Lysine Acetylome Analysis Reveals Photosystem II Manganese-stabilizing Protein Acetylation is Involved in Negative Regulation of Oxygen Evolution in Model Cyanobacterium Synechococcus sp. PCC 7002*

Nε-Acetylation of lysine residues represents a frequently occurring post-translational modification widespread in bacteria that plays vital roles in regulating bacterial physiology and metabolism. However, the role of lysine acetylation in cyanobacteria remains unclear, presenting a hurdle to in-depth functional study of this post-translational modification. Here, we report the lysine acetylome of Synechococcus sp. PCC 7002 (hereafter Synechococcus) using peptide prefractionation, immunoaffinity enrichment, and coupling with high-precision liquid chromatography-tandem mass spectrometry analysis. Proteomic analysis of Synechococcus identified 1653 acetylation sites on 802 acetylproteins involved in a broad range of biological processes. Interestingly, the lysine acetylated proteins were enriched for proteins involved in photosynthesis, for example. Functional studies of the photosystem II manganese-stabilizing protein were performed by site-directed mutagenesis and mutants mimicking either constitutively acetylated (K99Q, K190Q, and K219Q) or nonacetylated states (K99R, K190R, and K219R) were constructed. Mutation of the K190 acetylation site resulted in a distinguishable phenotype. Compared with the K190R mutant, the K190Q mutant exhibited a decreased oxygen evolution rate and an enhanced cyclic electron transport rate in vivo. Our findings provide new insight into the molecular mechanisms of lysine acetylation that involved in the negative regulation of oxygen evolution in Synechococcus and creates opportunities for in-depth elucidation of the physiological role of protein acetylation in photosynthesis in cyanobacteria.

metabolism and photosynthesis, in Synechocystis (22). Even now, lysine-acetylated proteome data for cyanobacteria are scarce.
Cyanobacteria constitute a large and morphologically diverse group among prokaryotes and can produce oxygen (25). Based on the endosymbiosis theory, plastids in eukaryotic algae and higher plants were derived from the ancestral cyanobacteria (26). Recently, cyanobacteria have been employed to produce renewable biofuels in an economically effective and environmentally sustainable manner (27). Synechococcus with high glycogen production capacity may become a suitable candidate for the development of biofuels (28 -29). In addition, it is one of the most studied model cyanobacterium with respect to photosynthesis under highintensity light and high salinity as well as other metabolic processes (28, 30 -31). The complete genome sequence of Synechococcus has been determined (see http://www.ncbi. nlm.nih.gov/), and it can be easily genetically transformed with a versatile system (32), making it a model organism to study biotechnological applications and photosynthesis (28). Our previous study has shown that lysine acetylation plays a very important role in carbon metabolism and photosynthesis in Synechocystis (22). Nevertheless, the molecular mechanism of lysine acetylation in the regulation of photosynthesis remains unclear in cyanobacteria.
To gain insights into the biological significance of lysine acetylation in the photosynthetic prokaryote Synechococcus, we analyzed its acetylproteome using immunoaffinity enrichment strategies integrated with high-accuracy LC-MS/MS and characterized the acetylation events in this model system. In this study, we identified 1653 unique lysine acetylation sites from 802 proteins. Interestingly, many acetylated proteins associated with photosynthesis were identified, followed by validation of enhancement of acetylation status in the whole cell lysate under high-intensity light exposure using Western blotting analysis. Functional study revealed that acetylation at K190 of the photosystem II (PSII) manganese-stabilizing protein (PsbO) negatively regulated oxygen evolution. These results provide clues to oxygen evolution process regulated by PsbO acetylation in cyanobacteria.

EXPERIMENTAL PROCEDURES
Cell Culture and Protein Extraction-The Synechococcus strain from Pasteur Culture Collection was cultured in Aϩ medium, gassing with 1% CO 2 (v/v) in filtered air. The strain was grown at 38°C under continuous illumination (33) (normal light: 250 mol photons/m 2 /s; high-intensity light: 2000 mol/m 2 /s). Cells were harvested in the exponential phase (OD 730 Ϸ 0.8) and then exposed to various stresses for 12 h, including A5 (-A5), calcium (-Ca), iron (-Fe), phosphate (-P), and nitrogen (-N) deficiencies. For high-intensity light treatment, cells in the exponential phase were exposed to 2000 mol photons/m 2 /s for 30 min. For high salinity treatment and photoheterotrophic growth conditions (ϩGly), cells in the exponential phase were resuspended in Aϩ medium containing 2.5 M NaCl for 30 min or enriched in Aϩ medium containing 10 mM glycerol, respectively (28). Following this, nicotinamide (10 mM) was added to the cultures for an additional 30 min to inhibit the endogenous protein deacetylase activities. The cells were harvested by centrifugation (6000 ϫ g at 4°C for 5 min), washed with Aϩ medium, and resuspended in lysis buffer containing 20 mM Tris-Cl (pH 7.5), 150 mM NaCl, 10 mM nicotinamide, 1% Triton X-100, and 1ϫ protease inhibitor mixture (Thermo Fisher Scientific, Waltham, MA). The suspension was sonicated (JY92-IIN Ultrasonic Homogenizer; Ningbo Scientz Biotechnology Co., Ltd., Ningbo, Zhejiang, China) at an output of 135 W for 30 min (5 s on followed by 5 s off) on ice-water. Cellular debris was removed by centrifugation (3000 ϫ g at 4°C for 10 min), and the supernatant was stored at Ϫ80°C. The protein concentration was determined by Bradford assay (Bio-Rad, Hercules, CA).
In-solution Trypsin Digestion, High-performance Liquid Chromatography Fractionation, and Enrichment of Acetylated Lysine Peptides-Total proteins (2 mg) were reduced, alkylated, and subjected to trypsin digestion, and the peptides were desalted and fractionated. Protein A/G agarose-conjugated antiacetyllysine antibody (PTM Biolabs, Inc., Chicago, IL) was used to enrich the acetylated peptides, as previously described (22).
LC-MS/MS Analysis-The enriched peptides were analyzed by a Q-Exactive mass spectrometer (Thermo Fisher Scientific) with an easy nLC-1000 system (Thermo Fisher Scientific) and at a linear gradient from 6 to 90% solvent B (98% acetonitrile/0.1% formic acid, v/v) over 40 min. Mass spectrometer was operated in a data-dependent mode with an automatic switch between MS and MS/MS acquisition. Full MS spectra from 350 to 1800 mass-to-charge ratio (m/z) were acquired at a resolution of 70,000 at m/z ϭ 200 in profile mode. Following every survey scan, up to 20 most intense precursor ions were picked for MS/MS fragmentation by higher energy C-trap dissociation with 28% normalized collision energy. Lock mass at m/z 445.12003 was enabled for full MS scan. The dynamic exclusion duration was set to 15 s with a repeat count of 1 exclusion window. The electrospray voltage applied was 2 kV. Automatic gain control was used to prevent overfilling of the ion trap, and 5 ϫ 10 4 ions were accumulated for generation of MS/MS spectra.
Data Analysis-The mass spectra were searched against the Synechococcus database (/cyanobase/GCA_000019485.1/genes.faa/ release date 2008/03/14; 3186 entries in the database) combined with the standard MaxQuant contaminants database using MaxQuant with integrated Andromeda v1.5.3.30 search engine (34). Enzyme specificity was set to full cleavage by trypsin, with two maximum missed cleavage sites permitted. Minimum peptide length was set to 6. Mass tolerances of precursor and fragment ion were set to 6 ppm and 0.05 Da, respectively. Carbamidomethylation of cysteine was set as a fixed modification, whereas oxidation of methionine, deamidation of asparagine and glutamine, and N-terminal acetylation and acetylation of lysine were set as variable modifications. The estimated FDR thresholds for proteins, peptides, and modification sites were specified at maximum 1%. The lysine-acetylated peptides identified were manually inspected using a modified method to exclude isobaric trimethylation (35). Moreover, all acetylation sites assigned to the peptide C-terminal were removed. All the raw data were deposited in a publicly accessible database PeptideAtlas (www.peptideatlas.org) (36).
Bioinformatics Analysis-The identified acetylproteins were grouped into biological process and molecular function classes based on the Gene Ontology (GO) terms using Blast2GO (37)(38)(39)(40). The protein subcellular localization was analyzed by PSORTb v3.0 program (41). GO enrichment was performed by BiNGO v2.44 (42)(43). The aligning peptide sequences with 12 residues surrounding the acetylated lysine residue were extracted, excluding all peptides where the acetylated lysine was located less than seven amino acids from the C-terminal end of the peptide, and these lysine-acetylated motifs were analyzed by motif-x (44). A heat map was generated to further analyze lysine acetylation sites (45)(46)(47). Secondary structures were predicted using NetSurfP (48 -49). Conservation of acetylproteins across species was determined using BLASTP (50). KEGG pathway was analyzed using KEGG mapper (51)(52). The three-dimensional structure of PsbO was modeled using CPHmodels-3.0 (53). The interaction network analysis was performed using STRING (54) and further analyzed using MCODE (55).
Production of Polyclonal Antibodies against Photosystem I Subunit VII (PsaC) and PsbO, Immunoprecipitation, and Western Blotting Analysis-Anti-PsaC and anti-PsbO polyclonal antibodies were generated by PEPTBIO Inc. (Wuhan, Hubei, China), against the synthetic peptides CKAGQIASSPRTED and GAVFLDPKARGTASG, respectively. Immunoprecipitation using anti-PsbO antibody and Western blotting analysis using anti-PsbO, anti-PsbC, and antiacetyllysine antibodies were performed as previously described (22).
Amino Construction of PsbO Point Mutants by Site-directed Mutagenesis-The kanamycin resistance cassette was inserted into SYNPCC7002_A0269 gene to create a knockout mutant (⌬psbO) and into SYNPCC7002_A1479 gene to generate a neutral platform as the alternative wild type (WT*). The trans-complemented psbO mutant and site-directed mutants were assembled with the cpcBA promoter, derived from Synechocystis (PcpcBA), in the front of SYNPCC7002_A0269 gene and the chloramphenicol resistance cassette at the termini. The site-directed mutagenesis was carried out with QuickChangeII mutagenesis kit (Agilent Technologies, Santa Clara, CA), according to manufacturer instructions. Transformation of the Synechococcus strain was performed as previously reported (32). Cells were screened on Aϩ agar plates with chloramphenicol (25 g/ml) or kanamycin (50 g/ml). Transformants were then verified by PCR and direct DNA sequencing. Primers used in this study are listed in supplemental Table S10.
Oxygen Evolution Rate Analysis-Photosynthetic oxygen evolution rates were measured by a Clark-type oxygen electrode (Hansatech Instruments Ltd., Norfolk, UK). Cells of all strains were collected in the exponential phase and resuspended in fresh Aϩ media, then diluted to a final OD 730 ϭ 1 for oxygen evolution measurements. For temperature treatments, cells were incubated at 28, 33, 38, and 43°C for 10 min. Cell suspensions were added to the electrode chamber with continuous stirring for each experiment. The various electrode chamber temperatures were maintained by a circulating water bath. Oxygen evolution rate was measured after 2 min in the dark with stimulating irradiance of 500 mol photons/m 2 /s and 10 mM NaHCO 3 .
Chlorophyll Fluorescence Analysis-Chlorophyll fluorescence measurements were made using a pulse-amplitude fluorimeter DUAL-PAM-100 (Heinz Walz GmbH, Effeltrich, Germany). All the mutants were collected in their exponential phase and diluted to OD 730 ϭ 1 for chlorophyll fluorescence analysis. Before measurements, cells were dark-adapted for 15 min at room temperature. The intensity of the actinic light was maintained at 50 mol photons/m 2 /s. The light intensity dependence of electronic conversion efficiency of photosystem II was detected with the stepwise increase in photosynthetically active radiation (0 -1000 mol photons/m 2 /s) through the fluorescence measure mode of a dual channel. For recording the transient post-illumination increase, the saturating pulse and actinic light last-ing for 2 min were terminated and the slow kinetics had been allowed to run for around 10 min.
Experimental Design and Statistical Rationale-The wild-type strain of Synechococcus was prepared from the same lysates as those used for the study of acetylation proteome by Yang et al. (47). Specifically, acetylated peptides from whole cell lysates were fractionated and enriched using immunoaffinity enrichment strategies. Eight samples were analyzed by high-accuracy nanoflow LC-MS/MS technology (Fig. 1B). FDR thresholds for proteins, peptides, and modification sites were set to maximum 1%. Data source name of the proteome of Synechococcus sp. PCC 7002 was GCA_000019485.1 (release date 2008/03/14) with 3186 entries in the database searched. GO analysis was performed using Blast2GO (41) and GO enrichment (42) using Hypergeometric test. Benjamini-Hochberg FDR correction was carried out at a significance level of 0.05. Lysine-acetylated motifs were analyzed by motif-x (44) at a significance level of 0.000001. Secondary structures were predicted using NetSurfP (48), with p values calculated by Wilcoxon test.
In the functional study, three replicates were prepared and analyzed for each of the ten biological samples: wild type (WT), WT* (SYNPCC7002_A1479 as a neutral platform), ⌬psbO, trans-complemented psbO mutant, and the site-directed mutants under the control of PcpcBA, including samples with a nonacetylated arginine to construct the mutants K99R, K190R, and K219R or a mimic-acetylated glutamine to construct K99Q, K190Q, and K219Q. The following analyses were performed on the samples: immunoprecipitation and Western blotting analysis of PsbO acetylation level; construction of ⌬psbO, WT*, trans-complemented psbO mutant, and site-directed mutants by site-directed mutagenesis; followed by growth rate, oxygen evolution rate, and chlorophyll fluorescence analyses. Mean and standard deviation of the three independent experiments are presented and the statistical tests used to analyze the data are indicated in the respective figures and/or manuscript sections.

Establishment of the Synechococcus Acetylome-To characterize the extent of lysine acetylation in Synechococcus,
Western blotting analysis of whole cell lysates using antiacetyllysine antibody was conducted. It showed multiple lysine-acetylated protein bands under different culture conditions ( Fig. 1A), suggesting the wide distribution of diverse acetylated proteins and a regulated process of lysine acetylation in this organism. To further comprehensively identify the large-scale data set of lysine-acetylated sites on Synechococcus proteins, the total proteins from different growth conditions before the total protein extracts were isolated, and the acetylated peptides were enriched using antiacetyllysine antibody. Then, the enriched modified peptides and sites were analyzed by a nanoflow LC-MS/MS (Fig. 1B). The suspicious acetylated peptides with unclear MS/MS spectra were manually removed (supplemental Table S1). In this study, we detected 802 proteins acetylated at 1653 unique sites. It is noteworthy that the identified acetylated proteins comprising 25.2% (802/3186) of Synechococcus proteome exceed the maximum previous findings of acetylome in bacteria (18, 20 -22). Synechococcus proteomic data also showed that 28.8% acetylated peptides and 58.5% acetylated proteins were enriched, compared with the identified total unique peptides and proteins, respectively (Fig. 1C). In addition, we es-timated the mass accuracy of the modified peptide and the average absolute mass error was 0.8614 ppm (Fig. 1D). We also found that the average acetylation degree was 1.97. Of all these acetylated proteins, ϳ55.1% had a single acetylated site, 21.1% carried two acetylated sites, 10% contained three sites, and 6.6% carried four sites. Notably, 58 proteins revealed five or more acetylated sites. For instance, phycobilisome core-membrane linker phycobiliprotein with 17 acetylation sites was the most intensively acetylated protein in this research. This data set provides a comprehensive view of lysine acetylation in Synechococcus.
Functional Classification of Acetylproteins-To better understand the role of identified acetylproteins in Synechococcus, we classified these proteins according to their subcellular localization, biological process, and molecular function (supplemental Fig. S1 and supplemental Table S2). Among the 802 acetylated proteins, 569 (70.9%) were functionally annotated to various biological processes. Many acetylated proteins were involved in cellular amino acid metabolism, translation, generation of precursor metabolites and energy, carbohydrate metabolism, photosynthesis, and other processes. With respect to the annotated molecular function, the acetylated proteins were assigned to several subcategories as ion binding, 252; oxidoreductase activity, 102; and lyase activity, 53. Based on subcellular localization, 491 (61.2%) acetylproteins were in the cytoplasm, 100 were assigned to cytoplasmic membrane, 10 to periplasm, 6 to outer membrane, and 3 to extracellular space. However, the localization of 22.9% acetylproteins was not predicted. We also performed a statistical test by applying a multiple testing correction, and the results have been presented in supplemental Table S3. We conducted GO enrichment analyses to examine which functional categories were targeted for lysine acetylation (supplemental Fig. S2 and supplemental Table S4). GO enrichment analysis of biological process category demonstrated that acetylproteins involved in cellular metabolism (p value ϭ 1.12E-02), metabolism (p value ϭ 1.12E-02), cellular process (p value ϭ 1.12E-02), and primary metabolism (p value ϭ 3.52E-02) were significantly enriched. These findings indicated the essential regulatory roles of lysine acetylated proteins in cellular metabolism in Synechococcus.
Analysis of Lysine Acetylation Sites-To evaluate the nature of acetylated lysine in Synechococcus, we compared the occurrences of 12 amino acid residues surrounding the acetylation site with those of all lysine that occur in the Synechococcus proteome. Previous studies have found preferences for amino acid residues around the acetylated lysine in prokaryotes (11, 18 -22, 24). Herein, we defined 11 enriched  Table S5), and they exhibit distinct abundances (Fig. 2B). Consistent with previous findings (18, 20 -22, 24), for three of these motifs, a positively charged lysine or arginine enrichment was observed on the ϩ4 or ϩ5 FIG. 2. Analysis of lysine acetylation sites. A, Probability logos of significantly enriched acetylation site motifs after aligning peptide sequences, including 12 residues surrounding the acetylated lysine residue, using motif-x. B, The number of identified peptides containing acetylated lysine in each motif. C, The relative abundance of amino acid residues flanking the acetylation sites represented by an intensity map. The intensity map shows the relative abundance for Ϯ6 amino acids from the Synechococcus lysine-acetylated site. The colors in the intensity map represent the log 10 of the ratio of frequencies within acetyl-13-mers versus nonacetyl-13-mers (red shows enrichment, blue shows depletion). D, Distribution of secondary structures containing lysine acetylation sites. The different secondary structures (␣-helix, ␤-strand and -coil) of acetylated lysine identified in this study were compared with the secondary structures of all lysine identified through proteomics.
position toward the C terminus. In addition, histidine (H), phenylalanine, or tyrosine (Y) on the ϩ1 position toward the C terminus appeared frequently, as previously described in bacteria (11,18,21,(23)(24). The motifs KacH, KacY, and KacF were also conserved in strawberry (56), implying some conserved motifs surrounding acetylated sites are shared by plants and bacteria. Interestingly, two negatively charged residues, including glutamate (E) on Ϫ1 position toward N terminus and aspartate (D) on Ϫ1/ϩ1 or Ϫ4/ϩ4 position toward N or C terminus, were observed. These results suggested that negatively or positively charged or aromatic amino acid residues might be functionally important for acetylation. To our knowledge, five of these motifs, including E on Ϫ1 position and D on Ϫ1/ϩ1 or Ϫ4/ϩ4 position, were the first significant motifs identified in bacteria. Considering that acetylated peptides with these three motifs account for 28.5% of all identified peptides, we speculate that proteins with such motifs might have some special functions in this model system. These results were also determined by an inspection of the intensity map analysis (Fig. 2C). We observed that the residues H and Y are preferred at ϩ1 position of acetylpeptides. The different enrichment of amino acid residues surrounding acetylation sites implies unique substrate preferences in Synechococcus and reflects the conservation of several target sequences of bacterial lysine acetyltransferase.
To investigate the local secondary structures of acetylproteins surrounding acetylation sites, we performed a structural analysis of the lysine-acetylated proteins. As demonstrated, ϳ49.5% of the modified sites were in regions of ␣-helices, 192 sites were in ␤-sheets, and 643 were in the unstructured regions (Fig. 2D). A similar distribution pattern of secondary structures was found in acetylated lysine in the Synechocystis proteome (22), which differs from most other bacterial acetylomes (18,20,23). Distinct from other bacteria, we deduced that acetylation prefers the ␣-helix structures in cyanobacteria.
Comparative Analysis with Other Bacterial Acetylome-Lysine acetylation is an evolutionarily conserved PTM in both prokaryotes and eukaryotes (1). To determine the conservation of the identified acetylated proteins, we performed a systematic analysis of the orthologs of acetylproteins by comparing the Synechococcus acetylproteome to several bacterial acetylproteomes (11-12, 15-17, 57 Table S6). Generally, the Synechococcus acetylome exhibited greater similarity to Gram-negative bacteria (486 proteins) than to Gram-positive bacteria (256 proteins). The comparative analysis of acetyl-proteomes revealed that many of the acetylated proteins might serve a conserved role in cellular metabolism and translation, which is consistent with these microorganisms with characterized acetylproteomes. These results point to an evolutionarily conserved and potentially vital role of acetylation in cellular function.
Protein-protein Interaction Network of the Identified Acetylproteins-To further understand the cellular processes regulated by acetylation, the STRING database was used to analyze the interaction network of these identified acetylproteins. The network showed that 748 acetylated proteins were network nodes, with 7474 physical and functional interactions (supplemental Table S7). Twenty-eight highly interconnected clusters of acetylated proteins were retrieved (supplemental Table S8). The top two clusters consisted of numerous acetylproteins involved in protein synthesis and photosynthesis (supplemental Fig. S3), with 1138 and 358 direct physical or indirect functional interactions, respectively, indicating the important regulatory roles of acetylation modification in the two processes. Interestingly, they have also been assigned to ribosome, aminoacyl-tRNA biosynthesis, and photosynthesis pathways via KEGG pathway analysis (supplemental Table  S9).
Lysine acetylation has been shown to play a major role in metabolism regulation in prokaryotes (18 -22). In the Synechococcus acetylproteomes, many of the identified acetylated proteins were involved in metabolic processes consistent with previous findings. As shown in supplemental Table  S9, 232 acetylated proteins were found to be involved in various metabolic pathways. Nearly all enzymes involved with each step of central metabolism were identified to be acetylated, indicating the role of reversible lysine acetylation in the regulation and control of cellular enzymatic activity at PTM level in this cyanobacterium. Moreover, 43 acetylated proteins were involved in carbon metabolism, 9 were photosynthesisantenna proteins, 27 involved in photosynthesis, 14 involved with carbon fixation photosynthetic organisms, 17 in glycolysis/gluconeogenesis, and 10 in citrate cycle (Fig. 3). Consistent with the Synechococcus proteogenomic study (58), our data also revealed the potential role of protein acetylation in photosynthesis (supplemental Fig. S3).
Amino Acid Sequences Alignment of PsbO-The manganese (Mn)-stabilizing protein PsbO in the PSII complex in all oxygenic photosynthetic organisms plays central roles in maintaining the integrity and activity of the Mn cluster and is essential for optimal oxygen evolution (59). To validate the acetylation events in PsbO, we carried out Western blotting analysis of the whole cell lysate under different conditions using pan antiacetylation antibody and anti-PsbO antibody. Our findings confirmed PsbO acetylation through immunoprecipitation; nevertheless, PsbO acetylation levels remained unchanged under different light intensities (Fig. 4A). To better understand the potential role of acetylation in PsbO conformation, the three-dimensional structure of PsbO was mod-eled, and the location of the three identified acetylation sites have been marked accordingly in Fig. 4B. The conservation of amino acid sequences of PsbO among species from cyanobacteria to higher plants was analyzed. Fig. 4C shows multiple alignments of partial sequences of PsbO from five cyanobacteria and six higher plants. This comparison indicates that cyanobacteria have one distinct region differing from higher plants, presented as the cyano-loop labeled in a rectangle. Although multiple alignments showed that PsbO was poorly conserved, researchers have identified five highly conserved regions within this protein among oxygenic organisms (60). Two acetylated sites of PsbO, including K99 and K190, were in KL and DPKGR conserved regions, respectively. Considering the evolutionary conservation of acetylation in both prokaryotes and eukaryotes, we deduced that the modification events might occur in other oxygenic organisms and might have a significant impact on the structure or function of PsbO.

Site-directed Mutagenesis of PsbO Acetylation Sites-To
investigate whether the acetylated sites were involved in oxygen evolution within PSII, site-directed mutagenesis of PsbO at K99 in KL region, K190 in DPKGR region, and K219 was performed. Before construction of the point mutants, a neutral site platform and a psbO null mutant were constructed as positive and negative controls, respectively. Synechococcus was adapted to grow on glycerol as an organic carbon source to favor mixotrophic growth, allowing the construction of the mutants lacking PSI and/or PSII, as previously reported (61) (Fig. 5). A neutral site platform with inactive SYNPCC7002_A1479 was obtained as the alternative wild type WT* ( Fig. 5A; supplemental Fig. S4). However, the inactivation of psbO in vivo behaved distinctly from WT*. The schematic of the vector for these site-directed transformations is presented in Fig. 5A. In this study, we converted lysine to either a nonacetylated arginine to construct the mutants K99R, K190R, and K219R or to a mimic-acetylated glutamine to construct the mimic mutants K99Q, K190Q, and K219Q. All mutants were verified by PCR and direct DNA sequencing ( Fig. 5B; supplemental Fig. S5).
To validate PsbO expression level in both WT* and mutants, we analyzed PsbO abundance among these strains by Western blotting analysis (Fig. 5C). The PsbO mutant forms were expressed in each case, and the translational levels of PsbO in K99R/K99Q, K190R/K190Q, and K219R/K219Q mutants were generally the same as that in WT*. Additionally, the intrinsic core of PSI among these mutants and wild type were further analyzed using the anti-PsaC antibody. PsaC was expressed in equal amount among the wild type and mutant strains, indicating that PSI was not affected by the inactivation or site-directed mutagenesis of psbO.
K190 in the conserved DPKGR region of PsbO is present in the head domain I, as shown in Figs. 4B and 4C, and is involved in many interactions with PSII members (60). Furthermore, our results supported the speculation that lysine acetylation at K190 in PsbO might have an anciently conserved role in the maintenance of PSII system and the thylakoid membrane, since its divergent evolution from G. violaceus (62). PsbO mutations have been previously identified to be associated with a decrease in the accumulation of PSII components in C. reinhardtii (63) and A. thaliana (64 -65), suggesting that deletions or amino acid substitutions might affect the intrinsic degradation of the PSII complex.
Characterization of Growth Rate of Mutant Strains-PsbO plays an essential role in the variability of both prokaryotic and eukaryotic systems (63, 66 -67). Knockout and/or knockdown mutations in psbO in Synechocystis (66,68), C. reinhardtii (63), and A. thaliana (64 -65, 67, 69) have been constructed for variability investigation. Mutations involving removal of PsbO from PSII were found to have different effects in prokaryotes FIG. 5. Construction and detection of the ⌬psbO site-directed mutants. A, Schematic shows the construction of ⌬psbO site-directed mutant strains by homologous recombination. The kanamycin resistance (K m r ) cassette was inserted into the SmaI restriction sites in psbO gene and the chloramphenicol resistance (Cm r ) cassette was inserted into the modified SmaI restriction sites in SYNPCC7002_A1479 gene, added to the appropriate ends of the flanking sequences (see Experimental Procedures). B, C, Agarose gel electrophoresis demonstrates complete segregation of alleles of P cpcBA -A0269::Cm r using A0269kn-F/R and A1479kn-F/R primers. D, Western blotting analyses of total cellular extracts from Synechococcus wild type and ⌬psbO site-directed mutants were performed using antibodies specific to PsbO and PsaC. PsbO, photosystem II manganese-stabilizing protein; ⌬psbO, PsbO knockout mutant; PsaC, photosystem I subunit VII; P cpcBA , cpcBA promoter derived from Synechocystis. and eukaryotes. The Synechocystis psbO mutant is capable of slow growth, but is more vulnerable to photoinhibition (70). However, the absence of PsbO in C. reinhardtii causes a photoautotrophic growth failure and the loss of the ability to assemble functional PSII reaction centers (63). The psbo1 A. thaliana mutant exhibits retardation of photoautotrophic growth as well (67). To confirm that the growth of Synechococcus depends on the involvement of acetylated sites located on PsbO, we tested the ability of psbO point mutants to grow photoautotrophically. The different strains exhibited variation in growth in our study (supplemental Fig. S6). We failed to construct a fully segregated null mutation of psbO under photoautotrophic growth conditions, indicating the essential role of PsbO in photoautotrophic growth. Contrarily, the ⌬psbO mutant could grow under normal irradiation (ϳ250 mol⅐m Ϫ2 ⅐s Ϫ1 ) when cells were supplied with 10 mM glycerol. The present experiments clearly demonstrated a decreased photoheterotrophic growth rate of all the site-directed mutants constructed under normal light compared with WT*, but an increased growth rate compared with the ⌬psbO mutant, which might be ascribed to the partial complementation of PsbO. Among these point mutants, strains mimicking constitutively acetylated states (K99Q, K190Q, and K219Q) grew under normal light irradiation, like the mutants with nonacetylated states (K99R, K190R, and K219R). Our observations indicate a minor role of PsbO acetylation in the variability of Synechococcus.
Detection of Oxygen Evolution Rate in Whole Cell under Various Temperature Conditions-PsbO stabilizes the Mn cluster that catalyzes the oxidation of water to molecular oxygen in photosynthetic organisms (59) and serves to optimize oxygen production at the physiological calcium and chloride concentrations (71). To characterize the role of PsbO acetylation in oxygen evolution in Synechococcus, the oxygen evolution rate in both WT* and mutants was examined at different temperatures. The ⌬psbO mutant has altered kinetic properties in the S-states of oxygen-evolving centers in Synechocystis (66,(72)(73). Our data showed that WT*, K99R/ K99Q, K190R/K190Q, and K219R/K219Q cells exhibit increased oxygen evolution with elevation in temperature and reach the maximal rate at 38°C, after which the rate of oxygen generation remarkably reduced at 43°C (Fig. 6). However, the mutants ⌬psbO and K190Q, with slow growth rates, exhibited dramatically decreased oxygen evolution with increase in temperature from 28 to 43°C. On removal of PsbO, the initial oxygen evolution rate per cell was about 53.7, 18.2, 7.5, and Ϫ9.8% of the wild-type rate at 28, 33, 38, and 43°C, respectively, whereas on substitution of K190 with uncharged residue Q, it was 55.2, 42.9, 31, and Ϫ9% at 28, 33, 38, and 43°C, respectively. Nevertheless, K190R and other mutations had a marginal effect on this process. It has been verified that K160R mutant protein (Synechococcus K190R in PsbO) restores oxygen evolution more effectively to some extent than the K160Q mutant protein in vitro in Synechocystis, indicating participation of this residue in electrostatic interaction (74). Our findings suggested that PsbO acetylation at K190 negatively regulates oxygen evolution in Synechococcus at high temperatures.
Detection of Electron Transport Rate through PSII (ETR(II)) and Transient Rise in Fluorescence-To investigate whether linear electron flow (LEF) was affected by PsbO acetylation, room temperature pulse amplitude modulated fluorescence was studied to determine the ETR(II) among the Synechococcus psbO site-directed mutants. The light response curves of ETR(II) were the same for these strains, except for the psbO null mutant, for which they not detectable, as shown in Fig.  7A. On the other hand, the K99R/K99Q, K190R/K190Q, and K219R/K219Q mutants showed a decrease in ETR(II) when compared with WT*. Nevertheless, no significant differences in ETR(II) were observed among these mutants.
The increase in fluorescence was undetectable in WT, trans-complemented psbO, and K190R strains, whereas a trace activity was detected in K99R/K99Q (Fig. 7B). In contrast, ⌬psbO and K190Q mutants were identified by significant enhancement of postillumination increase in chlorophyll fluorescence. These results reflected the less efficient electron transfer in LEF, in contrast to the elevated cyclic electron flow (CEF) in the mimic-acetylated mutant. This enhancement of transient rise in fluorescence was because of PsbO dysfunction. Considering that post-illumination increases in chlorophyll fluorescence are thought to be involved in PSI-cyclic electron transport (75), we speculate that PsbO acetylation mediated an increase in ATP generation to compensate for the decrease in NADPH production.

DISCUSSION
Protein lysine acetylation has been extensively studied in both eukaryotes and prokaryotes. Lysine acetylation is a highly dynamic and reversible PTM. This process can be carried out by acetyltransferases and deacetylases within the cells (1). In this study, 1653 acetylated sites and several new motifs were identified on acetylated proteins in Synechococcus. Several potential acetyltransferases (SYNPCC7002_A0110, SYNPCC7002_ F0063, SYNPCC7002_A2436, SYNPCC7002_A0096, and SYNPCC7002_A2015) and a deacetylase (SYNPCC7002_ A2791) have also been identified among the acetylated proteins. Additionally, the motif analysis revealed several new motifs that might be recognized by protein acetyltransferases. However, no lysine acetyltransferases or deacetylases were predicted or confirmed in this model cyanobacterium. Recently, it was demonstrated that lysine acetylation can also be chemically catalyzed (76 -77). Therefore, it is uncertain whether this PTM occurs enzymatically, nonenzymatically, or both ways in Synechococcus.
The proteomic profiling of lysine acetylation revealed that the acetylated proteins were involved in the regulation of many metabolic pathways (i.e. tricarboxylic acid cycle, glycolysis, and photosynthesis). Moreover, we also identified acetylated proteins involved in the two-component system and microbial metabolism in diverse environments, suggesting the evolution of adaptive responses to biotic and abiotic stresses. Thus, we speculate that acetylation is a reserve of carbon sources for cells resistant to unfavorable environment. Interestingly, our findings have indicated that protein lysine acetylation plays vital roles in photosynthesis in this model cyanobacterium, in accordance with the previous findings in Synechocystis (22) and Arabidopsis (6).
Photosynthesis serves as a system for sensing environmental stresses, besides its principal function as a solar energy converter (78 -79). It has been reported that various PTMs might be regulating photosynthetic efficiency in triggering acclimation to environmental cues (80). Lately, characterization of PTMs in photosynthesis has provided a large quantity of information for the study of this process. For example, several reversible modifications, including serine/threonine/ tyrosine phosphorylation (47,81), N terminus and internal lysine acetylation (6), lysine and arginine methylation (82), and S-thiolation (83), have been well established as key regulators of Calvin cycle, ATP synthesis, electron transport, and oxygen evolution. Recently, Synechococcus proteogenomic analysis has revealed 23 different PTMs in photosynthesis, including lysine acetylation (58). Our proteomic study also suggested potential roles of lysine acetylation in photosynthesis in Synechococcus.
The precise structure of oxygen generating PSII, with a high resolution of ca. 1.9 Å, has revealed the geometric arrangement of the Mn 4 CaO 5 cluster, which may facilitate the understanding of photolysis of water within PSII (84). However, the existence and function of PTMs, such as acetylation, in oxygen-evolving complex (OEC) have not yet been clarified with this structure. In our research, several proteins in OEC were found to be lysine acetylated, including PsbO. PsbO is responsible for maintaining optimal oxygen evolution from PSII during photosynthesis. It is poorly conserved, but has five highly conserved regions, including K99 in KL region and K190 in DPKGR region, discovered in all oxygenic photosynthetic organisms. K190 is substituted by arginine in G. violaceus, indicating the origin of this region and its role in PsbO stability. Three acetylated sites of PsbO (K99, K190, and K219) in Synechococcus were confirmed as targets for acetylation. It has been reported that PsbO in Synechocystis is acetylated at K82 (K83 in Synechococcus) and K209 (L213 in Synechococcus) (22). Nevertheless, no in-depth functional studies have been performed on PsbO acetylation in cyanobacteria.
In this study, we first focused on the role of PsbO acetylation in regulation of cellular growth under optimal growth conditions. The consequences of removal or site-directed mutagenesis of PsbO in vitro or in vivo have been analyzed in detail, to reveal its conserved role in photosynthetic organisms. Deletion or site-directed mutation of PsbO in PSII has been found to impact the photoautotrophic growth (63)(64)(65)(66).
The Synechocystis psbO mutant is capable of slow growth, but evolves oxygen at low rate, and is very sensitive to highintensity lights under photoautotrophic conditions (66). Furthermore, the absence of PsbO in C. reinhardtii leads to the failure of photoautotrophic growth and functional PSII reaction center assembly (63). On the other hand, the psbo1 A. thaliana mutant exhibits photoautotrophic growth retardation (67). Removal of PsbO, in this research, showed similar physiological phenotypes. The growth rate of ⌬psbO mutant was partially restored by trans-complemented and site-directed mutations. Characterization of such strains showed that these mutations hindered photoheterotrophical growth when placed in a genetic background lacking PsbO. It has been studied that lack of PsbO in PSII preparations results in withdrawal of two out of the four Mn atoms comprising the water-oxidizing complex (85)(86). Apart from this, PsbO has been hypothesized to be coupled with PSII turnover (87)(88)(89). In our study, PsbO acetylation was proposed to play a regulatory role in PSII and cell growth.
Next, we mainly focused on the role of PsbO acetylation in oxygen evolution under different temperature conditions. Thus, the oxygen evolution behaviors of ⌬psbO mutant and point mutant strains were further analyzed. Wild-type cells could evolve oxygen at 28 to 43°C normally, whereas K190Q mutation in PsbO of Synechococcus suppressed oxygen evolution significantly, like ⌬psbO mutant, and the mutant could not produce oxygen as efficiently as WT* and other cells. It has been confirmed previously that oxygen evolution rate is inhibited by high temperature treatment in tobacco (90). Additionally, oxygen production was dramatically reduced on substitution of K160 in Synechocystis (K190 in Synechococcus) with uncharged residues in vitro. Contrarily, the K160R mutant protein could reconstitute oxygen evolution more effectively, implying the participation of this acetylated site in electrostatic interaction (74). In vitro reconstitution assays have shown that spinach PsbO is unique in its thermostability (91). Conversely, our data revealed that Synechococcus PsbO exhibits less thermotolerance in in vivo assays. Taken together, our observations confirmed and extended the earlier results that ⌬psbO mutant shows a decrease in oxygen evolution under high temperature conditions. Five oxidation states of the OEC, known as S 1 , S 2 , S 3 , S 4 , and S 0 , have been previously characterized, with the S 4 state being responsible for water oxidation and oxygen evolution (92). Many researchers have conducted experiments to investigate the molecular mechanism of regulation of S-state transitions occurring at OEC (65,(93)(94)(95). The psbo1 A. thaliana mutant exhibits longer S 2 and S 3 states (65). Thermoluminescence experiments reveal that the S 2 state has a higher stability in the psbo1 mutant than in wild type, and electron paramagnetic resonance measurements suggest a lower yield of the S 2 multiline signal in A. thaliana (94). Flash oxygen yield measurements reveal that the S 2 and S 3 states are more stable in the ⌬psbO mutant of Synechocystis and that there is at least 5-fold reduction in S 3 3[S 4 ]3 S 0 transition in this mutant (96). Similar results have also been obtained for spinach PSII lacking PsbO, in vitro (97). Thus, we hypothesize that acetylation of PsbO at K190 may be one of the mechanisms that regulate S state transitions in Synechococcus.
The slow generation of oxygen in the absence of PsbO was verified to be a consequence of a decrease in the rate of electron transfer. Electron paramagnetic resonance measurement monitored the retardation in the rate of photooxidized Y Z reduction in the ⌬psbO mutant from Synechocystis, indicating slower electron transfer from the Mn 4 CaO 5 cluster to photooxidized Y Z (98). On the other hand, the psbo1 A. thaliana mutant exhibits reduction in electron transfer between quinones Q A and Q B (65,69). Consistent with the earlier observations, ETR(II) detected a decrease in LEF rate in the PSII of ⌬psbO and point mutants in our study. Additionally, CEF was measured, and the absence of PsbO and the substitution of K190 by glutamine resulted in an enhanced transient rise in fluorescence. These results suggested a positive correlation between acetylation and CEF.
To reveal the role of lysine acetylation in the phenotypic changes in photosynthesis, site-directed mutagenesis of psbO was performed to mimic the electrostatic charge of a nonacetylated lysine (K3 R) or an acetylated lysine (K3 Q) (supplemental Table S11). It is not surprising that the nonacetylated mutant has a similar phenotype to the native strain with respect to the very low acetyl stoichiometry (2, 99 -100). Compared with the mutants mimicking nonacetylated states, the in vivo characterization of the K190Q mutant mimicking acetylation exhibited a decrease in oxygen evolution and an enhanced cyclic electron transport rate under normal growth conditions. Although photosynthetic performance of the regulation of acetylation at K190 in PsbO was characterized, the mechanisms that altered the ability of PsbO to reactivate oxygen evolution were not well established. To further investigate the regulatory role of acetylation in oxygen evolution, we might need to apply several strategies to study acetylation dynamics and accurate acetylation stoichiometry of PsbO, to understand the mechanisms.

DATA AVAILABILITY
The raw data and annotated peptide spectra of acetylated peptides have been deposited in the PeptideAtlas database and can be accessed with the identifier PASS00772 (http: //www.peptideatlas.org/PASS/PASS00772).