Transcriptional regulation of CmpR, the LysR family protein involved in CO2-responsive gene regulation in the cyanobacterium Synechococcus elongatus

The cmpR gene of the cyanobacterium Synechococcus elongatus PCC 7942 encodes a LysR family protein that activates the cmpABCD bicarbonate transporter operon under the conditions of CO2 limitation. Using the luxAB reporter genes, the promoter of the cmpR gene was shown to be transiently activated under the low-CO2 conditions. Much higher levels of luxAB expression were observed in a ∆cmpR background even under high-CO2 conditions, revealing a repressive role of cmpR in regulation of the cmpR promoter. Exposure of the ∆cmpR strain to the low-CO2 conditions further increased the luxAB expression level, showing the involvement of an additional transcriptional activator responding to CO2 limitation. From the ∆cmpR strain, two mutant strains showing slower growth, lower contents of phycocyanin and even higher levels of PcmpR-luxAB expression than the parental strain were isolated and shown to have a partial deletion in the ORF of cpcE, encoding the alpha subunit of phycocyanobilin lyase. Expression of plasmid-borne cpcE in the mutants restored the growth and reduced the luciferase expression level to those observed in the parental strain, while partially restoring the phycocyanin content. These results indicated occurrence of multiple layers of regulation of cmpR expression in addition to the autoregulation by CmpR. Correspondence to: Tatsuo Omata, Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya-464-8601, Japan, Tel: +81 52 789 4106; Fax: +81 52 789 4107; E-mail: omata@agr.nagoya-u.ac.jp


Introduction
Though the atmospheric CO 2 concentration is rising because of combustion of fossil fuels, it is not high enough to saturate ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco), the enzyme responsible for photosynthetic CO 2 fixation, under ambient air in vitro; O 2 serves as the alternative substrate of Rubisco to CO 2 and hence interferes with the carboxylation of ribulose 1,5-bisphosphate (RuBP). To cope with this problem, photosynthetic organisms have developed various "CO 2 concentrating mechanisms" (CCMs) to raise the CO 2 concentration in the vicinity of Rubisco to enhance the carboxylation reaction.
Cyanobacteria are photoautotrophic prokaryotes capable of plant-type, oxygenic photosynthesis. They have an efficient CCM, which consists of two steps: (i) intracellular accumulation of high concentrations of ; and (ii) localized conversion of HCO 3 to CO 2 in the carboxysomes, the polyhedral inclusion bodies to which Rubisco is sequestered. Intracellular accumulation of HCO 3 is mediated by active HCO 3 transporters and specialized NADPH dehydrogenase complexes that convert CO 2 to HCO 3 in an energy-dependent manner [1][2][3][4][5]. The intracellular conversion of CO 2 to HCO 3 lowers the intracellular CO 2 level low, allowing for uptake of CO 2 from the external medium via passive diffusion. For both the HCO 3 and CO 2 uptake mechanisms, those having high-affinity for the respective substrates are distinguished from those having low-affinity for the substrates. The genes related to the high-affinity HCO 3 and CO 2 uptake mechanisms are generally repressed under high-CO 2 (HC) conditions, i.e., when cells are incubated under air supplemented with 1-5% (v/v) of CO 2 . Since transcription of these genes are sharply induced by incubation of the HC-grown cells under low-CO 2 (LC) conditions, i.e., under air containing 0.04% or lower CO 2 concentrations, they are thought to be LC-inducible genes [6][7][8].
LysR-type transcriptional regulators (LTTRs) of CbbR subfamily were first identified in chemoautotrophic bacteria and phototrophic bacteria as the regulator of the genes encoding the enzymes of the Calvin-Benson-Bassham (CBB) cycle [9]. In cyanobacteria, CbbR homologs were shown to be the major player in regulation of the LCinduced expression of the CCM-related genes; Synechocystis sp. PCC 6803 has two closely-related CbbR homologs, one of which (CmpR) was identified as the activator of the cmp operon, which encodes the subunits of an ABC-type high-affinity HCO 3 transporter [10], while the other (NdhR or CcmR) was shown to serve as a repressor under HC conditions, inhibiting expression of several CCM-related genes including its own gene, the sbtA gene encoding the Na + -dependent high-affinity HCO 3 transporter, and the ndhF3 operon encoding the components of the high-affinity CO 2 uptake mechanism [6,7,11]. Synechococcus elongatus PCC 7942, on the other hand, has only one CbbR homolog designated CmpR, which was shown to be the activator of the cmpABCD operon [10]. The CmpR protein of S. elongatus PCC 7942 was shown by in vitro experiments to bind to the promoter of the cmpABCD operon in the presence of 2-phosphoglycolate (2-PG), one of the products of the RuBP oxygenation reaction [12]. It is therefore supposed that 2-PG serves as the signal of CO 2 deficiency in vivo to activate transcription of the cmpABCD operon. It has remained unclear whether the S. elongatus CmpR protein regulates its own expression as many of the LTTR proteins do. Using a luxAB reporter gene transcriptionally fused to the cmpR promoter, it is shown in this study that the protein not only activates the cmp operon under LC conditions, but represses its own expression under HC conditions. Studies on the reporter strain reveal an additional layer(s) of regulation of the promoter, involving CmpR-independent mechanism(s) responding to the LC conditions. Genetic studies on the strain further point to a novel relationship between the integrity of the light harvesting system and the activity of the promoter.

Strains and growth conditions
A derivative of S.elongatus strain PCC7942 (SPc), which is cured of the resident pUH24 plasmid [13], was used as the parental strain of all the mutants used in this study. For construction of a ΔcmpR mutant, a 0.8-kbp DNA region upstream of cmpR (nucleotides -798 to +3 with respect to the translation start site) and a 0.9-kbp region downstream of cmpR (nucleotides +1 to +888 with respect to the termination codon) were amplified by PCR using the primer pairs P1/P2 and P3/ P4, respectively ( Table 1). The chloramphenicol resistance gene was also amplified by PCR using the primers P5 and P6, which respectively included the sequences complementary to the primers P2 and P3. The three DNA fragments were joined by overlap extension PCR using the primers P1 and P4. The resulting DNA fragment was used to transform the SPc strain to construct a chloramphenicol-resistant ∆ccmR mutant (MR5) via replacement of the cmpR gene with the chloramphenicol resistance gene ( Figure 1A). After three serial streak purifications of single colonies, successful genome segregation was confirmed by PCR using the primers P9 and P10 ( Figure 1B).
For construction of a transcriptional fusion of the promoter of cmpR (P cmpR ) and the luxAB ORFs encoding a bacterial luciferase, the 115-bp DNA region upstream of cmpR (nucleotides -115 to -1 with respect to the translation start site, Figure 1C), including 45 bases of the 3' portion of the synpcc7942_1309 ORF, was amplified by PCR, using the primers P7 and P8 carrying added NotI and BamHI recognition sequences, respectively (Table 1), and cloned between the NotI and BamHI sites of the pAM1414 plasmid carrying a promoter-less luxAB gene cluster [14]. The P cmpR ::luxAB fusion thus constructed was transferred into the neutral site 1 (NS1) locus on the chromosome of the SPc and MR5 strains to construct the CW and CT strains, respectively.
The liquid medium used for cultivation of the cyanobacterial strains was a modification of the BG11 medium containing 15 mM KNO 3 and 20 mM HEPES-KOH (pH 8.2) [15]. Cells were grown at 30℃ under continuous illumination provided by fluorescent lamps at a light intensity of 50 μE m -2 s -1 . Cultures were aerated with ambient air (low-CO 2 conditions) or air supplemented with 2% (v/v) CO 2 (high-CO 2 conditions). Solid medium was prepared by addition of 1.5 % (w/v) Bactoagar (Difco) to the liquid medium. Where appropriate, chloramphenicol (5 µg ml -1 ), spectinomycin (20 µg ml -1 ) and kanamycin (15 µg ml -1 ) were added to the media.
For random mutagenesis in the CT strain, cells in liquid medium were irradiated with a 15W UVC lamp (254 nm) for 4 min (survival rate ~ 0.2%) and then treated for 24 h with 40 µg ml -1 of ampicillin in the light under LC conditions. The cells were then spread onto solid medium prepared in rectangular plates of 128 mm × 86 mm in size and incubated under the HC conditions for 10 days. The agar plates, which contained a total of 17,000 colonies, were subsequently incubated under low CO 2 conditions and screened for mutant colonies showing altered expression levels of bioluminescence.

Bioluminescence measurements
Time-course measurements of bioluminescence from colonies of the P cmpR ::luxAB reporter strains were performed essentially as described previously [16,17] using an automated bioluminescence monitoring apparatus (model LL04-1,Churitsu Electric Co., Nagoya, Japan); Cells of CT and CW strains were plated on solid medium and incubated under HC conditions for 7 days. Agar pieces carrying 10 to 20 colonies were sealed in 35-mm petri dishes together with 30 μl of 3% n-decanal solution in cooking oil and 100 μl NaHCO 3 (10%, w/v), and incubated under illumination for 16 h. Expression of P cmpR ::luxAB was induced by removing the NaHCO 3 solution from the petri dish, and bioluminescence was measured at 30-min intervals. The data were analyzed using the SL00-01 bioluminescence-analyzing software (SL00-01,Churitsu Electric Co.) [18]. Bioluminescence intensity was expressed as cps per colony.
Identification of the mutant colonies derived from the UVCirradiated population of the luxAB reporter strain CT was carried out using a high-throughput real-time bioluminescence monitoring system equipped with a CCD image sensor (Churitsu Electric Corp., Nagoya, Japan; http://www.churitsu.co.jp/products/bio/highthroughput.html) essentially as described by Kondo and Ishiura [19].
For measurements of bioluminescence from the E3 and E11 mutants, which form only tiny colonies on solid medium, cells grown in liquid medium under HC conditions were collected by centrifugation, re-suspended in fresh medium to give an OD 730 value of 0.5, and incubated under illumination with aeration with ambient air. 0.5 ml aliquots of the cell suspension were occasionally withdrawn from the culture, mixed with 20 µl of 3% n-decanal solution, incubated for 5 min and then subjected to photon counting for 30 s using an AQUACOSMOS/VIM system (Hamamatsu Photonics, Hamamatsu, Japan).

Re-sequencing analysis of genomic DNA
Genomic DNA was extracted from Synechococcus cells using a Wizard Genomic DNA Purification Kit (Promega), after treatment of the cells with lysozme as follows; Cells were collected and re-suspended in 450 μl of a buffer containing 50 mM Tris-HCl (pH 8.0), 50 mM NaCl and 5 mM EDTA and after mixing with 50 μl of a lysozyme solution (50 mg ml -1 ), incubated at 37°C for 1 h. The cells were subsequently collected by centrifugation at 15,000 g at RT for 2 min and lysed by mixing with 0.6 ml of the Nuclei Lysis Solution of the Promega kit. Resequencing of the genomic DNA was performed using a SOLiD 5500 system (Life Technologies) as described previously [20].

Expression of the plasmid-borne cpcE in Syenchococcus.
An 859-bp DNA fragment carrying the coding region of cpcE was amplified from the genomic DNA of S. elongatus PCC 7942 by PCR using the primers P11 and P12 carrying added NcoI and XbaI recognition sequences, respectively ( Table 1). The amplified DNA fragment was digested with NcoI and XbaI and ligated between the NcoI and XbaI sites of the shuttle expression vector pSE1 [21]. The resulting plasmid was transformed into the cells of E3 and E11 to yield the strains E3R and E11R.

RNA extraction and semi-quantitative RT-PCR
25 ml aliquots of the cultures of S. elongatus (OD 730~0 .5) were used for RNA extraction, following the protocol reported by Ruffing [22]. From the lysate/ethanol mixture thus obtained, RNA was purified using a SV Total RNA Isolation System (Promega). The isolated total RNA was used for the synthesis of cDNA using a SuperScript III First-Strand Synthesis System (Life Technologies). The obtained cDNA was used as the template for semi-quantitative RT-PCR analysis using the primer pairs specific to luxA, cpcF and rnpB (P13-P18; Table 1). The cycle numbers for PCR amplification was 30 for luxA, 30 for cpcF, and 26 for rnpB.

CmpR-dependent and independent mechanisms of cmpR regulation
As previously reported for the cmpR insertional mutant MR4 [10], cells of the cmpR deletion mutant MR5 grew as fast as the parental SPc strain both under LC and HC conditions (data not shown). Accordingly, the P cmpR ::luxAB strains CW and CT, which were constructed from SPc and MR5 respectively, formed colonies of essentially the same size (data not shown). Low CO 2 response of the CW and CT strains were therefore compared by measuring the bioluminescence emitted from colonies on the surface of agar plates, using a real-time bioluminescence monitoring system ( Figure 1D). When HC-grown cells of the CW strain, which carries the wild-type cmpR gene, was transferred to LC conditions, there was a sharp induction of bioluminescence ( Figure 1D, CW, open symbols). After reaching a peak at t ~ 4 h, the intensity of bioluminescence declined to a lower level. These results indicated that transcription of cmpR is activated by CO 2 limitation. Unlike the CW strain, the reporter strain deficient in cmpR (CT) showed high level of bioluminescence even when grown under the HC conditions ( Figure  1D, CT, t = 0), showing that CmpR autoregulates its own expression as many other LTTRs do. After transfer of CT cells to LC conditions, further increase in the luciferase expression level was observed ( Figure  1D, CT, open symbols); The bioluminescence level increased for about 15 h after the transfer and remained at the high level thereafter. The results indicated the presence of a CmpR-independent mechanism(s) for LC-responsive activation of cmpR.

Characterization of the mutants with altered P cmpR regulation
To gain insight into the CmpR-independent mechanism(s) of cmpR regulation, CT cells were treated with UVC irradiation, subjected to an ampicillin enrichment procedure under LC conditions, plated on solid medium and screened for colonies showing altered levels of bioluminescence, using a high-throughput bioluminescence monitoring system equipped with a CCD image sensor. Among the mutant candidates, two (designated E3 and E11) were found to reproducibly show slow growth compared to the parental CT strain irrespective of the CO 2 conditions (Figure 2). Since these mutants formed only tiny colonies, their bioluminescence intensity was compared with that of the parental CT strain using cells grown in liquid cultures and re-suspended in fresh growth medium to give an OD 730 value of 0.5 (Figure 4). The results confirmed that the expression level of luciferase was much higher in the E3 and E11 strains than in the parental CT strain (Figure 4).
To identify the mutations that caused activation of the cmpR promoter in the E3 and E11 mutants, genome re-sequencing analysis of the CT, E3 and E11 mutants was performed ( Table 2). The CT strain had no mutation as compared with the parental SPc strain except for the replacement of the cmpR gene with the chloramphenicol resistance gene and the insertion of a spectinomycin resistance gene and the PcmpR::luxAB reporter construct into the NS1 site. E3 and   . Induction of bioluminescence from the CT strain and its derivatives. Cells grown under the HC conditions in liquid medium were collected by centrifugation, re-suspended in fresh liquid medium to give an OD 730 value of 0.5 at t = 0, and aerated with ambient air containing 0.05% CO 2 . 500 μl aliquots from the cultures were withdrawn at the time points indicated, mixed with 20 μl of 3% n-decanal and subjected to photon counting. Data shown are from one of the three sets.  E11 respectively had several mutations on the genome, but they had a common deletion of 121-bp in the cpcE gene, which forms an operon with the cpcF gene located downstream. The cpcE-cpcF operon encodes the subunits of the heterodimeric phycocyanin-α-subunit phycocyanobilin lyase, which mediates attachment of phycocyanobilin to the α subunit (CpcA) of phycocyanin (PC) [23], a blue-colored light-harvesting protein consisting of the α and β subunits encoded by the highly conserved cpcA and cpcB genes. In accordance with the previously reported properties of the cpcE deletion mutant of S. elongates [24], both E3 and E11 showed significantly reduced amounts of PC as revealed by their absorption spectra, showing lower absorption peak at 625 nm as compared with the parental CT strain (Figure 3).
To determine whether the CpcE deficiency was the cause of the growth defect and the enhanced activation of P cmpR under LC condition, a plasmid carrying the cpcE ORF downstream of the P trc promoter (see methods) was introduced into E3 and E11 to construct the E3R and E11R strains. Expression of cpcE from the P trc promoter could only partially restore the PC content (Figure 3), presumably due to the polar effect of the 121-bp deletion in the chromosomal cpcE gene on the expression of the cpcF gene. Nevertheless, growth of the E3R and E11R strains were comparable to that of CT, indicating that the amount of PC was sufficient to support normal photosynthetic growth under the given conditions (Figure 2). The level of bioluminescence was also reduced to that observed in the parental CT strain by expression of cpcE.
Semi-quantitative RT-PCR analysis of the transcripts from the luxA gene showed that the gene is expressed at a much higher level in E3 than in CT under the LC conditions. Also, expression of plasmid-borne cpcE reduced the transcript level to a level comparable to that in CT. These results confirmed that it was the activity of the P cmpR promoter that responded to the presence or absence of the cpcE gene ( Figure 5).

Discussion
The CmpR protein of S. elongatus PCC 7942 activates transcription of the cmpABCD operon, which encodes the subunits of a high-affinity ABC HCO 3 transporter, under the conditions of CO 2 limitation [10]. Micromolar concentrations of 2-PG were shown to enhance the binding of the protein to the promoter of the cmp operon in vitro, suggesting that the product of the RuBP oxygenation reaction acts as a co-inducer of CmpR under the LC conditions [12]. In addition to its role as an activator of gene expression, the present results show that CmpR acts as a repressor of its own expression under the HC conditions ( Figure  1). Thus, CmpR has two distinct modes of interaction with DNA; the one involving 2-PG as a co-activator and the other independent of 2-PG. The CcmR (NdhR) protein of Synechocystis sp. PCC 6803, the homolog of CmpR, is involved in repression of various CCM related genes under HC conditions [6,7,11]. By comparison of the promoters of these genes, a sequence ATAG-N 8 -CTAT has been proposed as the consensus binding motif for CcmR [6]. Since the cmpR promoter of S. elongatus has two of this sequence ( Figure 1C), CmpR presumably binds to these sites under HC in a 2-PG independent manner to repress transcription.
Since the P cmpR ::luxAB fusion is activated under LC conditions even in the CT strain deficient in CmpR ( Figure 1D, CT), it is deduced that S. elongatus has a CmpR-independent mechanism for LC-responsive activation of P cmpR . In an attempt to identify the genes relevant to this mechanism, we aimed at isolation of mutants that are defective in growth under LC conditions because of impaired LC response. The UVC-treated cells were hence subjected to an ampicillin enrichment procedure under LC conditions. The two mutant strains that showed a slow-growth phenotype were, however, found to grow slowly irrespective of the CO 2 conditions ( Figure 2) and to show stronger P cmpR induction than the parental CT strain under LC conditions (Figures 4 and 5). The cpcE mutation in the E3 and E11 strains is clearly responsible for the slow-growth phenotype and hyper-induction of P cmpR (Figures 2 and 5). The slow growth is presumed to be due to the reduced capacity of the cells to absorb light energy (Figure 3), but the underlying molecular mechanism that links CpcE and P cmpR activity is currently unclear. It is commonly supposed that the function of CCM becomes more important under high-light conditions, where the energy supply tends to exceed the supply of CO 2 to Rubisco. The activation of P cmpR in the cpcE mutant cells having smaller capacity of light absorption than the parental strain is therefore novel. Possible involvement of CpcE, a heat repeat protein, in direct interaction with P cmpR cannot be excluded, but it seems more likely that the reduced photosynthetic electron transport and the resulting metabolic change have relevance to the activation of P cmpR in the CmpR-deficient cpcE mutants. Although further study is required, the present results reveal a novel link between the light-harvesting capacity of the cells and regulation of the key regulator of CCM.