Evidence for the Involvement of Loosely Bound Plastosemiquinones in Superoxide Anion Radical Production in Photosystem II

Recent evidence has indicated the presence of novel plastoquinone-binding sites, QC and QD, in photosystem II (PSII). Here, we investigated the potential involvement of loosely bound plastosemiquinones in superoxide anion radical (O2 •−) formation in spinach PSII membranes using electron paramagnetic resonance (EPR) spin-trapping spectroscopy. Illumination of PSII membranes in the presence of the spin trap EMPO (5-(ethoxycarbonyl)-5-methyl-1-pyrroline N-oxide) resulted in the formation of O2 •−, which was monitored by the appearance of EMPO-OOH adduct EPR signal. Addition of exogenous short-chain plastoquinone to PSII membranes markedly enhanced the EMPO-OOH adduct EPR signal. Both in the unsupplemented and plastoquinone-supplemented PSII membranes, the EMPO-OOH adduct EPR signal was suppressed by 50% when the urea-type herbicide DCMU (3-(3,4-dichlorophenyl)-1,1-dimethylurea) was bound at the QB site. However, the EMPO-OOH adduct EPR signal was enhanced by binding of the phenolic-type herbicide dinoseb (2,4-dinitro-6-sec-butylphenol) at the QD site. Both in the unsupplemented and plastoquinone-supplemented PSII membranes, DCMU and dinoseb inhibited photoreduction of the high-potential form of cytochrome b 559 (cyt b 559). Based on these results, we propose that O2 •− is formed via the reduction of molecular oxygen by plastosemiquinones formed through one-electron reduction of plastoquinone at the QB site and one-electron oxidation of plastoquinol by cyt b 559 at the QC site. On the contrary, the involvement of a plastosemiquinone formed via the one-electron oxidation of plastoquinol by cyt b 559 at the QD site seems to be ambiguous. In spite of the fact that the existence of QC and QD sites is not generally accepted yet, the present study provided more spectroscopic data on the potential functional role of these new plastoquinone-binding sites.


Introduction
Photosystem (PSII) is a heterodimeric multiprotein-pigment complex embedded in the thylakoid membrane of photosynthetic organisms such as cyanobacteria, algae and higher plants. Recent X-ray crystallographic structural analyses of PSII from the cyanobacteria Thermosynechococcus elongatus and Thermosynechococcus vulcanus demonstrated that PSII consists of 20 protein subunits, 35 chlorophylls, 12 carotenoids and 25 lipids per monomer [1][2][3]. During oxygenic photosynthesis, PSII functions as a water-plastoquinone oxidoreductase that oxidizes water to molecular oxygen and reduces plastoquinone to plastoquinol [4][5]. In these reactions, four electrons extracted from water by a water-splitting manganese complex on the electron donor side of PSII are transferred to the primary and secondary electron acceptors on the electron acceptor side of PSII [6][7][8][9]. It is well established that the primary and secondary electron acceptors are plastoquinones tightly and loosely bound to the Q A and Q B sites, respectively. One-electron reduction of plastoquinone at the Q B site forms plastosemiquinone (Q B •− ), which is subsequently stabilized by the protonation of proximal amino acid side chains (Q B H • ), whereas the sequential one-electron reduction and protonation of Q B H • forms plastoquinol (Q B H 2 ). Several biochemical studies have suggested that PSII contains two plastoquinone-binding sites in addition to the Q A and Q B sites [10][11][12]. Based on the study on photoreduction of cytochrome b 559 (cyt b 559 ) in the presence of exogenous plastoquinone, a third plastoquinone-binding site referred to as Q C was proposed to be located closed to cyt b 559 [10]. Later, the effects of herbicides and ADRY agents on the redox properties of cyt b 559 provided more biochemical data on the existence of Q C site [11][12]. Consistent with biochemical studies, the crystal structure of PSII at 2.9 Å resolution revealed the existence of Q C site [2]. However, the Q C site was not reported in the most recent PSII crystal structure at 1.9 Å resolution [3]. Hasegawa and Noguchi proposed that the affinity of plastoquinone to the Q C site is lower compared to the Q B site [13]. In agreement with this proposal, it has been recently suggested that ambiguity in the existence of Q C site might be due to the different purification and crystallization procedures [14]. Recently, Kaminskaya and Shuvalov [15] identified a fourth plastoquinone-binding site denoted as Q D . The authors concluded that the Q C site depicted in the PSII crystal structure is in a highly hydrophobic environment, while the Q D site is located in a more polar environment. The urea-type herbicide DCMU (3-(3,4-dichlorophenyl)-1,1-dimethylurea) blocks Q B to Q B •− reduction at the Q B site, whereas the phenolic-type herbicide dinoseb (2,4-dinitro-6-sec-butylphenol) prevents the oxidation of plastoquinol (Q D H 2 ) to plastosemiquinone (Q D H • ) by cyt b 559 at the Q D site [15]. The limitations on electron transport both on the electron donor and acceptor sides of PSII are associated with the formation of reactive oxygen species (ROS) [16][17][18][19]. Under high-light conditions, when light absorption by chlorophylls exceeds the utilization of excitation energy, the over-reduction of the electron acceptor side of PSII leads to leakage of electrons to molecular oxygen. The reduction of molecular oxygen results in the formation of superoxide anion radical (O 2 •− ), which either spontaneously dismutates to hydrogen peroxide (H 2 O 2 ) or forms bound peroxide through interactions with the non-heme [20] or heme iron in cyt b 559 [21]. ) [24], free plastosemiquinone (PQ •− ) [25] and the ferrous heme iron in the low-potential (LP) form of cyt b 559 [26].
Due to a highly negative redox potential (Em (Pheo/Pheo •− ) = -505 to -610 mV, pH 6.5 to 7) [27][28], the reduction of molecular oxygen by Pheo •− is likely. The favorable thermodynamic properties for reduction of molecular oxygen by Pheo •− are limited by kinetic restrictions. Forward electron transport from Pheo •− to Q A •− is much more rapid than diffusion-limited reduction of molecular oxygen; thus, the reduction of molecular oxygen by Pheo •− is less likely. However, under certain circumstances, such as limitation of electron transport from Pheo •− to Q A •− , the Pheo •− lifetime is prolonged, and the reduction of molecular oxygen is more likely.
In contrast to Pheo •− , reduction of molecular oxygen by Q A ; however, the probability of its formation by the interaction of free plastoquinone and free plastoquinol is very low [25]. It has been proposed that the reduction of molecular oxygen by ferrous heme iron in the LP form of cyt b 559 produces O 2 •− and may be thermodynamically feasible because the LP form of cyt b 559 has a low midpoint redox potential (Em = -40 to +80 mV, pH 7) [21,26,33]. Herein, we studied whether loosely bound plastosemiquinones are involved in the light-induced O 2 •− formation in PSII membranes using an electron paramagnetic resonance (EPR) spin-trapping spectroscopy. We provide evidence that O 2 •− is produced via one-electron reduc- Materials and Methods 1. PSII membrane preparation PSII membranes were isolated from fresh spinach leaves using the method reported previously by Berthold et al. [34] with modifications described by Ford and Evans [35]. The isolated PSII membranes were dissolved in a buffer solution containing 400 mM sucrose, 10 mM NaCl, 5 mM CaCl 2, 5 mM MgCl 2 and 50 mM Mes-NaOH (pH 6.5) and stored at -80°C until further use. For PQ-supplemented PSII membranes, exogenous short-chain platoquinone containing one isoprenoid units in the side-chain (PQ-1) was added to the PSII membranes prior to illumination. 30 μM PQ-1 was added to the PSII membranes as an ethanol solution (the final concentration of ethanol did not exceed 1%).

Optical measurements
The redox properties of cyt b 559 were studied using an Olis RSM 1000 spectrometer (Olis Inc., Bogart, Georgia, USA). The redox states of cyt b 559 in PSII membranes (150 μg Chl ml -1 ) were determined based on the changes in the absorbance at 559 nm upon stepwise additions of 50 μM potassium ferricyanide, 8 mM hydroquinone, 5 mM sodium ascorbate and sodium dithionite in a cuvette at room temperature using the method in Tiwari and Pospíšil [21] with certain modification. The redox forms of cyt b 559 in the PSII membranes were determined by subtracting the control from the treatment spectra: for the HP form of cyt b 559 , the hydroquinone-reduced spectra were subtracted from the ferricyanide-oxidized cyt b 559 ; for the IP form of cyt b 559 , the ascorbate-reduced spectra were subtracted from the hydroquinone-reduced cyt b 559 ; and for the LP form of cyt b 559 , the dithionite-reduced spectra were subtracted from the ascorbate-reduced cyt b 559 . In photoreduction measurements, the photoreduced HP form of cyt b 559 (PH) was calculated based on the difference between the absorbance spectra measured after illumination for 180 s and the dark-adapted ferricyanide oxidized spectrum and hydroquinone-reduced spectra were subtracted from photoreduced HP form of cyt b 559 to get unreduced HP form of cyt b 559 . The PSII membranes were illuminated with continuous white light (1000 μmol photons m -2 s -1 ) in the cuvette, which was rotated by 90°at intervals of 15 s.

High-pressure liquid chromatography
The loosely bound plastoquinone was measured using the method in Wydrzynski and Inoue [36]. A 1 ml aliquot of the PSII membranes (300 μg Chl ml -1 ) was mixed with 3 ml of heptane and 30 μl of isobutanol, followed by vortexing for 1 h in the dark. The mixture was then centrifuged at 4000 x g for 10 min. The plastoquinone content in the upper organic layer was determined by HPLC based on the method of Kruk and Karpinski [37].

Superoxide anion radical production in unsupplemented PSII membranes
Light-induced O 2 •− formation in the unsupplemented PSII membranes was measured using EPR spin-trapping spectroscopy. For spin-trapping, we used the spin trap compound EMPO, which reacts with O 2 •− to form an EMPO-OOH adduct [38]. No EMPO-OOH adduct EPR signal appeared immediately after addition of EMPO to the unsupplemented PSII membranes in the dark (Fig. 1A). Illumination of the unsupplemented PSII membranes in the presence of EMPO resulted in the production of an EMPO-OOH adduct EPR signal (Fig. 1A). To prevent EMPO-OH adduct formation, the strong iron chelator Desferal was used to decrease the level of free iron available to produce HO • through the Fenton reaction [26,39]. Fig. 2B shows the time profile for the EMPO-OOH adduct EPR signal measured for the unsupplemented PSII membranes. These results demonstrate that the illumination of unsupplemented PSII membranes results in the formation of O 2 •− .

Superoxide anion radical production in PQ-supplemented PSII membranes
To study the role of loosely bound plastosemiquinone in O 2 formation was measured in the presence of exogenous PQ-1. Because PQ-1 is smaller than the natural molecule PQ-9, PQ-1 can better penetrate the membrane and substitute for PQ-9 as an electron acceptor in PSII. The observation that the addition of PQ-1 to EMPO did not generate any EPMO-OOH adduct EPR spectrum indicates that PQ-1 does not directly interact with EMPO (data not shown). In the dark, the addition of PQ-1 to the PSII membranes in the presence of EMPO did not produce an EPR signal; however, exposure of PQ-supplemented PSII membranes to white light resulted in the formation of an EMPO-OOH adduct EPR signal (Fig. 1C). The time profile of the EMPO-OOH adduct EPR signal measured after addition of exogenous PQ-1 to the PSII membranes revealed that the intensity of the EMPO-OOH adduct EPR signal was enhanced by 70% as compared to unsupplemented PSII membranes (Fig. 1D). These results indicate that plastosemiquinones are involved in light-induced O 2 •− production in PSII.  The relative intensity (mean ± SD, n = 3) of the light-induced EMPO-OOH adduct EPR signal measured using unsupplemented and PQ-supplemented PSII membranes. The other experimental conditions were the same as described in Fig. 1. doi:10.1371/journal.pone.0115466.g002

The effects of DCMU and dinoseb on superoxide anion radical production in unsupplemented PSII membranes
To investigate where loosely bound plastosemiquinones involved in O 2 •− production are formed, the effects of two herbicides, DCMU (bound at the Q B site) and dinoseb (bound at the Q D site) on the EMPO-OOH adduct EPR signal were studied in the unsupplemented PSII membranes. When the unsupplemented PSII membranes were illuminated in the presence of DCMU, the EMPO-OOH adduct EPR signal was suppressed by 50%, whereas the remaining EPMO-OOH EPR signal (50%) was insensitive to DCMU ( Fig. 2A and C). In previous studies [24,26,40,41], the relative proportion of DCMU-sensitive and DCMU-insensitive O 2 •− production in PSII varied, likely due to the endogenous platoquinone content. In addition to the EMPO-OOH adduct EPR signal, the EPR spectrum measured in the presence of DCMU comprises an EMPO-R adduct EPR signal formed by the interaction between EMPO and a carboncentered radical, the origin of which is unknown. These observations reveal that 1) the DCMU-sensitive EMPO-OOH adduct EPR signal corresponds to O 2 •− formed at or after the Q B site (i.e., reduction of molecular oxygen by loosely bound plastosemiquinones formed by one-electron reduction of plastoquinone and one-electron oxidation of plastoquinol) and 2) the DCMU-insensitive EMPO-OOH adduct EPR signal corresponds to O 2 •− , which is formed before the Q B site (i.e., reduction of molecular oxygen by Pheo •− and Q A

•−
). When dinoseb was added to the unsupplemented PSII membranes prior to illumination, the EMPO-OOH adduct EPR signal was enhanced by 25% ( Fig. 2A and C). Due to the fact that the occupation of the Q D site does not eliminate O 2 •− production, the production of O 2 •− by reduction of molecular oxygen by plastosemiquinone at the Q D site is ambiguous.

The effects of DCMU and dinoseb on superoxide anion radical production in PQ-supplemented PSII membranes
Addition of DCMU to PQ-supplemented PSII membranes decreased the EMPO-OOH adduct EPR signal by 55% ( Fig. 2B and C). Similar to unsupplemented PSII membranes, in PQsupplemented PSII membranes, 1) O 2 •− is formed at or after the Q B site via reduction of molecular oxygen by plastosemiquinone formed via one-electron reduction of plastoquinone and one-electron oxidation of plastoquinone and 2) O 2 •− is formed prior to the Q B site by reduction of molecular oxygen by Pheo •− and Q A •− . The intensity of the EMPO-OOH adduct EPR signal after the addition of DCMU was higher for the PQ-supplemented PSII membranes than for the unsupplemented PSII membranes (Fig. 2C). When dinoseb was added to the PQsupplemented PSII membranes prior to illumination, the EMPO-OOH adduct EPR signal was enhanced by 17% ( Fig. 2B and C). The intensity of the EMPO-OOH adduct EPR signal after the addition of dinoseb was higher for PQ-supplemented PSII membranes compared to unsupplemented PSII membranes (Fig. 2C). Similar to the unsupplemented PSII membranes, the effect of dinoseb on O 2 •− production in PQ-supplemented PSII membranes indicate that the Q D site is unlikely involved in O 2 •− production.

Different redox forms of cyt b 559 in the unsupplemented and PQsupplemented PSII membranes
To determine the different redox forms of cyt b 559 , we measured changes in absorption at 559 nm in the unsupplemented and PQ-supplemented PSII membranes. The different redox forms of cyt b 559 were discerned by examining the hydroquinone-reduced minus ferricyanideoxidized (HP) spectra, ascorbate-reduced minus hydroquinone-reduced (IP) spectra, and dithionite-reduced minus ascorbate-reduced (LP) spectra. In the unsupplemented PSII membranes, 40% of cyt b 559 was in the hydroquinone-reducible HP form, 22% was in the sodium ascorbate-reducible IP form, and 38% was in the dithionite-reducible LP form (Fig. 3A). In the supplemented PSII membranes, the levels of the hydroquinone-reducible HP, sodium ascorbate-reducible IP and dithionite-reducible LP forms of cyt b 559 were 42, 12 and 46% (Fig. 3B). These observations confirm the presence of the HP, IP and LP forms of cyt b 559 in both unsupplemented and PQ-supplemented PSII membranes.

Cyt b 559 photoreduction in the unsupplemented and PQsupplemented PSII membranes
To observe the light-induced reducible redox form of cyt b 559 , cyt b 559 photoreduction was measured in both unsupplemented and PQ-supplemented PSII membranes. When the unsupplemented PSII membranes were exposed to white light, the HP form of cyt b 559 was reduced (Fig. 3C). Addition of hydroquinone to the unsupplemented PSII membranes after illumination did not further reduce the HP form of cyt b 559 ; however, addition of sodium ascorbate and sodium dithionite reduced the IP and LP forms of cyt b 559 (Fig. 3C). Similarly, exposure of PQsupplemented PSII membranes to white light reduced the HP form of cyt b 559 (Fig. 3D); however, addition of hydroquinone to PQ-supplemented PSII membranes after illumination did not further reduce the HP form. Addition of ascorbate and dithionite to PQ-supplemented PSII membranes reduced the IP and LP forms of cyt b 559 (Fig. 3D). These results demonstrate that illumination of the unsupplemented and PQ-supplemented PSII membranes reduced the HP form of cyt b 559 .

The effects of DCMU and dinoseb on cyt b 559 photoreduction in the unsupplemented and PQ-supplemented PSII membranes
To confirm the involvement of the Q B site in cyt b 559 photoreduction via mobile plastoquinol, cyt b 559 photoreduction was measured in the presence of DCMU. Addition of DCMU to unsupplemented or PQ-supplemented PSII membranes prior to illumination fully prevented photoreduction of the HP form of cyt b 559 (Fig. 4A and 4B). These results indicate that DCMU prevents photoreduction of HP form of cyt b 559 due to inhibition of plastoquinol formation.
To confirm the involvement of the Q D site in the photoreduction of cyt b 559 , cyt b 559 photoreduction was measured in the presence of dinoseb. Illumination of PSII membranes in the presence of dinoseb did not cause cyt b 559 photoreduction in both unsupplemented (Fig. 4C) and PQ-supplemented PSII membranes (Fig. 4D). These results suggest that dinoseb convert HP form to LP form of cyt b 559 and prevents reduction of cyt b 559 at the Q D site due to inhibition of plastoquinol oxidation.

Quantifying loosely bound PQ and chlorophyll in PSII
To correlate the PQ-binding site and O 2 •− formation in PSII membranes, the content of loosely bound plastoquinone was measured by HPLC. HPLC analysis of the chlorophyll content indicated approximately 250 chlorophyll molecules per reaction center (RC), consistent with values in the literature (i.e., 200-300 Chl/RC) [35,42]. HPLC analysis of plastoquinone levels demonstrated that two of three plastoquinones per RC were extractable from the PSII membranes. These observations suggest that one plastoquinone is tightly bound (Q A ) and two plastoquinones are loosely bound (Q B and Q C or Q D ).

Discussion
Several lines of evidence have been provided that O 2 •− is formed through one-electron reduction of molecular oxygen on the electron acceptor side of PSII [16,17]. As the operational redox potential for the O 2 /O 2 •− redox couple is close to 0 mV or even positive due to the difference in concentration of molecular oxygen and O 2 •− , O 2 •− formation requires a suitable electron donor with a redox potential lower than the operational redox potential of O 2 /O 2 •− redox couple, and thus consequently, a high reducing power to reduce molecular oxygen. It was suggested that various cofactors on the electron acceptor side of PSII can fulfil such thermodynamic criteria and thus might serve as potential electron donors to molecular oxygen. Although light-induced O 2 •− formation in PSII has been examined by measuring oxygen consumption [43][44][45], ferricytochrome c reduction and the xanthine/xanthine oxidase assay [22], voltametric methods [23] and EPR spin-trapping spectroscopy [20, 24, 26, 40-41, 46, 47], the molecular mechanism underlying light-induced O 2  are lower than the operational redox potential of O 2 /O 2 •− redox couple (close to 0 mV or even positive), the reduction of molecular oxygen by plastosemiquinones is feasible. Based on the presented data, we propose that O 2 •− is produced by one-electron reduction of molecular oxygen by plastosemiquinones formed by one-electron reduction of plastoquinone at the Q B sites and one-electron oxidation of plastoquinol at the Q C site but most likely not the Q D site (Fig. 5).

Involvement of the Q B site in O 2 •− production
In the EPR spin-trapping data obtained using the urea-type herbicide DCMU, the EMPO-OOH adduct EPR signal was only partially suppressed, which indicates that molecular oxygen is reduced prior to the Q B site ( Fig. 2A and B) , which subsequently forms the more stable Q B H • by protonation of proximal amino acids. Subsequent Q B H • reduction and protonation yield Q B H 2 , which moves out through the channels [11]. However, if protonation of Q B •− by

Involvement of the Q C site in O 2 •− production
Based on X-ray crystal structural analyses of the PSII complex, Q B H 2 exchange by plastoquinone at the Q B site was proposed to occur via plastoquinol diffusion through channel I (bottom channel) and II (upper channel) [2]. During this process, Q B H 2 liberates from the Q B site and diffuses through the bottom channel to the Q C site located in the vicinity of the heme iron of cyt b 559 at distance of 17 Å from the head group of plastoquinol. Plastoquinol binding at the Q C site was proposed to favour electron donation to the ferric heme iron of cyt b 559 [46]. Illumination of PSII membranes caused the photoreduction of the HP form of cyt b 559 , demonstrating that Q C H 2 is oxidized by the ferric heme iron of cyt b 559 to form Q C H • . Here, we propose that Q C H • reduces molecular oxygen to O 2 •− . Because addition of dinoseb to the PSII membranes partially enhanced O 2 •− formation (Fig. 2C), we propose that the ferrous heme iron of LP cyt b 559 reduces molecular oxygen, which forms O 2 •− . Fig. 4C and D) show that the HP form of cyt b 559 was converted to the LP form in the presence of dinoseb, as previously demonstrated by Kaminskaya and Shuvalov [15]. In addition to binding of dinoseb to Q D site which has been claimed in the recent past , it is also known to bind to Q B site. In such a case, the formation of Q C H • is unlikely formed by oxidation of Q c H 2 ; however, the alternative reaction pathway for formation of Q C H • occurs. Consistent with this proposal, the formation of Q C H • by one-electron reduction of plastoquinone cannot be excluded [26] and thus the involvement of Q C H • and LP form of cyt b 559 in O 2 •− formation via the Q C site might be considered.

Involvement of the Q D site in O 2 •− production
The observation that the phenolic-type herbicide dinoseb, which binds at the Q D site enhanced EMPO-OOH adduct EPR signal further indicates that Q D H formed by plastoquinol oxidation at the Q D site is not involved in O 2 production ( Fig. 2A). Q D H 2 oxidation by the heme iron of the HP form of cyt b 559 and deprotonation by proximal amino acids results in the formation of Q D H • . Kaminskaya and Shuvalov [15] recently suggested that Q D H • is stable at the Q D site, and the midpoint redox potentials of the Q D /Q D H • redox couple are more positive than those of the Q B /Q B •− redox couple (Em = -45 mV, pH 7). Consistent with this proposal, we assume that the reduction of molecular oxygen by Q D H • is not feasible and thus O 2 •− formation at the Q D site is ambiguous.

Conclusion
The data presented in this study demonstrate that loosely bound plastosemiquinones at the Q B and Q C sites are involved in the formation of O 2 •− via one-electron reduction of molecular oxygen. Loosely bound plastosemiquinone Q B •− is formed via one-electron reduction of plastoquinone at the Q B site; however, one-electron oxidation of plastoquinol by cyt b 559 at the Q C site forms Q C H • . By contrast, the results indicated that O 2 •− formation from plastosemiquinones at the Q D site was ambiguous. In addition to loosely bound plastosemiquinone, previous studies have reported the formation of O 2 •by free plastosemiquinone in the PQ pool [25,[43][44][45]. The interaction of plastoquinol with plastoquinone in the PQ pool was suggested to result in the formation of free PQ •− , which reduces molecular oxygen to form O 2 •− . Further studies are needed to elucidate a unifying mechanism for O 2 •formation which involves PQ pool.