

Abstract

A potassium channel (SynK) of the cyanobacterium Synechocystis sp. PCC 6803, a photoheterotrophic model organism for the study of photosynthesis, has been recently identified and demonstrated to function as a potassium selective channel when expressed in a heterologous system and to be located predominantly to the thylakoid membrane in cyanobacteria. To study its physiological role, a SynK-less knockout mutant was generated and characterized. Fluorimetric experiments indicated that SynK-less cyanobacteria cannot build up a proton gradient as efficiently as WT organisms, suggesting that SynK might be involved in the regulation of the electric component of the proton motive force. Accordingly, measurements of flash-induced cytochrome b₆f turnover and respiration pointed to a reduced generation of ΔpH and to an altered linear electron transport in mutant cells. The lack of the channel did not cause an altered membrane organization, but decreased growth and modified the photosystem II/photosystem I ratio at high light intensities because of enhanced photosensitivity. These data shed light on the function of a prokaryotic potassium channel and reports evidence, by means of a genetic approach, on the requirement of a thylakoid ion channel for optimal photosynthesis.

The unicellular photoheterotrophic transformable cyanobacterium Synechocystis sp. PCC 6803 is the first photosynthetic organism for which the complete genome sequence was known. This prokaryote is characterized by an intracellular membrane system with thylakoids, where both photosynthesis and respiration take place. Cyanobacteria provide suitable model systems for studies on photosynthesis because these prokaryotes perform oxygenic photosynthesis using an apparatus similar to that found in chloroplasts of higher plants and algae. Moreover, cyanobacterial cultures can easily be exposed directly to defined stress conditions, and they are able to acclimate to a wide range of environments. Cyanobacteria are considered to represent the progenitors of chloroplasts, and Synechocystis has been widely used for genetic and biochemical studies of photosynthesis and various related metabolic processes (1).

In a recent study, we identified in the genome of Synechocystis sp. PCC 6803 SynK (slr0498), displaying the amino acid sequence (TMTTVGYGD) that is typical of all known K⁺ channels (2). This sequence forms a structural element known as a selectivity filter, which prevents the passage of Na⁺ ions but allows K⁺ ions to pass across the membrane at rates approaching the diffusion limit. SynK was found to function as potassium-selective channel when expressed in mammalian cells. Furthermore, SynK complemented K⁺ uptake in a K⁺-transporter-deficient Escherichia coli strain, and its localization within cyanobacteria was determined. SynK represents a thylakoid-located ion channel identified in cyanobacteria and is conserved in various photosynthetic cyanobacteria species (2).

During photosynthesis, in cyanobacteria a light-driven flux of protons from the cytoplasm to the luminal side of the thylakoid membrane occurs via photosytem (PS)II water splitting and plastoquinone reduction and the cytochrome b₆f turnover. In cyanobacteria, respiration that also occurs in the thylakoid leads to the flux of protons as well. These processes lead to the formation of a pH gradient and to the development of a transmembrane electrical potential. Ion fluxes across thylakoid membranes might contribute to regulation of photosynthesis and respiration by modulating the electric component of the transthylakoid proton motive force (pmf) (composed of osmotic component ΔpH and of electric component ΔΨ). Recent works indicate an approximate 50% contribution of the electric field to steady-state transthylakoid pmf both in higher plants and eukaryotic algae (e.g., refs. 3, 4). In higher plants, it has been postulated that the efflux of cations from the lumen toward the stroma or flux of anions in the opposite direction (5, 6) would permit dissipation of the transmembrane electrical potential while conserving the pH gradient. In chloroplasts, ΔA₅₁₅ decay measurements highlighted a major, slow component, attributable to the proton efflux via ATP synthase, but also a faster kinetic component, presumably attributable to flow of ions across the thylakoid (7). Magnesium (6) and potassium (8) have been proposed to act as dominant counterions. In higher plants, a K⁺ flux out of the thylakoid was measured upon illumination (9, 10) and tetraethylammonium⁺, a potassium channel inhibitor, was shown to reduce whole-chain photosynthetic electron transport by restriction of conductance through thylakoid membrane K⁺ channels on isolated thylakoids (11). In this latter case, the authors proposed that restriction of K⁺ efflux in the light would lead to an increased membrane potential (the lumen becoming more positive) across the thylakoid concomitantly with light-induced proton pumping. The buildup of positive charge within the lumen would increase the electrical gradient against which proton pumping must occur. This back pressure on proton pumping occurring because of increased positive charge in the lumen would impose an increased restraint on electron transport. However, direct genetic proof in favor of the “counterbalance” hypothesis is still missing, also because no gene(s) encoding for thylakoid-located potassium channel(s) have been identified to date for higher plants. Potassium could be an efficient counter ion in cyanobacteria, where the basic processes of photosynthesis are considered to occur similarly to higher plants, thanks to its high concentration in the cytoplasm (around 200 mM) (12).

In the present paper, we used reverse genetics to unravel the physiological role of a potassium channel identified in cyanobacteria. We show that removal of the thylakoid-located potassium channel SynK modifies photosynthetic activity and leads to a more photosensitive phenotype. This report indicates, by genetic means, the role of a thylakoid-located ion channel in modulation of photosynthetic activity.

Results

To understand the physiological role of SynK in cyanobacteria, a mutant deficient in SynK was generated exploiting the capacity of Synechocystis sp. PCC 6803 to spontaneously integrate foreign DNA into its genome (present in dozen of copies) by homologous recombination. This results in targeted gene replacement, without random mutations/insertions. A ΔSynK mutant was generated by inserting a kanamycin-resistance cassette into the Synechocystis genome, between nucleotide 122 [numbering from the A₁TG codon, as deduced from the work of Mitschke et al. (13)] and 479 of the slr0498 ORF. The linear DNA fragment used to transform cyanobacteria was designed to produce a large deletion of the ORF (from amino acids 41–159, including the pore region), by double-homologous recombination (Fig. 1A). Its insertion in the correct position of the genomic DNA was verified by PCR (Fig. 1B) and sequencing of the entire region (flanking regions included). Complete segregation of the recombinant chromosomes, usually reached by means of three to six subcloning passages, could be obtained in the ΔSynK mutant strain after the third subcloning and the cells were able to grow on plates with 50 μg mL⁻¹ kanamycin. Fig. 1C shows that both the SynK monomeric and its functionally active tetrameric forms (102 kDa) were completely missing from the mutant strain. Earlier work reports that prokaryotic potassium channels tend to migrate as tetramers even in the presence of SDS (e.g., refs. 2, 14). The homoplasmic character of the ΔSynK mutant makes it suitable to investigate the physiological role of this channel in a photosynthetic organism.

Fig. 1.

Construction of ΔSynK Synechocystis strain. (A) Schematic diagram of the construction of SynK knock-out. A large, central portion of the slr0498 sequence, enclosed between SpeI and BstXI sites, was deleted and substituted by a kanamycin-resistance cassette (as detailed in Materials and Methods). Positions of PCR primers used to verify correct insertion of the cassette and flanking regions through specific recombination are indicated. (B) Agarose gel electrophoresis of analytical PCR amplifications, performed on genomic DNAs from kanamycin-resistant control and ΔSynK strains: molecular mass marker (1); control DNA amplified with primers 1 and 3 (primer 1 anneals upstream of the SynK gene, whereas the latter anneals in the kanamycin cassette, almost 2,000 kbp away; therefore, no band can be detected) (2); control DNA amplified with primers 2 and 3 binding inside the kanamycin cassette (3); ΔSynK genomic DNA amplified with primers 1 and 3 comprising thus the flanking region of SynK and a portion of the kanamycin cassette (4); ΔSynK DNA amplified with primers 2 and 3 (to amplify the kanamycin cassette in the mutant strain) (5). (C) Western blot of total protein extracts from cyanobacterial cells, solubilized in sampling buffer. Cells were loaded at equal OD₇₃₀ (0.2 OD). The blot was developed with specific anti-SynK antibody: the monomeric (26 kDa) and the tetrameric (around 100-kDa apparent molecular mass) forms of the channel were detected in control but not in ΔSynK cells. The same blot was stripped and redeveloped with anti-PsaA antibody (55 kDa) (lower row) to show equal protein loading.

As mentioned above, cation fluxes could be responsible for lowering the electric (ΔΨ) component of the pmf (Δμ_H⁺). Assuming that SynK mediates such ion flux, its removal in ΔSynK should, therefore, enhance the ΔΨ in both dark and light conditions. We first assessed the consequences of the removal of SynK on the Δμ_H⁺ present in the dark, which is built by the activity of the respiratory chain present in thylakoid membranes. This possibility was evaluated by measuring the turnover of the cytochrome b₆f complex under single-turnover flashes regime. This parameter is controlled by the size of the pmf and, therefore, can be used to assess qualitatively the dark Δμ_H⁺ in vivo (15). Indeed, when dark adapted cells are exposed to single-turnover flashes, the pmf built in the light is negligible compared with the dark one generated by respiration. As shown in Fig. 2A, rereduction of oxidized cytochrome f by plastoquinol was largely enhanced by removal of the ΔpH with the H⁺/K⁺ exchanger nigericin in WT type cells, whereas a smaller effect was seen in ΔSynK (see also Fig. 2B). Nigericin mediates an electroneutral exchange of K⁺ for H⁺ and, thus, equalizes the concentration gradients for these two ions (16). Conversely, an opposite effect was seen when the ΔΨ was dissipated by addition of the protonophoric uncoupler carbonylcyanide-p-trifluoromethoxyphenyl hydrazone (FCCP) to nigericin-treated cells, the effect being larger in the mutant. In both cases, complete inhibition was seen upon treatment with 2,5-dibromo-3-methyl-6-isopropyl-p-benzoquinone (DBMIB), confirming that the measured reactions were triggered by plastoquinol oxidation (Fig. 2 A and B). Overall, these data suggest that the ΔΨ is bigger in ΔSynK cells under respiratory conditions, as expected in the frame of the “counterbalance” hypothesis. Similar conclusions were reached measuring directly the respiratory activity (oxygen consumption) of the cells (Fig. 2C): in WT cells, respiration was enhanced by nigericin (indicating removal of the “respiratory control” on electron flow), whereas a larger effect was seen in ΔSynK when both components of the pmf were suppressed by the simultaneous addition of nigericin and FCCP, as expected if the ΔΨ was larger in this mutant strain.

Fig. 2.

Formation of ΔpH across the thylakoid membrane is compromised in the ΔSynK mutant cyanobacteria. (A) Cytochrome f redox changes in Synechocystis WT (left) and ΔSynK (right) cells grown at 5 μmol of photons m⁻² s⁻¹. Cells were harvested in the exponential phase (OD ∼1) and concentrated 5 times before measurements in A, B, and C. Kinetics of cytochrome f redox changes. Squares, control; empty circles, nigericin 10 μM; filled circles, nigericin 10 μM plus FCCP 5 μM; downward triangles, DBMIB 5 μM. (B) Effect of nigericin, FCCP, and DBMIB on cytochrome b₆f turnover. Data as in A were deconvoluted using an exponential decay to evaluate the rates of cytochrome f rereduction. Enhanced rates indicate abolishment of the “photosynthetic control” by the ionophores. Inhibition of cytochrome f turnover by DBMIB demonstrates that this reaction depends on PQH₂ oxidation. (C) Consequences of nigericin and FCCP on respiration in WT and ΔSynK cells. They were dark incubated for 15 min in the presence of 10 mM glucose. Respiration was measured using a Clark-type electrode (Hansatech). Enhancement of respiration rates measured upon addition of nigericin and FCCP represent their suppression of the “respiratory control.” Data are relative to three independent measurements (± SD). Asterisks in B and C refer to significant changes with respect to WT and mutant strains under control condition (P < 0.05). (D) Representative AO fluorescence traces obtained from WT (black trace) and mutant (gray trace) organisms grown at 5 μmol of photons m⁻² s⁻¹. Downward deflections correspond to quenching of the AO fluorescence attributable to its protonation within the lumen following application of light, whereas upward deflections indicate dissipation of ΔpH when the light is switched off. Light with intensity of 700 μmol of photon m⁻² s⁻¹ for the indicated time was applied. (Right) ΔF/F values for WT (n = 10) and mutant (n = 11) cells. A ΔF/F ratio of 0.1864 ± 0.0071 (n = 10) vs. 0.149 ± 0.0074 were found in WT and ΔSynK, respectively. The difference is statistically significant (t test, P < 0.05).

Having ascertained that absence of SynK modifies the composition of the Δμ_H⁺ established in the dark, we investigated its possible role in photosynthesis. To our knowledge, no direct methods exist to measure the Δμ_H⁺ in the light in cyanobacteria. Thus, we used the fluorescent probe acridine orange (AO) to measure thylakoid membrane energization (17) in intact cells grown under nonphotoinhibitory conditions (5 μmol of photons m⁻² s⁻¹). In the light, we found that fluorescence quenching (i.e., the establishment of the pmf) was lower in the mutant than in the WT (Fig. 2D). In both lines, AO was taken up efficiently (Fig. S1A), and the signal was removed by addition of FCCP to dissipate the pmf (Fig. S1B) and by 20 μM 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) to block photosynthetic electron flow (Fig. S1C), confirming that AO quenching was measuring the proton gradient established in the light. We note that AO quenching was not affected by addition of Cs⁺ (80 mM) to intact WT cells (Fig. S2). Cesium is a general potassium channel inhibitor [SynK has been shown to be blocked by 10 mM Cs⁺ in patch clamp experiments (2)] capable of crossing the outer membrane (18), presumably through large porins, but it does not permeate the bacterial cytoplasmic membrane. Thus, we conclude that the plasma membrane (PM)-located SynK and/or other PM-located Cs⁺-sensitive potassium channels are not involved in the regulation of the pmf in the light. Overall, the above observations can be explained assuming that ΔSynK has a reduced capacity to generate a ΔpH in the light because it develops a higher ΔΨ (as already seen in the dark). Alternatively, it can be conceived that the reduced AO quenching observed in the mutant could reflect a reduced photosynthetic activity, attributable to a possible damage to some components of the electron transport chain.

The latter possibility was excluded by direct measurements of the photosynthetic activity. In low-light-grown cells (5 μmol of photons m⁻² s⁻¹), electron transfer efficiency was only slightly smaller in ΔSynK at low-intensity actinic light, as indicated by measurements of electron flow based on fluorescence-derived parameters [electron transfer rate (ETR)] (Fig. 3A). Both ETR (Fig. 3A) and oxygen evolution (Fig. S3A) decreased at higher actinic light intensities. In parallel, no changes in the composition of the photosynthetic apparatus were detected in low-light-grown cells, as indicated by 77K fluorescence spectroscopy (Fig. 3B) and Western blot analysis of the major complexes of the photosynthetic chain (Fig. 3C and Fig. S4A). The ratio of the peak intensities at 685/695 nm relative to 725 nm, which provide a rough estimation on the PSII/PSI ratio, suggests a very similar ratio for the WT and ΔSynK cultured at 5 μmol of photons m⁻² s⁻¹ (Fig. 3B). We also excluded changes in the partitioning of electrons between linear and cyclic electron flow. This was shown by the identical effect of DCMU (which blocks PSII and therefore only the linear electron flow) and of DBMIB (which blocks the cytochrome b₆f and therefore both linear and cyclic electron flow) on P₇₀₀ redox changes (Fig. S3B) in WT and ΔSynK cells.

Fig. 3.

Lack of SynK potassium channel reduces overall photosynthetic activity. (A) Apparent electron transport rate (ETR) determined by chlorophyll fluorescence measurements at different light intensities (abscissa) and calculated by using the DUAL-PAM program set are reported for WT and mutant cells grown at 5 μmol of photons m⁻² s⁻¹. Cells were kept at the indicated light intensities for 60 s between saturating pulses. Shown are mean ± SD values (n = 4 different cultures for WT and n = 4 different cultures for mutant). (B) A 77K low temperature fluorescence emission spectra measured in cells grown at 5 μmol of photons m⁻² s⁻¹. Data were normalized on PSI emission. Data are relative to three independent measurements (± SD). (C) Western blot analysis of subunit contents of photosynthetic complexes. WT and mutant cells grown in BG-11 supplemented with 5 mM glucose at 5 μmol of photons m⁻² s⁻¹ for 48 h were assayed. WT (lanes 1 and 3) and ΔSynK (lanes 2 and 4) samples were loaded at equal Chl (lanes 1 and 2) (1 μg) or at equal protein (lanes 3 and 4) (52 μg) concentrations on the same gel. Chl a is more abundant in PSI than in PSII (Chl a PSI: Chl a PSII ≥ 4). The upper part of the blot was developed with anti-PsaA antibody (PSI) (55 kDa), whereas the lower part of the same blot was used with anti-CP47 (PSII) (47 kDa). The figure shows that independently of the loading based on Chl or protein concentration, no change in PsaA/CP47 ratio could be observed between WT and mutant. (D–F) Experiments were performed as in A–C but on cells grown at 50 μmol of photons m⁻² s⁻¹. WT (lanes 1 and 3) and ΔSynK (lanes 2 and 4) samples were loaded at equal Chl (lanes 1 and 2) or at equal protein (lanes 3 and 4) concentrations on the same gel. In D, data were obtained from n = 7 different cultures for WT and n = 8 different cultures for mutant. In F, independently of the loading based on Chl or protein concentration, a similar change in PsaA/CP47 ratio could be observed between WT and mutant.

When activity was measured at high actinic light intensities, a decreased ETR was seen in the mutant, suggesting also a possible light sensitivity (Fig. 3A). Consistent with this, a strong reduction of photosynthesis was seen in cells grown at higher photon densities (50 μmol of photons m⁻² s⁻¹). AO fluorescence quenching was lower in the mutant than in the WT (Fig. S5A). Both ETR (Fig. 3D) and oxygen evolution (Fig. S5B) significantly decreased in the mutant, suggesting that damage to the photosynthetic apparatus had occurred. In fact, in this strain, the efficiency of ETR decreased with increasing time of actinic illumination, as expected in case of photoinhibition (Fig. S5 C and D). Biochemical assessment of different components of the photosynthetic chain indicated that PSII was down-regulated in ΔSynK, as shown by the specific decrease of the CP47 component of this photosystem (Fig. 3F and Fig. S4B). The lower PSII/PSI ratio was confirmed by measurements of 77 K fluorescence spectra (Fig. 3E, arrow). Furthermore, flow cytometric analysis indicate an increased chlorophyll content per cell in the mutant cells with respect to WT, in accordance with the observed change in PSII/PSI ratio (Fig. S6).

Despite the occurrence of photoinhibition, mutant cells maintained a similar membrane organization as in the WT, as assessed by transmission electron microscopy (TEM) (Fig. 4A and Fig. S7). However, their growth was visibly reduced with respect to that of the WT in these conditions, already after 4 d (Fig. 4B). Difference in growth was especially appreciable in the cells grown under stress conditions, e.g., at low temperature or at 500 μmol of photons m⁻² s⁻¹ light intensity in the presence of glucose. Indeed, photomixotrophic growth conditions are known to produce excess photooxidative damage (19, 20). The finding that this difference is present also in the case of cells grown without glucose indicates that the lack of SynK did not cause a general glucose intolerance but had an impact on the efficiency of photosynthesis. Fig. S8 illustrates the effect of long-term incubation (7 d) at 500 μmol of photons m⁻² s⁻¹ in liquid culture. These results are in agreement with the higher susceptibility of the mutant versus WT toward high light intensity and are suggestive of a pronounced oxidative stress in the mutant organisms cultured under stress conditions. Overall, growth defects were also observed after 6 d in the dark in the presence of external carbon source (5 mM glucose). This indicates that the changes in the pmf were already disadvantageous for cells in conditions where photosynthesis was not the major engine for cell growth.

Fig. 4.

Mutant cells do not display altered membrane organization but are more photosensitive with respect to WT organisms. (A) TEM of Synechocystis sp. PCC 6803 cells. WT and mutant cells were grown in BG-11 at 50 μmol of photons m⁻² s⁻¹ supplemented with 5 mM glucose for 48 h. (Scale bar: 0.1 μm.) cw, cellular wall; tm, thylakoid membrane. (B) Growth of WT and ΔSynK cells in triplicate on solid BG11 medium supplemented with 5 mM glucose [except when indicated without glucose (w/o)] under the indicated conditions. Photos were taken at day 4 except for growth in the dark and at low temperature (6 d). Optical densities (at 730 nm) at 0 time point are indicated. In the first two cases (dark and low light), spots obtained at 0.01 starting OD were omitted, because of lack of spots under these conditions characterized by slow growth. Spot tests repeated other three times gave the same results.

Discussion

The contribution of ion channels to the regulation of photosynthesis and the physiological role of potassium channels in prokaryotes are still largely unexplored. Here we show that SynK, a voltage-dependent potassium channel of Synechocystis, previously shown to be located in the thylakoid membrane in this organism, has a regulatory effect on setting the pmf across the thylakoid membrane. In particular, SynK seems to modulate the balance between the osmotic (ΔpH) and electric (ΔΨ) components of the transthylakoid proton gradient. Indeed, its removal in the mutant causes an increase in the ΔΨ. Using this mutant, we show that potassium flux via SynK importantly contributes to ion homeostasis and affects the efficiency of photosynthetic and respiratory electron transfer, which occur in the same membranes in cyanobacteria.

The light-driven transthylakoid pmf plays several essential roles in photosynthesis: it controls the rate of electron flow, the efficiency of light harvesting (by affecting the capacity of thermal dissipation of “excess excitation”), and feeds ATP synthesis (reviewed in ref. 4). Removal of SynK could therefore impact on photosynthesis at different levels. It has been proposed (4) that proton pumping into the lumen results in the building of a large ΔΨ component, which drives proton efflux through the ATP synthase (21), preventing the accumulation of high proton concentrations in the lumen. In this scenario, movements of thylakoid membrane-permeable counter ions, such as efflux of cations or influx of anions through ion channels would partially dissipate the ΔΨ, thereby allowing establishment of significant ΔpH. The ionic composition, as well as the activity of ion-flux pathways, in thylakoids is expected to determine the degree to which movement of ions can dissipate the ΔΨ (4). In accordance with this proposal, in ΔSynK, we find an increased ΔΨ and a decreased ΔpH. Beside enhancement of the activation of ATP synthase, the higher ΔΨ might have also additional effects. For example, it has been reported to restrain electron transport at the level of the cytochrome b₆f complex (22). Because electron transfer is more sensitive to ΔpH than ΔΨ, at high acidic luminal pH, the electron transfer would slow down because of the photosynthetic control, thus limiting the total pmf. In ΔSynK, a pmf is built in the light not by increased luminal proton concentration but rather by enhancing the potassium concentration in the lumen, because of the compromised exit of potassium from this compartment. Based on the data presented here, it is reasonable to assume that despite the changes in the balance between ΔpH and ΔΨ, the overall pmf is similar in the two strains, allowing a similar ATP synthesis. Because availability of this energy source often limits carbon assimilation in vivo (23), one may expect rather similar photosynthetic rates at moderate light intensities, as was indeed observed. Consistent with this, plants treated with low concentrations of nigericin which was able to collapse ΔpH without collapsing ΔΨ (nigericin allows exit of protons in exchange with entry of positively charged potassium, thus ensuring maintaining ΔΨ) do not exhibit major changes in the rate of electron flow through PSII (24). On the other hand, photosynthetic activity and electron flow are largely impaired in ΔSynK cultured at 50 μmol of photons m⁻²s⁻¹, demonstrating sustained photoinhibition. In plants, an acidic luminal pH is a key signal for initiating nonphotochemical quenching-mediated photoprotection. Thus, the reduced proton gradient in ΔSynK might a priori be responsible in part for the observed photosensitivity. On the other hand, no such pH dependent photoprotective mechanism exists in cyanobacteria (25) and other mechanism should be responsible for the observed loss of photosynthetic activity in ΔSynK cells. Previous work in microalgae (26) has demonstrated that the presence of a ΔΨ enhances recombination in PSII between chlorophyll cation P₆₈₀⁺ with downstream electron transport cofactors pheophytin (Phe⁻) or the primary stable electron acceptor plastoquinone Q_A⁻. This would promote the formation of ^TP₆₈₀, singlet oxygen, and photoinhibition (27). In our case, photoinhibition and reactive oxygen species (ROS) accumulation likely occur as indicated by the complete bleaching of the mutant culture grown at high light, by the decrease in photosynthetic oxygen evolution and by the changes in the PSII/PSI ratio. Therefore, it appears that the strategy consisting in enhancing the ΔpH at the expense of the ΔΨ under stress conditions in plants (28) could not only serve the purpose of increasing NPQ but also of decreasing ΔΨ-mediated charge recombination in PSII. In summary, lack of SynK potassium channel in cyanobacteria impacts ΔpH formation already at low-light intensities and leads to photoinhibition at higher light intensities. We cannot exclude that changes in the ion homeostasis within the cell could also, at least in part, alter photosynthesis and the interaction between the respiratory and photosynthetic activities (e.g., refs. 29–31), thus accounting for the observed reduced fitness.

All of the over 10 genomes of different cyanobacteria species completely sequenced to date contain at least 1 gene predicted to encode a potassium channel. However, for none of these putative channels the in vivo function was clarified, and in vitro activity was proven only for SynK (2). In general, although electrophysiological studies have been performed with success on native bacterial membranes (e.g., ref. 32), only very few studies addressed the physiological function of ion channels in prokaryotes. In cyanobacteria, the transport of various ions and solutes is linked to the formation of pmf consisting of pH gradient and membrane potential across the membrane. Although no information is available on regulation of photosynthesis by ion channels for cyanobacteria, in the case of higher plants, pharmacological experiments indicate involvement of potassium (11), calcium (33), and chloride-selective thylakoid channels (34) in the regulation of photosynthesis. Unfortunately, only very few ion channels/transporters of the higher plant thylakoid have been identified from molecular point of view, hampering a systematic study of their physiological roles. We have recently identified the two-pore potassium (TPK) channel TPK3, which seems to be related to SynK, in the thylakoid of Arabidopsis by using a specific monoclonal antibody (2). The electrophysiological activity, as well as the physiological role of TPK3, has still to be determined and homozygous mutants lacking TPK3 are not available in the seed banks to our knowledge. Chloride channel (ClC)-e also seems to be located in thylakoids (35), although another study located it to the chloroplast outer envelope membrane by proteomics (36). Plants lacking ClC-e did not display altered growth and showed a very subtle photosynthetic phenotype, and the underlying mechanism has not been investigated (35).

In summary, although the existence of ion channels in thylakoids has been known for several decades, their molecular identity is still largely unclear. Identification of SynK in the cyanobacterial thylakoids allowed us to understand that the potassium flux across this channel is necessary for maximally efficient photosynthesis. It has to be determined whether TPK3 plays a similar role in Arabidopsis. In higher plants, however, the high gene redundancy for potassium channels may hinder photosynthesis-related phenotype.

Materials and Methods

Synechocystis sp. PCC 6803 was maintained under photoheterotrophic growth conditions at 30 °C under white light in BG-11 medium at different light intensities. In the mutant cultures, kanamycin was included in the medium at a final concentration of 50 μg ml⁻¹. A kanamycin-resistance gene (kan^r), derived by PstI digestion from plasmid pUC4K (Pharmacia), was cloned into pEGFP-ΔSynK plasmid (2) and then used to transform WT Synechocystis sp. PCC 6803. SDS/PAGE and Western blots were performed according to standard protocol. Oxygen evolution was measured using Clark electrode, AO fluorescence was detected according to ref. 17, and chlorophyll fluorescence was measured using a Dual-PAM-100 (Walz). Cytochrome b₆f kinetics were measured using a JTS-10 spectrophotometer (Biologic) as described previously (37). P₇₀₀ redox changes were assessed at 705 nm, using the same instrumental setup. Electron microscopy was performed as in ref. 2, whereas flow cytometry was obtained by using FACSAntoII (BD-Pharmingen). For more experimental details, see SI Materials and Methods.