Light-Induced Oxidative Stress, N-Formylkynurenine, and Oxygenic Photosynthesis

Light stress in plants results in damage to the water oxidizing reaction center, photosystem II (PSII). Redox signaling, through oxidative modification of amino acid side chains, has been proposed to participate in this process, but the oxidative signals have not yet been identified. Previously, we described an oxidative modification, N-formylkynurenine (NFK), of W365 in the CP43 subunit. The yield of this modification increases under light stress conditions, in parallel with the decrease in oxygen evolving activity. In this work, we show that this modification, NFK365-CP43, is present in thylakoid membranes and may be formed by reactive oxygen species produced at the Mn4CaO5 cluster in the oxygen-evolving complex. NFK accumulation correlates with the extent of photoinhibition in PSII and thylakoid membranes. A modest increase in ionic strength inhibits NFK365-CP43 formation, and leads to accumulation of a new, light-induced NFK modification (NFK317) in the D1 polypeptide. Western analysis shows that D1 degradation and oligomerization occur under both sets of conditions. The NFK modifications in CP43 and D1 are found 17 and 14 Angstrom from the Mn4CaO5 cluster, respectively. Based on these results, we propose that NFK is an oxidative modification that signals for damage and repair in PSII. The data suggest a two pathway model for light stress responses. These pathways involve differential, specific, oxidative modification of the CP43 or D1 polypeptides.


Introduction
In plants, algae and cyanobacteria, Photosystem II (PSII) catalyzes the photo-oxidation of water to O 2 and protons [1]. The electrons derived from water are transferred sequentially to two quinone molecules, Q A and Q B , on the acceptor side of the reaction center [2]. The cyanobacterial PSII structure was solved to 1.9 Å resolution [2][3][4][5][6][7]. The membrane-spanning D1 and D2 proteins form the core of the reaction center. These proteins bind the catalytic oxygen evolving complex (OEC), which is a Mn 4 CaO 5 cluster, chlorophyll (chl), pheophytin, and the plastoquinones, Q A and Q B , [2]. The CP43 and CP47 proteins are also found in the core of PSII (reviewed in [8]). CP43 and CP47 span the membrane in the PSII complex, and these subunits contain flexible, hydrophilic loops that protrude into the lumen. Substitutions of amino acids in these loops have demonstrated their importance for complex assembly and protection from photoinhibition [8]. Calcium and chloride cofactors are essential for optimal activity under native conditions [9].
Light stress causes protein damage and suboptimal photosynthetic rates in PSII [10,11]. A decrease in steady state oxygen evolution, as well as accelerated D1 turnover, is the result. Recovery from photoinhibition involves PSII disassembly, proteolysis of damaged D1, and de novo synthesis of a new D1 protein.
Re-insertion of a new D1 subunit into the partially disassembled PSII complex and reassembly completes the repair cycle [12]. The signaling pathways for complex disassembly and D1 degradation remain unknown. However, post-translational oxidations of amino acids have been proposed to play signaling roles in this process [12].
Recently, we described a light-induced modification, NFK ( Figure 1A), resulting from the oxidative, post-translational modification (PTM) of W365 in the CP43 subunit [19]. A ,two fold increase in the yield of NFK365-CP43 was observed following high light illumination [19]. A concomitant two fold decrease in oxygen evolution was detected under the same conditions [19]. This result suggests a role for NFK and oxidative stress in plant photoinhibition.
NFK results from the reaction of the Trp side chain with several types of ROS, including singlet oxygen ( 1 O 2 ) [15,20], ozone (O 3 ) [21], and hydroxyl radicals (HO N ) [22,23]. NFK can also result from a metal-catalyzed radical mechanism, followed by reaction with O 2 [13]. ROS is produced in PSII, either by recombination reactions producing triplet chlorophyll ( 3 chl) or by reactions at the Mn 4 CaO 5 cluster ( Figure 1B). These are referred to as acceptor side ( 3 chl) and donor side (Mn 4 CaO 5 ) reactions.
In the acceptor side ROS mechanism, double reduction of Q A results in charge recombination and formation of the excited state 3 chl [24]. Energy transfer from 3 chl to ground state 3 O 2 results in 1 O 2 [24]. O 2 reduction to O 2 N2 may also occur under light stress conditions [25]. Dismutation to H 2 O 2 , followed by a single electron reduction, may produce HO N [25]. In the donor side ROS mechanism ( Figure 1B), release of the extrinsic proteins and OEC during light stress has been reported to stimulate H 2 O 2 production [26]. The one electron oxidation and reduction of H 2 O 2 was proposed to produce O 2 N2 and HO N , respectively [25]. In this work, we identify a specific, new oxidative modification of tryptophan in the D1 subunit, which is induced by light-stress. We provide evidence that N-formylkynurenine modifications in PSII are generated by ROS, which may be derived from the (B) shows models of PSII (top) and TW (bottom) PSII. Tris-washing removes the extrinsic subunits and OEC, or Mn 4 CaO 5 cluster. The core subunits (CP43, D1, D2 and CP47) and electron transfer cofactors (tyrosine z (Y z ), P 680 , pheophytin (Pheo), plastoquinone A (Q A ), and plastoquinone B (Q B )) are labeled. The water-splitting reaction at the OEC is shown. The ROS species generated in PSII and TW PSII during photoinhibition are indicated. The subunit colors are the same as in Figure 8. CP43 (pink); D1 (green); D2 (light blue); CP47 (dark blue); extrinsic subunits (violet, yellow, and orange). doi:10.1371/journal.pone.0042220.g001 Mn 4 CaO 5 cluster. To explain our results, we propose a twopathway model, in which NFK functions as a signal for D1 protein turnover, a key step in repair under high light stress.

Photoinhibition in PSII and Thylakoid Membranes (TMs)
A light intensity of 7,000 mmol photons m 22 s 21 was employed in these studies. This value is typical of conditions used in previous studies of plant light stress (4,000-7,000 mmol photons m 22 s 21 ) [27][28][29]. To evaluate the degree of photoinhibition under these conditions, the steady state rate of oxygen evolution was monitored as a function of illumination time. High light illumination of PSII membranes was conducted at a chlorophyll concentration of 1 mg/mL, pH 6.0, and 25u. Compared to the dark control (Figure 2A, blue), illumination induced a 3.460.4 fold decrease in oxygen evolution rate in PSII membranes (Figure 2A, blue and green). This agrees with our previous report of a 2.460.5 fold decrease under these conditions [19]. As expected, a lower light intensity of 500 mmol photons m 22 s 21 did not significantly decrease oxygen evolution rates (Figure 2A, blue and red).
At the same light intensity, thylakoid membrane (TM) samples were not significantly inhibited at a chlorophyll concentration of 1 mg/mL ( Figure 2B, black and blue). However, illumination at 0.1 mg/mL chlorophyll induced a 5.460.5 decrease in the steady state oxygen evolution rate ( Figure 2B, red and green).

Purification and MS/MS of NFK-containing Peptides in CP43
NFK has a unique absorption at 318 nm ( Figure 3B, dotted line), when compared to tryptophan ( Figure 3B, solid line), kynurenine ( Figure 3B, dashed line), or other modifications of the indole ring [19]. This unique absorption spectrum allows the identification and purification of NFK-containing tryptic peptides by HPLC (Figure 4). The HPLC chromatogram was monitored at 350 nm during purification of NFK-modified peptides to avoid overlap with the strongly absorbing 280 nm peak. In oxygenevolving PSII, two different NFK-containing peptides, peptides A and C, were identified ( Figure 4A and 4B). Typical absorption spectra, derived from the HPLC chromatograms, are shown in Figure 3A.
Using MS/MS (Table 1 and Figure S1), fraction C (retention time ,28 min.) was identified as NFK-365 in CP43 ( 363 AP(W*)- Figure 2. Steady state rates of oxygen evolution of PSII membranes (A) and TM (B) during high light illumination and in the dark. In A, PSII membranes were kept in the dark at 25uC for two hours (blue). PSII membranes were exposed to a white light intensity of 500 (red) and 7,000 (green) mmol photons m 22 s 21 for two hours at 25uC. In B, TM were kept in the dark (black and red) or exposed to a white light intensity of 7,000 mmol photons m 22 s 21 at chlorophyll concentrations of 1.0 (blue) or 0.1 mg/ml (green). Oxygen evolution was assayed every 30 minutes and normalized to time zero. The data shown are an average of three to six experiments. The error bars are plus and minus one standard deviation. See Materials and Methods, Photoinhibition, for experimental conditions. doi:10.1371/journal.pone.0042220.g002 LEPLRGPNGLDLSR 379 ), confirming our earlier result [19]. Fraction A contained the same NFK-W365 CP43 modification, but the peptide was shorter, 363 AP(W*)LEPLR 370 (Table 1). Only one NFK peptide was detected in Fractions A and C (Table S2). Representative MS/MS data are shown in the Supporting Information ( Figure S1).

CP43 NFK in Photoinhibition
To calculate the yield of NFK, the 350 nm peak was integrated, and the value was normalized to the total 220 nm absorption. This corrects for the yield of tryptic peptides [19]. These data are presented in the bar graph shown in Figure 5. As shown, formation of NFK-W365 in fraction C is light induced in oxygenevolving PSII. The yield increases by 2.160.6 ( Figure 5). This increase parallels the 3.460.4 fold decrease observed in the steady state oxygen evolution rate (Figure 2A, blue and green). However, fraction A ( Figure 4A and 4B) does not show a significant lightinduced increase (0.960.2, Figure 5 and Table S1).
In TW PSII, the Mn 4 CaO 5 cluster and extrinsic subunits are removed [30]. Under these conditions ( Figure 1B), without active oxygen evolution, no significant light induced increase is observed in fraction C (Figures 4C and 4D, Figure 5). Fraction A is not observed in TW PSII.

NFK365-CP43 is Observed in TM, but does not Show a Light-induced Increase
In TM samples, fractions A and C are observed in the dark and the light ( Figure 4E and 4F). Field grown spinach leaves, exposed to unregulated growth conditions, were used for the TM isolation. Thus, NFK modifications may be present in the dark, due to the  Table  S1 for average retention times and summary of light-induced changes. Fraction C corresponds to fraction 1 in ref [19]. doi:10.1371/journal.pone.0042220.g004 previous handling of the market spinach. The observation of NFK in TM demonstrates that the modification is not induced by detergent treatment. There is no significant, light-induced increase in these fractions ( Figure 5 and Table S1). This parallels the results of the oxygen evolution assays conducted under the same conditions (1 mg/mL chlorophyll) in Figure 2B (black and blue), which showed that the TM preparation was resistant to photoinhibition.

Photoinhibition at Increased Ionic Strength Results in a New NFK-containing Peptide, Peptide B
The photoinhibition experiment was conducted on PSII membranes in SMN buffer to which 2 mM NaCl was added. Compared to PSII in SMN buffer ( Figures 6A and 6B), a new peptide, peptide B, was observed in the light (Figures 6C and 6D). Fraction B was not observed in the dark ( Figure 6C). Under these conditions there was no significant increase in the yield of fractions A or C ( Figure 5). MS/MS identified peptide B as 313 VINT(W*)A-DIINR 323 in D1 (Tables 1 and S2, Figure S1). Only one NFK peptide was detected in this fraction (Table S2).
To test if the observation of peptide B depended on the identity of the cation or anion, the experiment was conducted in the presence of ZnCl 2 (0.15 mM) and Na 2 EDTA (1 mM). Peptide B was observed under both sets of conditions ( Figure 5). This result is not consistent with a role for a specific mono-or divalent ion.
We conclude that the small ionic strength increase underlies the observation of fraction B. The ionic strength of the SMN buffer, prior to the addition of salts, is calculated to be 34.9 mM. The ionic strength increased to 36.9 mM (2 mM NaCl or TMA), 35.4 mM (0.15 mM ZnCl 2 ), and 36.4 mM (1 mM Na 2 EDTA) when peptide B was observed in the light.
Approximately 7% of TW PSII reaction centers were reported to contain CP43 NFK-365 [19]. Assuming the same extinction coefficients (3750 M 21 cm 21 at 321 nm [36], the yield of D1 NFK-317 (fraction B) can be estimated. Comparison of HPLC peak intensities indicates that approximately 1% of the PSII centers contain D1 NFK-317 after photoinhibition.   (Table S1) for summary of fold changes and average retention times. doi:10.1371/journal.pone.0042220.g005 nation. Illumination in the presence of increased NaCl, ZnCl 2 , and Na 2 EDTA gave the same result ( Figure 7).

An Additional Light-induced NFK Modification is Observed in PSII
An additional NFK peptide was detected with a 34 min retention time (Tables S1 and S2). This fraction (D) increased in intensity in the light. More than one NFK peptide was detected in this fraction, with one identified as the D2 polypeptide 8 FTKDEKDLFDSMDD(W*)LR 24 and the other identified as the D2 polypeptide 14 DLFDSMDD(W*)LR 24 (Tables S1 and S2). In our previous work, which employed HPLC and affinity purification, an NFK modification of a light-harvesting subunit was detected with a similar retention time. Due to the complexity of this fraction, interpretation of the light induced increase in fraction (D) awaits further experimentation.

Summary
In this work, three NFK containing peptides, originating from the donor side of PSII, are identified. Fraction A, corresponds to 363 AP(W*)LEPLR 370 in CP43 and is observed in oxygen-evolving PSII and TM, but not in TW PSII. Fraction A showed no detectable light-induced increase in any sample that we examined. Fraction C, corresponds to 363 AP(W*)LEPLRGPNGLDLSR 379 in CP43, and is observed in oxygen evolving PSII, TW PSII, and TM. Fraction C showed a light induced increase in only one sample, oxygen-evolving PSII. Fraction B corresponds to 313 VINT(W*)ADIINR 323 in D1. It was observed only in oxygenevolving PSII, after illumination and under conditions of higher ionic strength. Figure 8 shows the position of the NFK modifications in the PSII structure from T. vulcanus [2]. NFK317-D1 is located ,24 Å away from NFK365-CP43. NFK 365-CP43 is 17 Å from the Mn 4 CaO 5 cluster; NFK 317-D1 is 14 Å from the cluster. Figure 8 also shows the position of NFK365-CP43 and NFK317-D1 relative to P 680 and YZ. YZ is an electron transfer intermediate during the water oxidizing reaction ( Figure 1B) [37]. YZ is oxidized by the primary chlorophyll donor, P 680 , and in its radical form, YZ is a strong oxidant [38]. However, NFK365-CP43 is 19 Å from YZ and 30 Å from P 680 . NFK317-D1 is 27 Å from YZ and 19 Å from P 680 . Thus, a YZ or P 680 radical-based mechanism for the formation of the NFK modifications seems unlikely.  (Table S1). Fraction C corresponds to fraction 1 in ref [19]. doi:10.1371/journal.pone.0042220.g006  Sequence conservation across photosynthetic organisms supports an important evolutionary role for both NFK-modified tryptophans. Although these core subunits are consistent in plants and cyanobacteria, some distinct structural distinctions remain. These differences mainly lay in the extrinsic [39] and light harvesting antennae [40] polypeptides. Thus, the location and role of NFK in cyanobacteria remains to be determined.

ROS and Specificity of NFK Modifications in PSII and Other Proteins
In our experiments, we attribute the formation of NFK to the reaction of the Trp side chain with ROS [41]. Studies in mitochondrial proteins have concluded that the NFK modification is a ROS-targeted mechanism [14,42,43]. In mitochondrial ATP synthase, the NFK modification was channeled to a single Trp residue (Trp-503) [14]. In the mitochondrial aconitase-2 protein, site-specific oxidation of Trp-373 was also observed [42]. The specificity of the post-translational NFK modification in PSII and other proteins suggests a selective physiological role for the modification.
We demonstrate here that removal of the Mn 4 CaO 5 cluster and extrinsic subunits prevents light-induced accumulation of NFK in the CP43 subunit. Previously, EPR spin trapping experiments have suggested that photoinhibited oxygen evolving PSII produces both 1 O 2 and HO N [44]. However, only HO N was detected in Mndepleted PSII [44]. Coupled with our data, this previous result is supportive of the conclusion that 1 O 2 [15], and not HO N [22], reacts with Trp to form NFK. Our results suggest a Mn 4 CaO 5 origin for the reactive oxygen species, which oxidizes the Trp side chain. However, we cannot rule out the possibility that extrinsic subunit removal or acceptor side alterations are contributing factors.

NFK in D1 Turnover and Photoinhibition
Photoinhibition is known to induce D1 protein damage and a high rate of D1 turnover [10,12,45]. Previous studies have identified intermolecular cross-links of D1 with D2, cytochrome b 559 , and CP43 [46,47]. In intact leaves and chloroplasts, D1 damage and turnover also occurred by D1 fragmentation and cross-linking [46]. These cross-links were proposed to participate in pathways for complete degradation of damaged D1 in vivo [46,47]. NFK can bind covalently to primary amine-containing side chains [19], such as arginine and lysine, and may participate in this proteolysis pathway. The Western blot analysis of PSII membranes, presented here, confirmed that D1 cleavage and oligomerization occurred when NFK accumulated either in CP43 or in D1. The 3.460.4 decrease in oxygen evolution rate of PSII membranes under the same conditions is further support for a correlation of photoinhibitory effects with NFK formation.
Reversible, light induced structural changes in the degree of spinach TM stacking (grana) have been observed by electron microscopy [48]. These dynamic alterations in structural organization may be involved in protection from light stress [49] and would not occur in isolated PSII membrane fractions. In future work, we will explore the impact of these topological changes. In these experiments, we compared TM with PSII for two reasons. First, we wished to examine the possibility that the NFK modification is induced by detergent treatment. Observation of the NFK modification in TM in the dark eliminates this possibility. The residual level of oxidative modification in the dark may be due to the use of market, field grown spinach, which is transported and harvested under uncontrolled conditions. Second, TM samples do not photoinhibit at the high chl concentrations necessary for the HPLC assay. Therefore, TM preparations provide an important negative control for the PSII experiments. We report that illumination of TM did not accumulate NFK, supporting the conclusion that the increase in NFK yield is caused by light stress. We attribute the resistance to photoinhibition in the TM to a shading effect [50], because illumination of TM with the same light intensity at a lower chlorophyll concentration (0.1 mg/ mL) significantly decreased activity.

Effects of Ionic Strength on Oxidative Modifications
In this work, we found the interesting result that small ionic strength increases had a dramatic effect on the pattern of NFK modifications. However, this change did not alter the degradation pattern of D1 as assessed with a C-terminal antibody. Although the D1 degradation pattern was not changed, the alternative D1-NFK modification to Trp-317 was induced by increasing ionic strength. The ionic strength effect may cause conformational changes in the extrinsic loops. Lowering of the thylakoid lumen pH during excess light involves protein conformational changes that may be necessary in non-photochemical quenching [51].
Concomitant Mg 2+ efflux into the stroma occurs during the transition from dark to light conditions [52,53]. Thus, ionic strength induced events are essential in regulatory pathways in TM and PSII. Further evidence for loop dynamics is provided by the inefficient tryptic cleavage of CP43 in TW PSII, noted here. For example, both the 363 AP(W*)LEPLR 370 peptide (fraction A) and the 363 AP(W*)LEPLRGPNGLDLSR 379 (fraction C) peptide were observed in intact PSII and TM. However, in TW PSII, the shorter CP43 peptide 363 AP(W*)LEPLR 370 was not detected. These results can be attributed to different conformations of the CP43 loop region. These conformational changes may be important in control of photoinhibitory responses in the chloroplast, where changes in the proton motive force can occur during illumination.

NFK in PSII Signaling and Repair
The signaling pathways for induction and control of D1 turnover are not known. Oxidative PTMs of aromatic amino acids have been proposed to participate in signaling. The NFK modifications identified here may function as these signals. We showed previously that substitutions at Trp-365 (Trp-352 in Synechocystis 6803) did not affect the steady-state rate of oxygen evolution under normal light-saturated conditions [54]. This result indicated that mutations at Trp-365 do not alter the structure of PSII or change the overall rates of electron transfer. However, the mutants displayed an increased rate of photoinhibition at higher light intensities (5,000 mmol photons m 22 s 21 ) [55]. Thus, the inability to form NFK in the mutants resulted in reduced repair during high light stress. Because the light-induced increases in NFK in CP43 and D1 appear to be mutually exclusive, we propose that these modifications occur on two different damage/ repair pathways. Inhibition of the CP43 pathway promotes the D1 oxidative pathway.
The primary proposed protease involved in D1 turnover, FtsH [56,57], has been proposed to recognize partially unfolded proteins [58]. Oxidation of Trp to NFK may promote partial protein unfolding required for signal recognition by the protease [56]. Replacement of NFK with unmodified Trp requires de novo protein synthesis [59]. Multiple NFK modifications may be required for continuous D1 turnover. Interestingly, an increase in CP43 degradation and cross-linking was observed during photoinhibition and donor side inactivation [29].

Conclusions
Our data provide evidence for specific oxidative modifications of PSII subunits. These PTMs are induced by high light stress and are under differential control of ionic strength. We propose that NFK plays a role in signaling for repair during D1 turnover. In a two-pathway signaling model for repair, inhibition of one NFK signaling pathway (the ''CP43'' pathway) stimulates repair by the alternative pathway (the ''D1'' pathway). These results provide new insight into redox signaling in oxygenic photosynthesis.

Photoinhibition
Photoinhibition experiments were performed on spinach TM, PSII membranes, and TW PSII membranes [19]. The final resuspension buffers were used during illumination (see Materials and Methods, Thylakoid, PSII, and TW PSII membrane preparations above). All samples were stirred and kept at 25uC with a water bath and dewar, during white light illumination with

UV-Visible Spectrophotometry
Optical spectra of model compounds tryptophan, NFK, and kynurenine were recorded at room temperature from 200-750 nm on a Hitachi (U3000) spectrophotometer [19]. The model compounds, 40 mM L-tryptophan (Sigma-Aldrich, St. Louis, MO), L-kynurenine (Sigma-Aldrich), and NFK [19,64] were suspended in H 2 O. The NFK-containing peptide optical spectra were derived from the chromatogram through the use of a Beckman (Brea, CA) System GoldH HPLC, equipped with a 125 solvent module, a 168 photodiode array detector (1 cm path length, 2 nm scan interval), and 32 Karat Software, version 7.0.

Tryptic Peptide Digestion and High Pressure Liquid Chromatography (HPLC) Assay
In-situ trypsin (Life Technologies, Carlsbad, CA) digestion of TM, PSII, and TW PSII was conducted as described [19]. HPLC separation, isolation of NFK-containing peptides, and quantitative NFK assay were carried out as previously described [19]. Retention times for Fractions A-D were 25, 26, 28, and 34 minutes and were reproducible to 60.6 min (Table S1). The amount of the NFK containing peptide was quantitated by integration of the 350 nm peak by the procedure previously described [19]. This area was normalized to the total 220 nm absorption. This normalization corrects for any differences in the yield of tryptic products (Table S1).

Tandem Mass Spectrometry (MS/MS) Peptide Analysis
PSII tryptic peptides were analyzed as described [19]. Representative MS/MS data are shown in Figure S1. Figure S1 Representative MS/MS spectra of NFK modifications in CP43 (A and C) and D1 (B) proteins. The peaks in blue represent the b-fragments. The peaks in red represent the y-fragments. The NFK modified W is indicated in the corresponding sequences. This residue carries the +32 m/z masss shift, which was unambiguously assigned to Trp-365 in CP43 (A and C) and Trp-317 in D1 (B).

(DOCX)
Table S1 # NOD, not observed in dark; NODL, not observed in dark or light. (DOCX) Table S2 For the identification of peptides, filter criteria were set to warrant a false discovery rate of less than 1% on the peptide level. In each of the three independent LC-MS/MS runs of the four fractions, more than 20000 MS/MS spectra were recorded. For fraction A-C, between 3500-5000 spectra were assigned to peptides from 50-80 proteins from S. oleracea. In fraction D only, 1200 spectra could be assigned to peptides of about 50 proteins. (DOCX)