An Rrf2-Type Transcriptional Regulator Is Required for Expression of psaAB Genes in the Cyanobacterium Synechocystis sp. PCC 6803 1[W][OA]

Photosynthetic organisms must regulate photosystem stoichiometry (photosystem I-to-photosystem II ratio) under various light conditions. Transcriptional regulation of the psaAB genes is a critical process for this photoacclimation in cyanobacteria. In the course of our screening of transcriptional regulators in the cyanobacterium Synechocystis sp. PCC 6803, we found that chlorophyll accumulation was impaired in an Rrf2-type regulator Slr0846 mutant. DNA microarray and primer extension analyses showed that the expression of psaAB genes was markedly decreased in the mutant. Consistently, the mutant exhibited lower photosystem I-to-photosystem II ratio under normal light conditions, suggestive of decreased accumulation of the photosystem I reaction center. Gel-shift assay conﬁrmed that the Slr0846 protein bound to a far upstream promoter region of psaAB . These phenotypes of the mutant varied substantially with light conditions. These results suggest that Slr0846 is a novel transcriptional regulator for optimal expression of psaAB . # and slr0846-3 5 # -GTTGCCAAAAGACCAACG-3 # and cloned into pCR2.1 vector (Invitro- gen). The chloramphenicol resistance cassette was inserted into slr0846 at Ssp I in the same direction as slr0846 . Mutants were generated by transformation of Synechocystis cells with this DNA and selected on BG11 plates containing 50 m g mL 2 1 chloramphenicol. Complete segregation was conﬁrmed by PCR with the same primers as mentioned above.

Oxygenic photosynthesis utilizes two photosystems that are tandemly arranged to form the linear electron transport chain from water to NADP + . These photosystems are equipped with distinct light-harvesting systems and, as a result, are driven by the distinct light qualities and quantities. Since the light environment varies depending on location, time, and weather, it is important to acclimate to these environmental changes. It should be noted that not only optimization of the light reactions but also avoidance of damages due to excessive light excitation are critical for survival of the phototrophic organisms. To end this, the oxygenic phototrophs must separately regulate accumulation of the two photosystems and light-harvesting antenna systems. It has been established that the regulated accumulation of PSI is critical for longterm acclimation to the light conditions in cyanobacteria (Kawamura et al., 1979;Fujita, 1997;Hihara et al., 1998), while rapid inactivation and recovery of PSII seem to be effective for short-term acclimation to down-regulate the whole electron flow under highlight conditions (Mohamed et al., 1993;Tichý et al., 2003). As short-term acclimation, state transitions work for optimal energy distribution between PSI and PSII especially under low light (Mullineaux and Emlyn-Jones, 2005).
Indeed, expression of psaA and psaB, which encode the PSI reaction center subunits, is tightly regulated in the acclimation processes (Hihara et al., 1998;Herranen et al., 2005). Their expression was affected by the quality of light; it might be caused by changing the redox state of the photosynthetic electron transport chain. Promoter analysis revealed that psaA and psaB genes are cotranscribed from two common promoters, which may be regulated differentially by several cis-and trans-elements (Muramatsu and Hihara, 2006). In addition, gel-shift assay suggested that the psaAB promoters have some putative protein binding sites. The psaAB gene expression may be regulated by several transcription factors that are activated individually; however, there was no report to specify the factors for psaAB except recent report of RpaB (Seino et al., 2008).
Several redox-sensitive transcriptional regulators have been reported in Synechocystis. Slr1738, which is an oxidant-responsive PerR-like transcriptional regulator, represses sll1621, which encodes a type 2 peroxiredoxin carrying glutathione-dependent peroxidase activity Hosoya-Matsuda et al., 2005). Another reactive oxygen species-responsive transcriptional regulator, SufR, represses expression of the suf operon, which encodes components of an iron-sulfur-cluster assembly Suf system (Shen et al., 2007). PedR senses the activity of photosynthetic electron flow and regulates a subset of photosynthetic genes (Nakamura and Hihara, 2006). On the other hand, redox-responsive two-component signal transduction pathways were also implicated. A His kinase, Hik33 (also called DspA), and two related response regulators, RpaA and RpaB, seem to distinctively respond to different stresses (high light, salt, osmolarity, and low temperature) as a multistress regulatory system (Murata and Suzuki, 2006). RpaA and RpaB are the response regulator-type transcriptional regulators, although the multistress sensing and signal transduction have not yet been elucidated. There are many other works to study transcriptional regulation of genes in Synechocystis sp. PCC 6803 (Li and Sherman, 2000).
In the course of our screening of transcriptional regulators in Synechocystis, we found that an Rrf2-type regulator Slr0846 is critical for maintenance of normal chlorophyll accumulation. The Rrf2 family is one of the typical winged helix-turn-helix superfamily of the prokaryotic transcriptional regulators (Aravind et al., 2005). In many cyanobacterial genomes, there are one or two types of rrf2 genes, which are clustered into two distinct subgroups. However, there are no reports to describe their role or function in cyanobacteria. In the Synechocystis genome, slr0846 is the only rrf2 gene. Here, we describe the phenotype of Dslr0846 mutant and DNA binding of Slr0846 protein. We propose that Slr0846 is a novel factor that acts as a transcriptional activator for PSI reaction center genes, psaA and psaB.

Disruption of the slr0846 Gene
We disrupted slr0846 by insertion of the chloramphenicol cassette. The mutation was fully segregated under photoautotrophic conditions after several streaks on BG11 plates (see "Materials and Methods"; data not shown) and the mutants grew photoautotrophically, and the mutant phenotype mentioned below was reproducibly observed for independent clones. This indicates that slr0846 is dispensable for photoautotrophic growth. Under light conditions of 25 mmol photons m 22 s 21 , the Dslr0846 mutant grew slightly slower than the wild type (Fig. 1A), while it stopped growing after transient growth under the high-light conditions (100 mmol m 22 s 21 ; Fig. 1B). Similar growth inhibition after transient growth was observed under high-light conditions in a peroxiredoxin sll1621 mutant  and sll1961 mutant, which is defective in regulation of photosystem stoichiometry (Fujimori et al., 2005). These facts imply that the growth inhibition is due to accumulation of photooxidative stresses.

Global Gene Expression Profiles of the Dslr0846 Mutant
As Slr0846 is a putative transcriptional regulator, we compared by DNA microarray analysis the gene expression profile of the Dslr0846 mutant with the wildtype cells grown under normal light condition of 50 mmol m 22 s 21 (Table I). Clearly, the expression levels of psaA and psaB genes, which encode PSI-A and PSI-B polypeptides of the PSI reaction center, respectively, were approximately 8-fold lower in the mutant than in the wild type. Besides them, the expression levels of cpcB and cpcA genes (phycocyanin band a-subunits) and rbcS gene (Rubisco small subunit) decreased to a lesser extent (Table I). On the other hand, most genes upshifted in the Dslr0846 mutant are currently unknown in function (Table I). It is of note that expressions of many other PSI genes and PSII genes were hardly affected in the mutant (Supplemental Table S1).

Slr0846 Binds to the psaAB Upstream Region
We examined whether Slr0846 directly binds to the promoter of the genes by gel-shift assays using a recombinant protein with an N-terminal His-tag. Regarding the psaA promoter, we divided the upstream region into several fragments as shown in Fig. 2A. It was found that only the psaA-R5/R6 fragment (data not shown) but not the psaA-R7/R8 bound the Slr0846 protein (Fig. 2B). When the former was further divided into two subfragments, Slr0846 bound to only the far upstream subfragment psaA-6/5 ( Fig. 2C) but not to psaA-10/11 (Fig. 2D). The binding to the psaA-6/5 subfragment was strongly inhibited by a specific competitor DNA (Fig. 2C, lane 5). By contrast, the other candidate genes such as cpcB or sll0528 did not show any retardation with Slr0846 (data not shown). It is also of note that the Slr0846 protein bound to an upstream region of slr0846 itself (Fig. 2E), although the microarray data were masked due to expression of the antibiotics resistance cassette.
Next, we studied the decay of the psaAB transcripts in the presence of a transcription inhibitor rifampicin by the primer extension analysis (Fig. 3B). In agreement with the previous report of northern-blot analysis (Herranen et al., 2005), the psaAB transcripts of the wild type were quickly degraded under the light conditions and relatively stable under the dark conditions. The Dslr0846 mutant also showed a similar tendency. These results suggest that Slr0846 is not involved in the stabilization of the transcripts but acts as a transcriptional activator for both P1 and P2 promoters.

PSI-to-PSII Ratio
To investigate the physiological role of Slr0846, we studied pigmentation of the Dslr0846 mutant. The chlorophyll content per cell of the mutant was approximately 60% of the wild type, while the phycocyanin content was approximately 80% (Table II). When the PSI-to-PSII ratio was deduced from the lowtemperature chlorophyll fluorescence at 720 and 695 nm (F 720 for PSI and F 695 for PSII; Fig. 4B), the ratio of the mutant was approximately 46% of the wild type (Table III). The low-temperature fluorescence with phycocyanin excitation also confirmed the reduced fluorescence from PSI (Fig. 4C). These results suggest that the decreased PSI content was peculiar phenotype of the Dslr0846 mutant, as PSI mainly contributes to the cellular chlorophyll content in cyanobacteria. It is also consistent with the selective decrease in the psaAB expression.

The Phenotypes of the Dslr0846 Mutant under Different Light Conditions
For evaluation of the regulatory role of Slr0846, we compared the photoautotrophic growth and the PSIto-PSII ratio between the wild type and the Dslr0846 mutant under various light conditions. The growth of the mutant was comparable to that of the wild type under normal light conditions of white light as mentioned above (Fig. 1A), while the growth of the mutant was severely restricted under the high-light conditions (Fig. 1B). However, the PSI-to-PSII ratio under the normal light conditions was much lower (less than half) in the mutant than in the wild type, while the ratio was approximately two-thirds of the wild type under the high-light conditions ( Fig. 5; Table III). These results suggest that Slr0846 was active under the normal light conditions and suppressed under the high-light conditions.
Since the light quality also affects the PSI-to-PSII ratio, we studied the properties under conditions of PSII light and PSI light. Under weak PSII light that mainly excites PSII, the Dslr0846 mutant grew significantly slower than wild type (Fig. 6A). Both cells accumulate chlorophyll under the PSII light at a level comparable to that under normal white light (Figs. 4A and 6B). Accordingly, the low-temperature chlorophyll fluorescence showed that accumulation of PSI in the wild type under the PSII light was comparable to that under the normal white light (Figs. 4B and 6C). The disruption of slr0846 resulted in suppression of the accumulation (Fig. 6C). On the other hand, under the weak PSI light, the Dslr0846 mutant grew slowly at a rate similar to wild type (Fig. 6D). Both cells accumulate much less chlorophyll under the PSI light than the white or PSII light (Fig. 6E). It is noteworthy that the chlorophyll content relative to phycocyanin in the mutant cells was almost identical to that in the wildtype cells. The low-temperature fluorescence revealed that the PSI-to-PSII ratio was drastically decreased both in the wild type and the mutant (Fig. 6F). Namely, the fluorescence ratio of the mutant was close to that of the wild type in the PSI light (Table III). Thus, it is apparent that disruption of slr0846 gene gave rise to minimal effects on growth and accumulation of chlorophyll and PSI under the PSI light.

The Role of Slr0846
In this study, we demonstrated that Slr0846 is the transcriptional activator of psaAB by microarray analysis of the Dslr0846 mutant. Slr0846 protein directly bound to the far upstream promoter region of psaA. Primer extension analysis showed that the transcription from the two sites was similarly affected in the mutant. These results clearly demonstrated that the expression of the psaAB is regulated by the new transcriptional activator Slr0846. Since the transcriptional regulation of the psaAB is one of the most important events in the acclimation of the photosystems at least in cyanobacteria (Fujita, 1997; Hihara

Phylogenetic Analysis of Cyanobacterial Rrf2 Family
The Slr0846 belongs to the Rrf2 family, which includes putative bacterial transcriptional regulators such as Rrf2 in Desulfovibrio vulgaris (Rossi et al., 1993), IscR in Escherichia coli, NsrR in Nitrosomonas europaea (Beaumont et al., 2004), CymR in Bacillus subtilis (Even et al., 2006), etc. Slr0846 of Synechocystis appears to be a distant homolog of IscR of E. coli that regulates expression of an isc operon involved in assembly of iron-sulfur clusters (Schwartz et al., 2001). IscR itself possesses an unstable iron-sulfur cluster that senses redox changes in cells for optimal assembly of various iron-sulfur clusters. The redoxresponsive iron-sulfur cluster is coordinated by three conserved Cys residues and a Glu residue (Zeng et al., 2008). However, these ligand residues are not conserved in Slr0846. The Slr0846 homologs are highly conserved among many other cyanobacteria except Prochlorococcus (Fig. 7). It is suggested that the regulation of psaAB expression is widely distributed in many cyanobacteria.
Besides them, there is another group of Rrf2 family proteins conserved in many cyanobacterial genomes (Fig. 7). Many of them, such as tlr0503 in Thermosynechococcus elongatus, are tandemly arranged in the genome with cysK, which encodes Cys synthase (tlr0504). These findings suggest that the second group of Rrf2 in cyanobacteria may have a role homologous to CymR that regulates expression of CysK and other enzymes in the Cys metabolism in B. subtilis (Even et al., 2006). It is of note that this group of genes is absent in the Synechocystis genome.

The Mechanism of Transcriptional Activation by Slr0846
Generally, bacterial transcriptional regulators act as an activator or repressor depending on the location of the binding site relative to the core promoter where the RNA polymerase binds (Browning and Busby, 2004). For example, IscR of E. coli binds to a slight upstream of 235 sequence and induces expression of some target genes, while it also binds to a site partly overlapped with 210 sequence and represses expression of other target genes (Giel et al., 2006). Our data revealed that the Slr0846-binding sequence is located more than 150 bp upstream from the transcription start point P1 in the psaA-6/5 region, although the precise binding motif has not yet been elucidated. cis-Elements in far upstream from the core promoter are not so popular but have been found in certain genes such as malK for CRP binding (Danot et al., 1996) and nir for NarL binding (Wu et al., 1998). In these cases, transcription factors such as CRP and NarL work together with other regulator proteins for transcriptional initiation of RNA polymerase. Slr0846 may also be associated with other transcriptional regulator(s) for induction of psaAB. . The spectra were normalized to the peak at 620 nm. B and C, Fluorescence emission spectra of the wild type (solid line) and Dslr0846 mutant cells (dashed line). After dark adaptation, cells were frozen and the spectra were recorded at 77 K with excitation of chlorophyll at 435 nm (A) or phycocyanin at 600 nm (B). The spectra were normalized to the PSII emission at 690 nm. The cis-elements of the psaAB have been analyzed by promoter-deletion study, and it was found that the two transcription start points (P1 and P2) of psaAB have multiple cis-elements and are transcribed in a different manner (Muramatsu and Hihara, 2006). The HNE2 (2417 to 2323 based on the position numbering in this study), which is nearly identical to the psaA-6/5 region of our study, is partly responsible for the highlight-induced repression, suggesting that a high-lightresponsive repressor binds to a sequence within the HNE2 region. Taken together, it is assumed that the activator Slr0846 and the yet unknown repressor may interact with each other within psaA-6/5 region.

Signal(s) That Slr0846 Recognizes
At the moment, the amino acid sequence of Slr0846 does not possess any peculiar features for signal sensing. The Rrf2 family including Slr0846 belongs to the winged helix-turn-helix superfamily, which may have evolved from the canonical helix-turn-helix proteins in evolution of prokaryotic transcriptional regulators (Aravind et al., 2005;Wilkinson and Grove, 2006); N-terminal region of Rrf2 family proteins usually acts for DNA binding and C-terminal region often serves for signaling. For example, the redox sensor IscR of E. coli possesses Cys residues for coordination of a redox-responsive iron-sulfur center in the C-terminal region. Some Rrf2 proteins bind a small molecule for transcriptional regulation (O-acetylserine for Bacillus CymR [Even et al., 2006] and nitric oxide  , respectively. The spectra were normalized to the peak at 620 nm. C and F, Fluorescence emission spectra of the wild type (solid line) and Dslr0846 mutant (dashed line). The cells were grown under PSII light (C) or PSI light (F) as in B and E. After dark adaptation, cells were frozen and the spectra were recorded at 77 K with excitation at 435 nm. The spectra were normalized to the PSII emission at 690 nm.
for Neisseria gonorrhoeae NsrR [Isabella et al., 2009]). Recently, Bacillus CymR was reported to bind an enzyme Cys synthase for regulation of its DNA-binding activity (Tanous et al., 2008). Consistently, sequence alignment shows that the N-terminal region of Slr0846 is widely conserved in cyanobacteria and other bacteria, while the C-terminal region is conserved within the cyanobacterial homologs. This suggests that the C-terminal region may act in signal transduction.
Based on the 77 K chlorophyll fluorescence (Table  III), the deduced PSI-to-PSII ratio of the Dslr0846 mutant under the PSI light was very low and comparable to that of the wild type. These facts suggest that the transcriptional activation of psaAB due to Slr0846 was minimal under the PSI light. When cells were grown under the PSII light, the ratio of the mutant was approximately 2.5-fold higher than that under the PSI light (1.56 versus 0.63 in Table III), while the ratio of the wild type increased over 4-fold (3.03 versus 0.73 in Table III). This suggests that Slr0846 contributes to the accumulation of PSI under the PSII light. Slr0846 may monitor light or redox conditions to optimize the linear electron transport. Although no sequence feature has been detected for binding of a chromophore or a redox component, Slr0846 might bind some signaling molecule derived from the intersystem or a protein like in CymR regulation (Tanous et al., 2008). It should be noted that yet unknown factor(s) are also involved in regulation of PSI accumulation between PSI and PSII light. On the other hand, the high-lightinduced down-regulation of PSI was not much affected in the Dslr0846 mutant and will be discussed with regard to a two-component regulator RpaB (see below).

Other Genes Regulated by Slr0846
The DNA microarray data suggest that the cpcBA genes also decreased by the Dslr0846 disruption (Table  I). In addition, phycocyanin content is lower in the Dslr0846 mutant than the wild type. However, Slr0846 did not bind to the upstream sequence of the cpcBA genes (data not shown). Thus, these effects are caused by a side effect of the decreased PSI. Most genes upregulated in the Dslr0846 mutant (Table I) are reported to be induced by several kinds of stress, such as strong light, UV light, high salt, and cold (Hihara et al., 2001;Huang et al., 2002;Marin et al., 2004;Shoumskaya et al., 2005). Expression of these genes may be indirectly regulated by Slr0846, perhaps via accumulation of PSI. In cyanobacteria, expression levels of the psaAB genes decrease under high light (Muramatsu and Hihara, 2003). Even though the Dslr0846 mutant has a lower PSI-to-PSII ratio than the wild type, the growth of the Dslr0846 mutant was inhibited under high-light condition. In this case, the Dslr0846 mutant easily turns into overreduced conditions because of the excess PSII.

Other Regulatory Factors of psaAB
The high-light-induced down-regulation of psaAB can be accounted for by a transcriptional regulator RpaB (Rre26), since it binds to a high-light-responsive element1 just upstream of the P1 promoter of psaA (Eriksson et al., 2000;Seino et al., 2008). RpaB is an OmpR-type response regulator, of which DNA binding may be activated by phosphorylation from the stress-responsive His kinase Hik33 under the normal light conditions (Tu et al., 2004;Kappell and van Waasbergen, 2007), and suppressed by dephosphorylation under high-light conditions (Seki et al., 2007). Hik33-RpaB system also seems to regulate high-lightinducible genes such as hliA by acting as a transcriptional repressor under normal light conditions. It is obvious that RpaB is a global regulator responsive for stresses including high light. Regarding the PSI genes, RpaB binds to promoter sequences of other PSI genes (i.e. psaC, psaD, psaE, psaF, psaK1, and psaL; Seino et al., 2008). On the other hand, the microarray analysis revealed that Slr0846 does not regulate these genes but only psaAB (Supplemental Table S1). Since  psaAB genes are essential for assembly of the whole PSI complex (Smart and McIntosh, 1991), Slr0846 may specifically regulate expression of psaAB genes in contrast with rather global regulation of RpaB on PSI genes.

Strains and Growth Conditions
The original motile strain of Synechocystis sp. PCC 6803 showing positive phototaxis was used as the wild type (Ikeuchi and Tabata, 2001). The slr0846 gene was disrupted by insertion of the chloramphenicol resistance gene derived from pACYC184. The wild type and the Dslr0846 mutant were grown at 30°C in BG11 medium supplemented with 20 mM TES-KOH (pH 7.8; Rippka, 1988) with bubbling of 1% (v/v) CO 2 under continuous illumination of white fluorescent lamps (30-50 mmol m 22 s 21 ). Alternatively, cells were cultured under weak orange LED light, which we designated as PSII light, with a l max of 610 nm and a 20 nm half band width (TLOH180P; TOSHIBA) at 10 mmol m 22 s 21 or weak red LED light, which we designated as PSI light, with a l max of 700 nm and a 20 nm half band width (L700-03AU; Marubeni) at 10 mmol m 22 s 21 . Cell density was monitored as optical density at 730 nm with a spectrophotometer (model UV-2400PC; Shimadzu).
Escherichia coli strain JM109 was used for cloning and subcloning of plasmids, while BL21 (DE3) was used for expression of His-Slr0846 with pET28a (Novagen). Cells were grown in Luria-Bertani medium. When required, kanamycin or ampicillin was added at a concentration of 20 or 50 mg mL 21 , respectively.

Gene Disruption
A DNA fragment of 1,465 bp containing slr0846 was amplified by PCR using primers slr0846-1 5#-GGATGTCCCCCTTAAATT-3# and slr0846-3 5#-GTTGCCAAAAGACCAACG-3# and cloned into pCR2.1 vector (Invitrogen). The chloramphenicol resistance cassette was inserted into slr0846 at SspI in the same direction as slr0846. Mutants were generated by transformation of Synechocystis cells with this DNA and selected on BG11 plates containing 50 mg mL 21 chloramphenicol. Complete segregation was confirmed by PCR with the same primers as mentioned above.

Analyses of the Photosynthetic Parameters
Chlorophyll content was calculated after extraction with 100% methanol as described (Grimme and Boardman, 1972). Absorption spectra were recorded using a spectrophotometer (model U-3500S; Hitachi) equipped with an end-on photomultiplier. Cells at log phase were harvested and resuspended at cell density of OD 730 = 2.0. Phycocyanin content was calculated from the absorption spectra as described (Arnon et al., 1974).
Seventy-seven K fluorescence spectra were recorded using a spectrofluorometer (RF-5300PC; Shimadzu). Cells at log phase were harvested and resuspended at 5 mg chlorophyll mL 21 . After dark adaptation for 10 min, cells were frozen in liquid N 2 . The bandwidth of the excitation light was 10 nm.

RNA Isolation
Cells were collected by centrifugation at 6,000g for 10 min at 4°C and stored in liquid N 2 . The frozen cells were thawed with 500 mL of a buffer containing 10 mM Tris-HCl (pH 8.0), 1 mM EDTA, and 2% (w/v) SDS at 65°C and immediately treated with 500 mL of phenol at 65°C for 15 min. After centrifugation, the supernatant was extracted several times with phenol/ chloroform. To eliminate trace amounts of contaminating DNA, RNA samples were incubated with RNase-free DNaseI (Takara) for 30 min at room temperature. After precipitation with ethanol, RNA was dissolved in water.