A Carbon Dioxide Limitation-Inducible Protein, ColA, Supports the Growth of Synechococcus sp. PCC 7002

A limitation in carbon dioxide (CO2), which occurs as a result of natural environmental variation, suppresses photosynthesis and has the potential to cause photo-oxidative damage to photosynthetic cells. Oxygenic phototrophs have strategies to alleviate photo-oxidative damage to allow life in present atmospheric CO2 conditions. However, the mechanisms for CO2 limitation acclimation are diverse among the various oxygenic phototrophs, and many mechanisms remain to be discovered. In this study, we found that the gene encoding a CO2 limitation-inducible protein, ColA, is required for the cyanobacterium Synechococcus sp. PCC 7002 (S. 7002) to acclimate to limited CO2 conditions. An S. 7002 mutant deficient in ColA (ΔcolA) showed lower chlorophyll content, based on the amount of nitrogen, than that in S. 7002 wild-type (WT) under ambient air but not high CO2 conditions. Both thermoluminescence and protein carbonylation detected in the ambient air grown cells indicated that the lack of ColA promotes oxidative stress in S. 7002. Alterations in the photosynthetic O2 evolution rate and relative electron transport rate in the short-term response, within an hour, to CO2 limitation were the same between the WT and ΔcolA. Conversely, these photosynthetic parameters were mostly lower in the long-term response of a few days in ΔcolA than in the WT. These data suggest that ColA is required to sustain photosynthetic activity for living under ambient air in S. 7002. The unique phylogeny of ColA revealed diverse strategies to acclimate to CO2 limitation among cyanobacteria.


Introduction
Oxygenic photosynthesis is the most popular anabolic biological activity on earth. From photon energy, water (H 2 O), and atomospheric carbon dioxide (CO 2 ), oxygenic phototrophs synthesize photosynthates, including glycogen, starch, and cellulose. This process involves both light and dark reactions in the photosynthetic electron transport system on the thylakoid membrane and the Calvin-Benson cycle in the stroma in chloroplasts, respectively. In the photosynthetic electron transport system, photon energy excites the reaction center chlorophylls (Chl), P680 and P700 in photosystems (PS) II and I, respectively, to drive photosynthetic linear electron transport. In PSII, H 2 O is oxidized to oxygen (O 2 ) by photo-oxidized P680, and electrons derived from P680 are transferred to photo-oxidized P700 in PSI through plastoquinone, cytochrome b 6 /f complex, and plastocyanin or cytochrome c 6 . These processes result in the formation of a proton gradient across the thylakoid membrane to produce ATP with chloroplast ATP synthase. Conversely, electrons originating in P700 are transferred to NADP + via ferredoxin and ferredoxin-NADP + reductase. Subsequently, NADPH and ATP are produced by photosynthetic linear electron transport. In the stroma, ribulose 1,5-bisphosphate (RuBP) carboxylase/oxygenase, also known as Rubisco, catalyzes the carboxylation of RuBP with

Growth of S. 7002 WT and ∆colA under Ambient Air and High CO 2
To investigate the effects of ColA on the growth of the cyanobacterium S. 7002, we constructed a mutant deficient in the gene SYNPCC7002_A2492, ∆colA (Supplemental Figure S1), and monitored the growth of S. 7002 WT and ∆colA under ambient air and high CO 2 conditions. Cell growth, reflected in the increase in optical density at 750 nm (OD 750 ) of the cultures, was not significantly different between the WT and ∆colA in either ambient air or high CO 2 conditions ( Figure 1A,B). However, the Chl content of the ∆colA culture was lower than that of the WT under ambient air but not high CO 2 conditions ( Figure 1A,B), which was recognized by the pale color of the ∆colA cell culture ( Figure 1C). These data suggest that ColA deletion causes a decrease in the amount of Chl in S. 7002 when CO 2 is limited.

Growth of S. 7002 WT and ΔcolA under Ambient Air and High CO2
To investigate the effects of ColA on the growth of the cyanobacterium S. 7002, we constructed a mutant deficient in the gene SYNPCC7002_A2492, ΔcolA (Supplemental Figure S1), and monitored the growth of S. 7002 WT and ΔcolA under ambient air and high CO2 conditions. Cell growth, reflected in the increase in optical density at 750 nm (OD750) of the cultures, was not significantly different between the WT and ΔcolA in either ambient air or high CO2 conditions ( Figure 1A,B). However, the Chl content of the ΔcolA culture was lower than that of the WT under ambient air but not high CO2 conditions ( Figure 1A,B), which was recognized by the pale color of the ΔcolA cell culture ( Figure 1C). These data suggest that ColA deletion causes a decrease in the amount of Chl in S. 7002 when CO2 is limited. . Closed and open symbols, respectively, show the optical density at 750 nm (OD750) and Chl content. Black symbols represent S. 7002 WT; red symbols represent, ΔcolA. Measurements were independently conducted three times, and the data are shown as the mean ± SD. Differences in Chl content between the WT and ΔcolA were analyzed using Student's t-test. Asterisks indicate statistically significant differences in Chl content between the WT and ΔcolA at p < 0.05. (C) Photographs of the culture flasks taken on the indicated days.
Next, we investigated the nitrogen (N) content in S. 7002 WT and ΔcolA, because in cyanobacteria, a decrease in Chl content can result from a decrease in N content [18]. Therefore, the . Closed and open symbols, respectively, show the optical density at 750 nm (OD 750 ) and Chl content. Black symbols represent S. 7002 WT; red symbols represent, ∆colA. Measurements were independently conducted three times, and the data are shown as the mean ± SD. Differences in Chl content between the WT and ∆colA were analyzed using Student's t-test. Asterisks indicate statistically significant differences in Chl content between the WT and ∆colA at p < 0.05. (C) Photographs of the culture flasks taken on the indicated days.
Next, we investigated the nitrogen (N) content in S. 7002 WT and ∆colA, because in cyanobacteria, a decrease in Chl content can result from a decrease in N content [18]. Therefore, the Chl contents in WT and ∆colA were evaluated based on the N content. As shown in Figure 1A, the Chl content per OD 750 in ∆colA was lower than that in the WT under ambient air but not high CO 2 (Figure 2A). The N content per the OD 750 in ∆colA was generally lower than that in the WT under ambient air, although this difference was not significant ( Figure 2B). Furthermore, the Chl content per N in ∆colA was significantly lower, by about 70%, than that in the WT in ambient air growth conditions ( Figure 2C), which indicates that the lack of ColA impacted the distribution of N under ambient air in S. 7002. Photo-oxidative damage in PSII and PSI is derived from photon energy absorbed by Chl [6,16], which suggests that oxygenic phototrophs can alleviate the potential risk of the photo-oxidative damage by decreasing the Chl content. It is possible that the lower Chl level in ∆colA is a secondary effect of the absence of ColA in cyanobacterial cells, allowing it to escape the excess supply of photon energy under CO 2 limitation conditions. Conversely, in both the WT and ∆colA, the Chl and N contents per OD 750 were likely to be lower in the cells grown in ambient air than in those grown in high CO 2 ( Figure 2). The first step in N assimilation is driven by nitrate reduction catalyzed by nitrate reductase with NAD(P)H as the electron donor, which is reversibly inactivated under CO 2 -limited conditions [19]. This process might support flexible modulation of N use in S. 7002 in response to changes in the amount of available CO 2 . Chl contents in WT and ΔcolA were evaluated based on the N content. As shown in Figure 1A, the Chl content per OD750 in ΔcolA was lower than that in the WT under ambient air but not high CO2 ( Figure 2A). The N content per the OD750 in ΔcolA was generally lower than that in the WT under ambient air, although this difference was not significant ( Figure 2B). Furthermore, the Chl content per N in ΔcolA was significantly lower, by about 70%, than that in the WT in ambient air growth conditions ( Figure 2C), which indicates that the lack of ColA impacted the distribution of N under ambient air in S. 7002. Photo-oxidative damage in PSII and PSI is derived from photon energy absorbed by Chl [6,16], which suggests that oxygenic phototrophs can alleviate the potential risk of the photo-oxidative damage by decreasing the Chl content. It is possible that the lower Chl level in ΔcolA is a secondary effect of the absence of ColA in cyanobacterial cells, allowing it to escape the excess supply of photon energy under CO2 limitation conditions. Conversely, in both the WT and ΔcolA, the Chl and N contents per OD750 were likely to be lower in the cells grown in ambient air than in those grown in high CO2 ( Figure 2). The first step in N assimilation is driven by nitrate reduction catalyzed by nitrate reductase with NAD(P)H as the electron donor, which is reversibly inactivated under CO2-limited conditions [19]. This process might support flexible modulation of N use in S. 7002 in response to changes in the amount of available CO2.  light grey bars indicate ∆colA. Measurements were conducted 8 to 10 and 2 to 3 days after inoculation in cells grown in ambient air and high CO 2 , respectively. Data are shown as the mean ± SD of three independent experiments. Differences between the WT and ∆colA were analyzed using Student's t-test. Asterisks indicate statistically significant differences between the WT and ∆colA at p < 0.05.

Oxidative Stress in S. 7002 WT and ∆colA under Ambient Air
We evaluated the degree of oxidative stress in the WT and ∆colA grown in ambient air. Typical high-temperature thermoluminescence of the cyanobacterial cells reflects the extent of oxidative stress originating from lipid peroxidation in the cells [20]. To measure this stress, cells grown in ambient air were analyzed after adaptation to the dark for 30 min. The luminescence peak, at approximately 140 • C, was higher in ∆colA than in the WT ( Figure 3A), indicating that lipid peroxidation by ROS was occurring in ∆colA.

Oxidative Stress in S. 7002 WT and ΔcolA under Ambient Air
We evaluated the degree of oxidative stress in the WT and ΔcolA grown in ambient air. Typical high-temperature thermoluminescence of the cyanobacterial cells reflects the extent of oxidative stress originating from lipid peroxidation in the cells [20]. To measure this stress, cells grown in ambient air were analyzed after adaptation to the dark for 30 min. The luminescence peak, at approximately 140 °C, was higher in ΔcolA than in the WT ( Figure 3A), indicating that lipid peroxidation by ROS was occurring in ΔcolA. Next, we quantitatively measured the carbonylated proteins in the WT and ΔcolA grown in ambient air. Protein carbonylation is caused by sugar-and lipid-derived reactive carbonyls. Sugarderived reactive carbonyls, including methylglyoxal, glyoxal, and 3-deoxy glucosone, are enzymatically and non-enzymatically produced through sugar metabolism, such as glycolysis and the Calvin-Benson cycle [21,22]. On the other hand, lipid-derived reactive carbonyls, including acrolein, 4-hydroxy-2-nonenal, and malondialdehyde, are produced via lipid peroxidation by ROS [23]. These carbonyls react with amino acid residues, such as lysine and cysteine, to inactivate protein function [24,25]. The amount of carbonylated proteins is an indicator of oxidative stress. Crude soluble and membrane extracts were obtained from cyanobacterial cells grown in ambient air. Protein carbonylation was detected using an antibody specific to 2,4-dinitrophenylhydrazine (DNPH), which conjugates with carbonyl groups. The absence of ColA resulted in an increase in protein carbonylation in both crude soluble and membrane fractions of S. 7002 ( Figure 3B). We detected the Next, we quantitatively measured the carbonylated proteins in the WT and ∆colA grown in ambient air. Protein carbonylation is caused by sugar-and lipid-derived reactive carbonyls. Sugar-derived reactive carbonyls, including methylglyoxal, glyoxal, and 3-deoxy glucosone, are enzymatically and non-enzymatically produced through sugar metabolism, such as glycolysis and the Calvin-Benson cycle [21,22]. On the other hand, lipid-derived reactive carbonyls, including acrolein, 4-hydroxy-2-nonenal, and malondialdehyde, are produced via lipid peroxidation by ROS [23]. These carbonyls react with amino acid residues, such as lysine and cysteine, to inactivate protein function [24,25]. The amount of carbonylated proteins is an indicator of oxidative stress. Crude soluble and membrane extracts were obtained from cyanobacterial cells grown in ambient air. Protein carbonylation was detected using an antibody specific to 2,4-dinitrophenylhydrazine (DNPH), which conjugates with carbonyl groups.
The absence of ColA resulted in an increase in protein carbonylation in both crude soluble and membrane fractions of S. 7002 ( Figure 3B). We detected the increased signal particularly around 50 kDa in the crude fractions of ∆colA, as indicated by the white arrows in Figure 3B. Cyanobacteria harbor a variety of enzymes that scavenge reactive carbonyls [26,27]. Therefore, the amounts of reactive carbonyls in cyanobacterial cells are determined by both the production and detoxification rates in cells. The impacts of ColA deletion on metabolic alterations of reactive carbonyls should be investigated in future studies.

Responses of Photosynthesis in S. 7002 WT and ∆colA to CO 2 Limitation
Based on a lower conversion of N into Chl and the oxidative stress in ∆colA under ambient air conditions, ColA was assumed to have a functional role in the suppression of photo-oxidative damage in acclimating to the situations where photosynthetic CO 2 assimilation is suppressed. To investigate the effects of ColA on photosynthesis in response to CO 2 -limited conditions, we measured the responses of photosynthesis to CO 2 limitation over both the short-(within an hour) and long-term (a few days) in S. 7002 WT and ∆colA.
To measure the short-term response, we followed a previously developed method (see the Section 3.8.) [1,11]. Time courses, for both O 2 in the reaction medium and the relative Chl fluorescence of the cells, were simultaneously measured during the transition of photosynthesis from CO 2 -saturated to CO 2 -limited phases in S. 7002 WT and ∆colA grown under high CO 2 ( Figure 4A,B). We calculated both the photosynthetic O 2 evolution rate ( Figure 4C) and relative electron transport rate (ETR) at PSII ( Figure 4D) from the data shown in Figure 4A,B. Unexpectedly, no difference was detected in the photosynthetic parameters between the WT and ∆colA in the transition to CO 2 -limited conditions ( Figure 4). We evaluated the photosynthetic O 2 evolution rate, based on N content, considering the difference in Chl content between the WT and ∆colA (Figure 2). The same photosynthetic responses in ∆colA and in WT suggest that ColA does not directly affect photosynthesis and the electron transport reaction. On the other hand, in both the WT and ∆colA, the O 2 evolution rates five minutes after adding sodium bicarbonate (NaHCO 3 ) (approximately 20 µmol O 2 mg −1 N h −1 ) were lower than those five minutes after turning on the actinic light (AL) (approximately 30 µmol O 2 mg −1 N h −1 ), corresponding to the relative ETR ( Figure 4C,D).
considering the difference in Chl content between the WT and ΔcolA (Figure 2). The same photosynthetic responses in ΔcolA and in WT suggest that ColA does not directly affect photosynthesis and the electron transport reaction. On the other hand, in both the WT and ΔcolA, the O2 evolution rates five minutes after adding sodium bicarbonate (NaHCO3) (approximately 20 µmol O2 mg −1 N h −1 ) were lower than those five minutes after turning on the actinic light (AL) (approximately 30 µmol O2 mg −1 N h −1 ), corresponding to the relative ETR ( Figure 4C,D).  Next, we investigated the photosynthesis response to CO 2 -limited conditions on the longer timeframe of a few days. For this measurement, WT and ∆colA cells were grown under high CO 2 conditions, until the OD 750 reached 2.0 to 3.0 ( Figure 1B), and then the cells were inoculated into freshly prepared A + medium with the OD 750 adjusted to 2.0, and then they were grown in ambient air. We measured both the photosynthetic O 2 evolution rate and relative ETR before, one day after, and three days after the transition to ambient air growth conditions. In both the WT and ∆colA, photosynthetic O 2 evolution rates and relative ETR gradually decreased with time in ambient air, but the extent of the decrease in photosynthetic parameters was significantly larger in ∆colA than in the WT ( Figure 5A,B). The dependence of these photosynthetic parameters on photon flux density before and three days after the transition was compared ( Figure 5C−F). Although little difference was found in these photosynthetic parameters between the WT and ∆colA before the cells were shifted to ambient air ( Figure 5C,D), ∆colA showed a significantly lower photosynthetic O 2 evolution rate and relative ETR, particularly at higher photon flux densities, than those in the WT three days after the cells were exposed to ambient air ( Figure 5E,F). Unfortunately, we could not determine the reason why photosynthetic activity decreased to a greater extent in ∆colA, but the greater oxidative stress in ∆colA shown in Figure 3 suggests that the absence of ColA accelerated the photo-oxidative damage and inactivated photosynthetic activity in S. 7002 in ambient air. We assessed the level of total oxidizable P700 (P m ), which is a marker of PSI activity [6], in the WT and ∆colA, which indicated that P m in ∆colA (0.52 ± 0.07, n = 3) was almost the same as that in the WT (0.53 ± 0.04, n = 3), three days after the transition to ambient air growth conditions. In this experiment, we adjusted the concentration of the cyanobacterial cells in the reaction medium to 80 µg N mL −1 . These data suggest that the larger decrease in photosynthetic activity in ∆colA may be due to the inactivation of PSII, rather than of PSI. relative ETR, particularly at higher photon flux densities, than those in the WT three days after the cells were exposed to ambient air ( Figure 5E,F). Unfortunately, we could not determine the reason why photosynthetic activity decreased to a greater extent in ΔcolA, but the greater oxidative stress in ΔcolA shown in Figure 3 suggests that the absence of ColA accelerated the photo-oxidative damage and inactivated photosynthetic activity in S. 7002 in ambient air. We assessed the level of total oxidizable P700 (Pm), which is a marker of PSI activity [6], in the WT and ΔcolA, which indicated that Pm in ΔcolA (0.52 ± 0.07, n = 3) was almost the same as that in the WT (0.53 ± 0.04, n = 3), three days after the transition to ambient air growth conditions. In this experiment, we adjusted the concentration of the cyanobacterial cells in the reaction medium to 80 µg N mL −1 . These data suggest that the larger decrease in photosynthetic activity in ΔcolA may be due to the inactivation of PSII, rather than of PSI.  Time courses of O2 evolution rate and relative electron transport rate (ETR) after the cells were shifted from high CO2 to ambient air conditions. The photosynthetic O2 evolution rate was measured at 1100 µmol photons m −2 ·s −1 . (C−F) Dependence of the O2 evolution rate and relative ETR on photon flux density (C,D) before and (E,F) three days after the transition. Photosynthetic parameters in the reaction medium containing the cells (10 µg Chl mL −1 and 10 mM NaHCO3) were measured independently three times. Data are shown as the mean ± SD. Black circles, S. 7002 WT; red triangles, ΔcolA. Differences between the WT and ΔcolA were analyzed using the Student's t-test. Asterisks indicate statistically significant differences between the WT and ΔcolA at p < 0.05.

Phylogeny of ColA Gene Homologs among Cyanobacteria
Gene homologs encoding ColA are conserved in some, but not all, cyanobacteria other than S. 7002. For example, S. 6803 harbors sll0218 as a ColA gene, whereas S. 7942 has no gene encoding ColA. The gene sll0218 in S. 6803 is highly expressed under CO2-limited conditions, like that observed rate and relative electron transport rate (ETR) after the cells were shifted from high CO 2 to ambient air conditions. The photosynthetic O 2 evolution rate was measured at 1100 µmol photons m −2 ·s −1 . (C−F) Dependence of the O 2 evolution rate and relative ETR on photon flux density (C,D) before and (E,F) three days after the transition. Photosynthetic parameters in the reaction medium containing the cells (10 µg Chl mL −1 and 10 mM NaHCO 3 ) were measured independently three times. Data are shown as the mean ± SD. Black circles, S. 7002 WT; red triangles, ∆colA. Differences between the WT and ∆colA were analyzed using the Student's t-test. Asterisks indicate statistically significant differences between the WT and ∆colA at p < 0.05.

Phylogeny of ColA Gene Homologs among Cyanobacteria
Gene homologs encoding ColA are conserved in some, but not all, cyanobacteria other than S. 7002. For example, S. 6803 harbors sll0218 as a ColA gene, whereas S. 7942 has no gene encoding ColA. The gene sll0218 in S. 6803 is highly expressed under CO 2 -limited conditions, like that observed for SYNPCC7002_A2492 in S. 7002 [16], and is proposed to stabilize PSII [28]. We aligned the amino acid sequences of ColA isozymes in S. 7002 and S. 6803 (Supplemental Figure S2) and found a 68% similarity between the primary structures of these gene products in the ClustalW analysis. We constructed the phylogenetic tree of ColA isozymes based on amino acid sequences of some cyanobacteria (Figure 6), which indicated that the genes for ColA are normally encoded between the genes for FLV4 and FLV2, designated the flv4-2 operon, by Aro' and colleagues, in cyanobacterial genomes [16,28]. In fact, all the cyanobacteria species registered in CyanoBase [29] possess genes for FLV2/4, similar to S. 6803, with S. 7002 as the only exception ( Figure 6). Curiously, S. 7002 has a gene for ColA alone, without those for FLV2/4. In S. 6803, the physiological role of ColA encoded by sll0218 is independent of that of FLV2/4, even though these genes exist in the same operon [28], supporting the idea that ColA functions alone in S. 7002 (Figures 1-5). The physiological functions of ColA in other FLV2/4-coding cyanobacteria species might provide some insights into the evolutional process of the relationship between ColA and FLV2/4.
We considered the transfer of ColA genes to some cyanobacteria species. In this hypothesis we assumed that the cyanobacteria, like S. 7002, harboring ColA alone diverged, and that ColA had been strongly selected for acclimation to CO 2 -limited conditions. Thereafter, the genes for FLV2/4 were generated by the duplication of genes for FLV1/3 that merged into the location of the gene coding ColA to form the flv4-2 operon in the genomes of some cyanobacteria, which would have been advantageous for cyanobacteria requiring mechanisms to prevent photo-oxidative damage under CO 2 limitation conditions. In this study, we constructed a mutant of S. 7942 (#HA-ColA) with the gene for ColA introduced into S. 7002 in the form of a hemagglutinin (HA)-fusion protein (Supplemental Figure S3) and compared the long-term response of photosynthesis to limited CO 2 to that of a control strain expressing HA alone (#HA) (Supplemental Figure S4). Unfortunately, the heterologous expression of HA-tagged ColA did not affect the response of photosynthesis to limited CO 2 in S. 7942 (Supplemental Figure S4), implying that another molecular mechanism is responsible for ColA-dependent acclimation of S. 7002 to CO 2 limitation.  Figure S4), implying that another molecular mechanism is responsible for ColA-dependent acclimation of S. 7002 to CO2 limitation. In this study, we found that ColA supports the acclimation of S. 7002 to ambient air growth conditions (Figures 1−5). ColA gene expression was 690-fold higher in cells grown under ambient air than in those grown under high CO2 conditions [17]. Interestingly, gene homologs for ColA are conserved only in a limited number of cyanobacteria, and, even among those cyanobacterial species, their ColA genes are incorporated into the flv4-2 operon (Figure 6) [28]. That is, S. 7002 is unique in possessing ColA alone. The deletion of ColA impacted the N use (Figure 2), oxidative stress status (Figure 3), and photosynthetic activity (Figures 4 and 5) in ambient air. These data imply that other In this study, we found that ColA supports the acclimation of S. 7002 to ambient air growth conditions (Figures 1-5). ColA gene expression was 690-fold higher in cells grown under ambient air than in those grown under high CO 2 conditions [17]. Interestingly, gene homologs for ColA are conserved only in a limited number of cyanobacteria, and, even among those cyanobacterial species, their ColA genes are incorporated into the flv4-2 operon (Figure 6) [28]. That is, S. 7002 is unique in possessing ColA alone. The deletion of ColA impacted the N use (Figure 2), oxidative stress status (Figure 3), and photosynthetic activity (Figures 4 and 5) in ambient air. These data imply that other oxygenic phototrophs without ColA use mechanisms that are different from that of S. 7002 to acclimate to limited CO 2 . We believe that ColA in S. 7002 is evidence of one of the diverse strategies of oxygenic phototrophs used to facilitate cyanobacterial life on earth, and that such a unique strategy for the acclimation to CO 2 limitation possibly gives us an opportunity to increase a biomass yield for biotechnological applications in the present atmosphere, where the partial pressure of CO 2 is about 40 Pa.
The latest research [28] has concluded that ColA functions in the process of PSII assembly and repair in S. 6803, which can explain the results in this study. The impairment of the PSII assembly and repair process causes the decreases in Chl and PSII content [28], possibly leading to oxidative stress. Additionally, PSII assembly and repair is a highly coordinated process that requires all components to be carefully controlled, which may be consistent with the failure to complement ColA into S. 7942 (Supplemental Figure S4). In S. 6803, a PSII assembly factor YCF48 is required for the expression of ColA (i.e., Sll0218) [28]. The isozyme for YCF48 is also found in the genome of S. 7002, which is encoded in SYNPCC7002_A0229 [29]. As described above, the cyanobacterium S. 7002 possesses ColA but not FLV2/4. Therefore, S. 7002 is expected to provide novel insights into the detailed role of ColA, as a distinct component and separate player from that of the FLV2/4 heterodimer, in protection of PSII in future studies.

Growth Conditions and Determination of Chla
Cultures of S. 7002 were kept under continuous light conditions (25 • C, 50 µmol photons m −2 s −1 from a fluorescent lamp) on A + medium agar plates [30]. For physiological measurements, the cells were inoculated into A + liquid medium and were grown on a rotary shaker (100 rpm) in continuous light (25 • C, 150 µmol photons m −2 ·s −1 from a fluorescent lamp) at high CO 2 (2% (v/v)) or ambient air. The optical density (OD) of the cultures at 750 nm was measured with a spectrophotometer (U-2800A, Hitachi, Tokyo, Japan).
Cells of the S. 7942 mutants, grown in BG-11 medium rather than A + medium, proliferated as well as S. 7002 [31].
For Chl measurement, cells from 0.1-1.0 mL of culture medium were centrifugally harvested and resuspended by vortexing in 1 mL 100% (v/v) methanol. After incubation at room temperature for 5 min, the suspensions were centrifuged at 10,000× g for 5 min. Total Chla in the supernatants were spectrophotometrically determined [32].

Bioinformatics
All of the cyanobacterial gene sequence data used in this study were obtained from CyanoBase (http://genome.microbedb.jp/CyanoBase) [29].
To construct the phylogenetic tree of gene homologs for ColA, we searched for gene homologs among cyanobacteria using the BLAST tool in CyanoBase. Amino acid sequences of these gene products were aligned in ClustalW [33]. The evolutionary history of these genes was inferred using the neighbor-joining method [34]. The percentages of replicate trees, in which the associated taxa clustered together in the bootstrap test with 500 replicates, are shown next to the branches [35]. The tree was drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Poisson correction method [36] and are expressed in the number of amino acid substitutions per site. Evolutionary analyses were conducted in MEGA7 [37].
To predict the transmembrane helices in ColA (Supplemental Figure S2), we used a classification and secondary structure prediction system (SOSUI) [38].

Statistical Analysis
Student's t-tests were applied to detect differences. All statistical analyses were performed using Microsoft Excel 2010 (Microsoft, Washington, DC, USA) and JMP8 (SAS Institute Inc., Tokyo, Japan).

Generation of Mutants
To construct the mutant ∆colA, the genomic region encoding SYNPCC7002_A2492 was amplified from the genomic DNA of S. 7002 by PCR (primer set: forward, CAGCCTCTCCGT CAATGGCTCAA; reverse, CCAGGCGGGATTATTACATGACC) and then cloned into the pGEM-T Easy vector (Promega, Tokyo, Japan).
The recombinant plasmid was linearized and amplified by inverse polymerase chain reaction (PCR) (primer set: forward, TGACCGTGTGCTTCTGAAGTGTCGATTGTTTCTGTGCT; reverse, TCTTACGTGCCGATCAAC GAAACAGCTCTACCAAAATCAG) and then applied to the In-Fusion cloning system (Takara, Shiga, Japan) with a chloramphenicol resistance gene cassette originating from the pACYC184 vector [11,39]. S. 7002 was transformed with the resulting plasmid using the standard procedure [40]. The mutant was selected on a 0.5% A + medium agar plate containing chloramphenicol (15 µg·mL −1 ), and complete segregation was confirmed by PCR (Supplemental Figure S1).
To introduce the HA-tagged ColA into S. 7942, we used pNSHA based on pNSE1 harboring a spectinomycin resistance gene in the fragment derived from the neutral site [41]. The coding region for ColA was amplified from the genome of S. 7002 and then cloned into pNSHA, digested with BamHI and HindIII, using an In-Fusion cloning system (Takara, Shiga, Japan). The resulting plasmid was used to transform WT S. 7942 according to the method used for S. 7002. Heterologous expression of HA-tagged ColA was achieved in the light for 2 h in the presence of isopropyl β-D-1-thiogalactopyranoside, which was examined by immunoblotting as described below.

Immunoblot Analysis
For the immunodetection of HA-tagged ColA, cyanobacterial cell cultures (20 mL) were harvested by centrifugation, and the pellets were suspended in 500 µL extraction buffer (50 mM HEPES-KOH, 800 mM sorbitol, 20 mM CaCl 2 , and 1 mM phenylmethylsulfonyl fluoride; pH 7.5). The suspensions were homogenized with glass beads using Bug Crasher GM-01 (Taitec Co., Aichi, Japan) at 4 • C for 20 min and thereafter were centrifuged at 8000× g for 15 min at 4 • C. The supernatants were incubated at room temperature for 1 h in Laemmli SDS sample buffer with 6 M urea and then centrifuged at 13,000× g for 10 min at 4 • C. The resulting supernatants (10 µL each) were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). After electrophoresis, the proteins were electrotransferred to a polyvinylidene fluoride membrane and detected by HA-specific antibody (Aviva Systems Biology, San Diego, CA, USA).
To analyze protein carbonylation, 20 mL cyanobacterial cell cultures were harvested by centrifugation, and the pellets were suspended in a 500 µL extraction buffer (50 mM HEPES-KOH, 800 mM sorbitol, 20 mM CaCl 2 , and 1 mM phenylmethylsulfonyl fluoride; pH 7.5). The suspensions were homogenized with glass beads using Bug Crasher GM-01 (Taitec Co., Aichi, Japan) at 4 • C for 20 min, and thereafter were centrifuged at 8000× g for 15 min at 4 • C, resulting in supernatants and pellets. The supernatants were centrifuged twice more at 13,000× g for 30 min at 4 • C, and the resulting supernatants were treated as crude soluble fractions. Conversely, the pellets were mildly vortexed in the extraction buffer with 2% (w/v) octyl β-D-glucopyranoside for 30 min at 4 • C. After centrifugation at 13,000× g for 30 min at 4 • C, the resulting supernatants were treated as crude membrane fractions.
Protein concentrations in both crude fractions were determined using a Pierce 660 nm Protein Assay (Thermo Scientific, Rockford, IL, USA) with bovine serum albumin as the standard. Both crude soluble and membrane fractions (12 and 4 µg protein per lane, respectively) were analyzed by SDS-PAGE. After electrophoresis and electrotransfer to a polyvinylidene fluoride membrane, the proteins were derivatized with 2,4-dinitrophenylhydrazine (DNPH) and then detected by polyclonal antibody specific to DNPH groups [42] purchased from Cosmo Bio (Tokyo, Japan).

Measurement of Nitrogen
Cyanobacterial cells were centrifugally harvested and dried overnight at 60 • C. Dried pellets were digested using the Kjeldahl method with sulfuric acid and hydrogen peroxide (H 2 O 2 ). Total nitrogen (N) content was determined using Nessler's reagent after adding sodium potassium tartrate and sodium hydroxide (NaOH) [18].

Thermoluminescence
Lipid peroxidation was monitored by measuring the in vivo thermoluminescence from 20 • C to 160 • C at a heating rate of 1 • C min −1 using a Thermoluminescence System TL 500 (Photon Systems Instruments, Brno, Czech Republic) [43].

Measurement of O 2 and Chl Fluorescence
Net uptake and evolution of O 2 were simultaneously measured with Chl fluorescence. Cell samples in a freshly prepared medium (2 mL, 10 µg Chl mL −1 ) were stirred with a magnetic microstirrer and illuminated with red actinic light (AL) (620 < λ < 695 nm) at 25 • C. A halogen lamp (Xenophot HLX 64625, Osram, München, Germany) with an LS2 light (Hansatech, King's Lynn, UK) was used as the AL source. O 2 was continuously monitored using an O 2 electrode (Hansatech, King's Lynn, UK) [1].
The relative Chl fluorescence originating from Chla was measured using a PAM-Chl fluorometer (PAM-101; Walz, Effeltrich, Germany) [1,44]. Pulse-modulated excitation was achieved using an light emitting diode (LED) lamp with a peak emission at 650 nm. Modulated fluorescence was measured at λ > 710 nm (Schott RG9 long-pass filter). The minimum Chl fluorescence (F o ) was determined based on illumination using a measuring light (ML). Steady-state fluorescence (F s ) was monitored under AL, and 1000-ms pulses of saturated light (λ > 620 nm; 10,000 µmol photons m −2 ·s −1 ) were supplied to determine the maximum variable fluorescence (F m ). The fluorescence terminology used in this study follows that used in a previous report [45]. The effective quantum yield of PSII, Y(II), was defined as (F m -F s )/F m [46]. The relative electron transport rate (ETR) at PSII was calculated as the product of Y(II) and the photon flux density of AL.
For the S. 7942 mutants, the cells were measured similar to the method used for the S. 7002 cells, with the exception that the reaction medium was replaced with 50 mM HEPES-KOH at pH 7.5.
For the experiments of Figure 4, we simultaneously measured O 2 and relative Chl fluorescence by the method previously reported [1,11,13]. Cyanobacterial cells, grown in a fresh medium under high CO 2 conditions, were applied to an O 2 electrode chamber without an additional inorganic carbon source and were then illuminated with red AL. Illumination with AL stimulated photosynthesis, which was accompanied by an increase in O 2 in the reaction medium ( Figure 4A,B). Thereafter, the CO 2 in the medium was gradually removed by cyanobacterial photosynthesis, as the diffusion of CO 2 from the atmosphere into the reaction medium was much slower than its consumption, as a result of photosynthetic CO 2 assimilation in the experimental system. When the CO 2 was depleted, O 2 in the reaction medium began to decrease ( Figure 4A,B), indicating that photosynthesis was suppressed in the transition to CO 2 -limited conditions. The addition of CO 2 , in the form of NaHCO 3 to the reaction medium, restored photosynthetic activity (Figure 4). During measurement, the top of the chamber remained open, which enabled O 2 and CO 2 to diffuse into or out of the reaction medium. This open system prevented an excessive increase in O 2 in the reaction mixture during the longer measurements, so that the level did not surpass the undetectable point of the O 2 electrode. We temporarily closed the chamber to exclude the effects of O 2 diffusion to determine the photosynthetic O 2 evolution rate as indicated by blue shadings (Figure 4A,B). With this experimental system, we could evaluate the response of photosynthesis to CO 2 limitation in vivo [1,11,13].

Measurement of Total Oxidizable P700
Total oxidizable P700 (P m ) of the cyanobacterial cell samples was measured using a Dual-PAM-100 (Heintz Walz, Effeltrich, Germany) at room temperature (25 ± 2 • C). The reaction mixture consisted of fresh A + medium and cells (10 µg Chl mL −1 ). The maximum P700 photo-oxidation level, obtained with saturated pulse light under far-red illumination, was measured and defined as P m according to the standard method [47]. For measurements, a 300-ms saturation pulse (10,000 µmol photons m −2 ·s −1 ) was supplied.
Supplementary Materials: The following are available online at www.mdpi.com/1660-3397/15/12/390/s1, Figure S1: DNA fragments amplified by polymerase chain reaction (PCR) showing complete segregation of the inactivated gene SYNPCC7002_A2492; Figure S2: Comparison of the amino acid sequences of ColA in Synechococcus sp. PCC 7002 and Synechocystis sp. PCC 6803; Figure S3: Immunoblotting to detect heterologous expression of ColA in Synechococcus elongatus PCC 7942; Figure S4: Effects of heterologous expression of ColA on the long-term response of photosynthesis in Synechococcus elongatus PCC 7942 to limited CO 2 .