Impact of PsbTc on Forward and Back Electron Flow, Assembly, and Phosphorylation Patterns of Photosystem II in Tobacco 1[W][OA]

Photosystem II (PSII) of oxygen-evolving cyanobacteria, algae, and land plants mediates electron transfer from the Mn 4 Ca cluster to the plastoquinone pool. It is a dimeric supramolecular complex comprising more than 30 subunits per monomer, of which 16 are bitopic or peripheral, low-molecular-weight components. Directed inactivation of the plastid gene encoding the low-molecular-weight peptide PsbTc in tobacco ( Nicotiana tabacum ) does not prevent photoautotrophic growth. Mutant plants appear normal green, and levels of PSII proteins are not affected. Yet, PSII-dependent electron transport, stability of PSII dimers, and assembly of PSII light-harvesting complexes (LHCII) are signiﬁcantly impaired. PSII light sensitivity is moderately increased and recovery from photoinhibition is delayed, leading to faster D1 degradation in D psb Tc under high light. Thermoluminescence emission measurements revealed alterations of midpoint potentials of primary/secondary electron- accepting plastoquinone of PSII interaction. Only traces of CP43 and no D1/D2 proteins are phosphorylated, presumably due to structural changes of PSII in D psb Tc. In striking contrast to the wild type, LHCII in the mutant is phosphorylated in darkness, consistent with its association with PSI, indicating an increased pool of reduced plastoquinone in the dark. Finally, our data suggest that the secondary electron-accepting plastoquinone of PSII site, the properties of which are altered in D psb Tc, is required for oxidation of reduced plastoquinone in darkness in an oxygen-dependent manner. These data present novel aspects of plastoquinone redox regulation, chlororespiration, and redox control of LHCII phosphorylation.

PSII, the oxygen-evolving pigment-protein complex of the thylakoid membrane system, is present in photosynthetic organisms from cyanobacteria to vascular plants. The proteins of the photochemically active reaction center (RC), the heterodimeric polypeptides PsbA (D1) and PsbD (D2) and the lowmolecular-weight (LMW) subunits PsbE and PsbF, and the aand b-chains of cytochrome b 559 coordinate the redox cofactors necessary for primary PSII photochemistry. Electron flow following charge separation occurs via several electron carriers of the RC, including pheophytin a, as well as the primary and secondary quinones, Q A and Q B . The striking similarities between cyanobacterial and land plant PSII core components, including the intrinsic light-harvesting antennae, the chlorophyll-binding proteins CP43 and CP47, and their interactions during photosynthetic charge separation, suggest a high degree of structural and functional conservation. Thus, much of the information obtained from crystallization of a cyanobacterial PSII can be applied to that of higher plants (Zouni et al., 2001;Kamiya and Shen, 2003;Ferreira et al., 2004;Loll et al., 2005;Nield and Barber, 2006).
Besides the RC, the PSII core contains a remarkably high number (16) of intrinsic or peripheral LMW peptides. Five of them are encoded in eukaryotes by nuclear genes (psbR, psbTn, psbW, psbX, and psbY1/ psbY2) and 11 by plastid genes (psbE, psbF, psbL, psbJ, psbK, psbM, psbI, psbTc, psbN, psbH, and psbZ), of which most are parts of conserved operons (for review, see Minagawa and Takahashi, 2004;Shi and Schrö der, 2004;Thornton et al., 2005;Mü h et al., 2008). PSII cores of higher plants and cyanobacteria form a functional dimer. In chlorophyll a/b lineages, the additional chlorophyll proteins light-harvesting complex II (LHCII), forming trimers CP24, CP26, and CP29, are associated with dimeric PSII complexes and, except for CP24, partially with PSI, depending on the energy balance between the two photosystems (Dekker and Boekema, 2005;Takahashi et al., 2006). This process is regulated by reversible phosphorylation of these chlorophyllbinding proteins. Excess light excitation of PSII relative to the accessibility of the electron sink induces back electron flow and charge recombination that generates singlet oxygen species, considered one of the causes of photoinactivation and related lightdependent degradation of the D1 protein  as well as of a genetically programmed stress response (Wagner et al., 2004). Another primary event of photodamage to PSII may be a direct inactivation of the oxygen-evolving complex (Hakala et al., 2005;Ohnishi et al., 2005).
The conservation of PSII LMW peptides throughout evolution from cyanobacteria to plastids suggests that those proteins that are not harboring any of the electron carrier components play distinct roles in the structural organization, dynamics, and regulation of PSII electron flow and its protein turnover. Possible roles for the LMW peptides in these processes have been discussed for cyanobacteria, algae, and land plants (Minagawa and Takahashi, 2004;Shi and Schrö eder, 2004;Thornton et al., 2005;Mü h et al., 2008). In this context, it is relevant to mention that inactivation of some of these genes often resulted in different phenotypes in distantly related organisms (Regel et al., 2001;Ohad et al., 2004;Schwenkert et al., 2006). Detailed analyses of knockout mutants of PSII LMW peptides in tobacco (Nicotiana tabacum) have revealed specific roles for PsbE, PsbF, PsbL, and PsbJ (Regel et al., 2001;Swiatek et al., 2003b;Ohad et al., 2004) as well as for PsbZ (Swiatek et al., 2001), PsbI (Schwenkert et al., 2006), and PsbM (Umate et al., 2007).
Two different LMW peptides associated with PSII were originally designated PsbT. PsbTc of 4 kD originates from a chloroplast gene, and the unrelated protein PsbTn of 11 kD originates from a nuclear gene (for review, see Shi and Schrö der, 2004;Mü h et al., 2008). PsbTc is an intrinsic, bitopic LMW protein harboring a single transmembrane helix and, different from most LMW peptides, exposes its N-terminal tail to the lumenal phase of the thylakoid membrane (Shi and Schrö der, 2004). The psbTc gene is located in a highly conserved operon, together with the PSII genes psbB and psbH, the cytochrome b 6 f complex genes petB and petD, encoding cytochrome b 6 and subunit IV, and psbN on the opposite strand. The operon is transcribed into a primary product of 5.7 kb that is processed in a complex way (Barkan, 1988;Kohchi et al., 1988;Westhoff and Herrmann, 1988).
As part of an ongoing project, individual thylakoid LMW peptides were systematically disrupted in tobacco to elucidate the effect of each in the same organism. Here, the first study (to our knowledge) of the effect of psbTc inactivation on PSII function in a higher plant is presented. Our data show that Q A and Q B site properties, forward and back electron flow, phosphorylation, assembly, and stability of PSII proteins are affected in the mutant. In addition, experiments performed with the wild type reveal that in vivo the Q B site of PSII is responsible for the oxidation of reduced plastoquinone and the reduction of oxygen in the dark, presumably to keep the acceptor side of PSII oxidized. This sheds new light on the nature of a terminal oxidase working during chlororespiration.

Inactivation of psbTc
To selectively inactivate psbTc, we inserted a terminatorless amino glycoside 3# adenyl transferase (aadA) cassette into a newly introduced restriction site in the 5# region of the gene to avoid expression of a functional or truncated protein (Fig. 1A). Homoplastomy of transformed lines was confirmed by PCR analysis of isolated chloroplast chromosomes (Fig. 1B). Primers used for PCR did not amplify the 1,021-bp product obtained for the wild type but exclusively amplified the aadA-containing product of 1,952 bp in DpsbTc. Insertion of the selection cassette into a pentacistronic operon could affect the expression, processing, and/or stability of transcripts originating from the same operon or, in this case, from the oppositely transcribed psbN. Therefore, the integrity and levels of well-known processed RNA species (Westhoff and Herrmann, 1988) were estimated by northern-blot analyses using strand-specific probes for psbTc, psbH, and psbN as well as random primed labeled probes for psbB and petB (Fig. 1C). These data confirmed the homoplastomic state of mutant plants, since spliced and unspliced pentacistronic transcripts, detected with psbB, psbTc, psbH, and petB probes, as well as dicistronic psbB-psbT transcripts, are expectedly up-shifted by the size of the aadA cassette. Although no monocistronic psbTc transcripts could be detected in the wild type, the psbB-psbTc intergenic region was processed in the mutant, resulting in the accumulation of a psbTc-aadA transcript of about 1.3 kb. The analysis further demonstrated that processing and expression levels of psbH-, psbN-, and petB-containing transcripts located downstream of psbTc are not notably affected by the aadA insertion (Fig. 1C). Mutant plants grew photoautotropically under greenhouse conditions on soil and appeared green, comparable to wild-type plants.
In contrast to the wild type, PsbTc could not be detected in mutant thylakoid membranes by mass spectrometry using a novel approach, which will be published elsewhere (Granvogl et al., 2008). To ensure that the phenotype was not caused by random nuclear background mutations and inherited solely via chloroplasts, we performed backcrosses with the wild type using three independently generated homoplastomic DpsbTc mutants as female parents. This resulted in a 100% offspring of the mutant phenotype with respect to photosynthetic parameters, such as reduction of the potential maximum quantum yield of PSII (F v /F m ), to about 0.68 6 0.03 compared with 0.82 6 0.01 in the wild type using at least 20 individuals from each cross. This unequivocally indicates that loss of PsbTc accounts exclusively for the phenotype. Three chosen mutant lines, DpsbTc-1 to -3, were used for detailed studies.
Composition of Thylakoid Membrane Proteins in DpsbTc-1 The protein composition of thylakoids isolated from DpsbTc and wild-type plants was estimated by immunoblot analysis using antisera raised against individual thylakoid proteins (Schwenkert et al., 2006). Levels of PSII proteins D1, D2, CP47, and PsbI and the extrinsic polypeptide PsbO of the oxygen-evolving complex in DpsbTc-1 were comparable with those of the wild type ( Fig. 2A). Also, no notable differences were found for the major LHCII (Lhcb1), LHCI (Lhca1), and minor antenna apoproteins CP29, CP26, and CP24 or for components of PSI, the cytochrome b 6 f complex, and ATP synthase ( Fig. 2A; data not shown). This corroborates that translation and accumulation of petB and petD transcripts downstream of psbTc-1 are not affected by the aadA insertion.
Assembly of PSII-LHCII Supercomplexes and Stability of PSII Dimers Are Affected in DpsbTc-1 to -3 To gain information on the assembly and stability of PSII in DpsbTc plants, controlled partial lysates of isolated thylakoids solubilized with 1% b-dodecylmaltoside were resolved by blue native (BN)-PAGE and subsequent SDS-PAGE. The protein patterns uncovered that PSII-LHCII supercomplexes were almost missing and levels of dimeric PSII complexes were reduced, whereas trimeric LHCII complexes, as well as monomeric PSII RC complexes lacking CP43, called RC47, were increased in DpsbTc-1 to -3 thylakoids compared with the wild type ( Fig. 2B; Supplemental  Fig. S1A). Levels and sizes of all other detectable thylakoid membrane complexes were not significantly affected. To estimate a possible loss of unstable PSII supercomplexes in the mutant during sample preparation, milder separation of thylakoid lysates by Suc density gradient centrifugation and BN-PAGE using 1.5% digitonin  was applied. Separation profiles confirmed reduced levels of PSII-LHCII supercomplexes in the mutant (Supplemental Fig. S2, A and B). Levels of monomeric RC47 complexes and trimeric LHCII complexes were increased in DpsbTc-1 (Supplemental Fig. S2, A and B), suggesting that assembly and/or stability of higher order PSII complexes are affected or that they could not be Figure 1. A, Strategy for inactivation of the chloroplast psbTc gene. The organization of the psbB operon with the genes psbB, psbTc, psbH, petB, petD, and psbN on the opposite strand is indicated. Arrows designate the direction of their transcription. The aadA selection cassette without termination signal and its 16S rDNA promoter (P) and ribosome-binding site (R) were inserted into the introduced HpaI restriction site (plastome position 76,694 bp) of the psbTc coding frame (plastome position 76,683-76,787 bp). B, PCR products of the psbTc region in the wild type (WT) and mutant indicate the homoplastomic state of the insertion. C, Northern analysis of the wild type and DpsbTc-1 using double-strand (petB and psbB) and single-strand (psbTc, psbH, and psbN) probes. Arrowheads indicate shifted psbTc-containing transcripts, confirming the homoplastomic state of the aadA insertion. Arrowheads 1 to 3 correspond to the up-shifted aadA containing primary spliced and unspliced transcripts in DpsbTc-1. Arrows 4 and 5 indicate the up-shifted psbB-psbTc-aadA and the appearing psbTc-aadA tran-script in the mutant, respectively. Ribosomal RNA (rRNA) was stained with methylene blue as a loading control. readily isolated in DpsbTc thylakoids. Therefore, a third approach was chosen to discern whether the formation or stability of PSII-LHCII supercomplexes and dimeric complexes is affected in DpsbTc-1. Leaves were labeled with [ 35 S]Met, and isolated thylakoids were analyzed by two-dimensional BN-SDS-PAGE (Fig. 2C). This approach showed that the mutant is capable of assembling RC47, PSII monomers, and dimers efficiently but not PSII-LHCII supercomplexes compared with the wild type. Since steady-state levels of dimers are severely reduced, we conclude that the stability of dimeric PSII complexes is affected by the mutation, presumably due to conformational changes induced by the absence of PsbTc. The relative amount of assembled PSII-LHCII complexes even after a short 5-min pulse resembles the pattern after a pulse of 40 min (Fig. 2C), indicating that the supercomplexes are not lost because of severe instability during the labeling period in DpsbTc-1. Improper assembly of LHCII may result secondarily from conformational changes of the PSII dimer (Fig. 2C).

Light Sensitivity: Photoinactivation and Recovery of PSII Activity
Upon excessive illumination, D1 protein is damaged and degraded, resulting in impaired PSII activity. Therefore, the extent of PSII light sensitivity and its ability to recover from photoinactivation were studied by measuring F v /F m as an estimation of the potential PSII quantum yield in wild-type and mutant plants. In order to distinguish between PSII damage and repair (de novo D1 synthesis) processes, the effect of the chloroplast protein synthesis inhibitor D-threochloramphenicol (CAP) was examined. For this, we chose saturating concentrations of CAP (200 mg mL 21 ) in darkness and incubation times (30 min) sufficient to completely block D1 synthesis (Supplemental Fig.  S1B). Wild-type and DpsbTc-1 to -3 leaf discs were illuminated at 1,200 mE m 22 s 21 after pretreatment with CAP. In the absence of CAP (control conditions), F v /F m was reduced to about 50% of the initial value in all three mutant lines compared with only 66% 6 1.2% in the wild type after 4 h of illumination ( Fig. 3A; Supplemental Fig. S1C). Addition of CAP induced a more pronounced PSII inactivation in both wild-type and DpsbTc-1 to -3 plants to 33% 6 1.3% and 20% to 23% of their initial values, respectively ( Fig. 3A; Supplemental Fig. S1C). This shows that recovery principally takes place and that PSII is more sensitive to high light exposure in the mutants than in wild-type plants.
To further analyze the efficiency of the recovery process, wild-type and DpsbTc-1 leaf disc samples were exposed to 1,200 mE m 22 s 21 until the PSII quantum Figure 2. A, Western-blot analysis of wild-type (WT) and DpsbTc-1 thylakoid proteins. The designation 100% corresponds to 5 mg of chlorophyll. B, Separation of wild-type and DpsbTc-1 thylakoid protein complexes by BN-SDS-PAGE. Complexes and silverstained proteins are labeled according to their identification by mass spectrometry. Comparable results were obtained from two further independently generated mutant plants (Supplemental Fig. S1A). D, Dimer; M, monomer; T, trimer. Note that PSII-LHCII supercomplexes are lacking and levels of PSII dimers are reduced. C, Autoradiograph of in vivo radiolabeled thylakoid membrane protein complexes separated by BN-PAGE and SDS-PAGE. Intact leaves labeled for 40 and 5 min showed impaired assembly of PSII-LHCII supercomplexes, whereas dimer assembly was not affected in DpsbTc-1. yield reached 0.17 in all samples and then were allowed to recover PSII activity under very low light intensity (3 mE m 22 s 21 ; Fig. 3B). After 2 h, recovery was about two times faster in the wild type compared with DpsbTc-1. Mutant leaf samples finally recovered 70.4% 6 8% of their initial PSII activity compared with 84.2% 6 7% in the wild type within 6 h, indicating a delayed recovery of PSII activity from photoinhibition in DpsbTc-1. To further analyze photosensitivity at the protein level, quantitative immunological analysis with D1 proteins of high light-treated wild-type and mutant plants was performed in the absence and presence of CAP (Fig. 3C). We calculated that D1 accumulates to about 85% in the wild type and to 70% in the mutant at 3 h after photoinhibition. After CAP and high light treatment, D1 was reduced to approximately 45% in the wild type and to 15% in the mutant. These data confirmed that the photoinhibitory effect in DpsbTc-1 was indeed due to D1 degradation and that CAP induced a more pronounced degradation of D1 in the mutant compared with the wild type, again indicating an increased light sensitivity of PSII in DpsbTc-1 (Fig. 3C).

Activities of PSII and PSI
The maximum quantum yield of PSII, F v /F m , was decreased in all independently selected and backcrossed mutant lines (DpsbTc-1 to -3) and ranged from 0.65 to 0.71 versus 0.82 6 0.01 in the wild type, which was due to an increased F o and/or a decreased F v , again indicating dissociation of the outer PSII antenna and/or malfunction of PSII electron transport activity. Further photosynthetic parameters have been measured in the three lines (Fig. 4). The photochemical quenching parameter (qP) as well as the effective PSII yield (F PSII ) were decreased at 5 and 45 mE m 22 s 21 red actinic light in DpsbTc-1 to -3 mutant plants (Fig. 4, A and C), indicating a partial loss of PSII function. Nonphotochemical quenching (NPQ) was also significantly lower at both light intensities, amounting to only about 50% in the mutants compared with the wild type at red light intensity of 45 mE m 22 s 21 , indicating a decreased transmembrane proton gradient due to lower PSII activity (Fig. 4B). The PSI oxidation state at 5 and 45 mE m 22 s 21 red light intensities was about 7-fold and 4-fold higher, respectively, in DpsbTc-1 to -3 compared with the wild type (Fig. 4D). Since NPQ is significantly decreased at both light intensities, the increased oxidation level of P700 can be solely attributed to a lower capacity of electron transfer from PSII to the plastoquinone pool in the mutants. This inference is supported by steady-state fluorescence measurements (F s #) elicited by 650-nm actinic light at an intensity of 34 mE m 22 s 21 on far-red background light (12 W m 22 ). The ratio F s # DpsbTc-1 to F s # wild type was 0.77, illustrating that the rate of electron flow toward plastoquinone is lower than that from plastoquinol to the electron sink via PSI in the mutant. These results are also supported by a lowered PSII activity ranging F v /F m ratios are expressed relative to the initial values before starting the experiment in the wild type (F v /F m approximately 0.82) and mutants (F v / F m approximately 0.68), respectively (see data for independent transformants in Supplemental Fig. S1C). B, PSII recovery from photoinhibition of wild-type and DpsbTc-1 mutant leaves exposed to photoinhibition at 1,200 mE m 22 s 21 until a F v /F m ratio of 0.17 was reached by each sample. PSII recovery was followed by further incubation of the samples at 3 mE m 22 s 21 for the times indicated. Total recovery after 6 h is depicted as the ratio F v /F m as a percentage of the original F v /F m in the wild type and mutant. Error bars in A and B represent SD of at least three independent experiments. C, Immunoblot analysis of the wild type and DpsbTc-1 with D1-specific antibodies before and 3 h after photoinhibition (3 h) at 1,200 mE m 22 s 21 with (+) and without (2) CAP. Five micrograms of chlorophyll was loaded, and equal loading was routinely checked either by immunoblot analysis with antibodies raised against cytochrome b 6 or by gel staining. from 65% to 76% in three mutant plants compared with wild-type thylakoids, measured as oxygen evolution under saturating light conditions in the presence of uncouplers.

Electron Flow within PSII as Measured by Thermoluminescence
Thermoluminescence (TL) measurements were performed to further investigate the effect of psbTc inactivation on electron flow within PSII. Activation energy required for back electron flow from Q B 2 to the S 2,3 states of the Mn 4 Ca donor side of PSII via P 680 + charge recombination is supplied by temperature rise and accompanied by light emission. The temperature at which recombination and light emission are maximal is related to the redox potential difference between the recombining charge-separated pairs and thus to the energy input required to drive the back electron flow (Krieger-Liszkay and Rutherford, 1998;Rutherford and Krieger-Liszkay, 2001). Charge recombination between Q B 2 and oxidized states of S 2 or S 3 occurs at about 35°C. The luminescence emission appearing at this temperature is designated the B band (Schwenkert et al., 2006;Ducruet et al., 2007). Blocking electron flow to the Q B site results in back electron flow from Q A 2 and charge recombination with the S 2/3 state, resulting in the Q band emission. Due to the lower energy gap between these recombining pairs, the Q band temperature occurs at about 3°C to 5°C. This is the case when phenolictype herbicides, such as Ioxynil, bind at the Q B site (Krieger-Liszkay and Rutherford, 1998;Cser and Vass, 2007;Umate et al., 2007). However, binding of urea-type herbicides such as 3-(3,4-dichlorphenyl)-1,1-dimethylurea (DCMU) affects the Q B site conformation and thus alters its interaction with the Q A site, resulting in a decrease of the Q A :Q A 2 /Q B 2 :Q B redox potential gap and thus an up-shift of the required energy input to drive back electron flow that occurs at 12°C to 15°C (Krieger-Liszkay and Rutherford, 1998;Rutherford and Krieger-Liszkay, 2001;Ohad et al., 2004).
The B band emission temperature was up-shifted to about 38°C to 42°C (Fig. 5A), while the Q band temperature emitted in the presence of DCMU was strongly down-shifted to about 3°C in mutant thylakoids (Fig. 5A). The TL emission temperature in the presence of Ioxynil also showed a severe down-shift of the Q band emission temperature to 210°C in DpsbTc-1 and DpsbTc-2 (Fig. 5B). Strikingly, the concentration of Ioxynil (5 mM) that is sufficient to saturate the Q B site in wild-type thylakoids, resulting only in emission of the Q band at 3°C, does not completely abolish the B band emission in the mutant, which shows an increase in the Q band and a decrease of the B band emission intensity with increasing Ioxynil concentration up to 20 mM (Fig.  5B). These results demonstrate that the absence of PsbTc induces significant conformational changes with respect to the Q A and Q B site properties.
The oxidation/reduction of the Mn 4 Ca cluster usually occurs with a period of four stable steps (S 0 -S 3 ) during oxygen evolution (Haumann et al., 2005), while that of the Q B site occurs in two steps (Q B and Q B 2 ). Following transition from light to darkness, back electron flow from the reduced electron carriers of PSII (pheophytin 2 , Q A 2 , and Q B 2 ) occurs, resulting in the accumulation of about 25% S 0 and 75% S 1 states and a 1:1 ratio of Q B 2 /Q B (Rutherford et al., 1982(Rutherford et al., , 1984Figure 4. Photosynthetic performance in the wild type (WT) and the three mutant lines DpsbTc-1 to -3. Photosynthetic parameters, as indicated by qP (A), NPQ (B), F PSII values (C), and P700 oxidation ratio measured as DA/DA max (D), were calculated from wild-type and mutant plants kept in the steady state at 5 and 45 mE m 22 s 21 red light intensities. Error bars represent SD of at least three independent experiments.
Krieger-Liszkay and Rutherford, 1998;Umate et al., 2007). The intensity of TL B band emission of darkadapted tobacco thylakoids, following single-turnover excitations (one to six flashes), is proportional to the number of recombining Q B 2 /S 2,3 pairs and oscillates with the number of exciting single-turnover flashes with maximal emission at flashes 2 and 6 in the wild type (Umate et al., 2007). Since deletion of PsbTc shows drastic effects on the Q A /Q B site interactions and charge recombination, one would expect changes also in the oscillation pattern of the B band emission with the number of excitation flashes. Results of experiments addressing this question indeed showed an alteration of the TL B band oscillation patterns for DpsbTc-1 and DpsbTc-2, exhibiting emission maxima at flashes 1 and 5 (Fig. 5C). Using a simulation program for oscillation patterns (Umate et al., 2007), the ratio Q B /Q B 2 present in dark-adapted thylakoids prior to the flash excitations was significantly increased in DpsbTc-1 (1.68 versus 0.90 in the wild type; Supplemental Table S1). These results support the conclusion that the back electron flow within PSII is altered in the mutant, presumably because of alterations of the Q B binding site properties.

Effect of the Mutation on Redox-Controlled Phosphorylation of LHCII and PSII Core Proteins
Plants have developed several mechanisms to efficiently adjust their energy conversion to different light qualities and quantities. Shifting their antenna from one photosystem to the other (state transition) allows optimal adaptation to changing light conditions. Thus, phosphorylation and dephosphorylation of a mobile pool of CP29, CP26, and LHCII are induced under light, which preferentially excites PSII (state II) and PSI (state I), respectively (Rochaix, 2007). Regulation of phosphorylation of the D1, D2, and CP43 proteins is less understood. Therefore, phosphorylation and antenna association experiments were performed (Fig. 6,  A and B). Strikingly, in the absence of PsbTc, the PSII core proteins D1/D2 were not phosphorylated and CP43 was only marginally phosphorylated under both state I and state II conditions in all three mutant lines ( Fig. 6A; Supplemental Fig. S1D). On the other hand, in contrast to the wild type, LHCII was already highly phosphorylated in darkness and dephosphorylated under state II light conditions in DpsbTc-1 to -3. To check whether phosphorylation in the mutants results from a reduced plastoquinone pool in darkness, state I conditions were adjusted with far-red light. This light treatment abolished LHCII phosphorylation in both wild-type and mutant plants (Fig. 6A). Mild solubilization of thylakoids with digitonin and separation of lysates by BN-PAGE allows the detection of LHCII-PSI supercomplexes under state II light conditions in the wild type . In contrast, in the mutant, LHCII-PSI supercomplexes were formed only in darkness, consistent with the phosphorylation pattern and the reduced state of the plastoquinone pool Figure 5. TL measurements of wild-type (WT) and DpsbTc-1 thylakoids. A, The B band occurring at 35°C in the wild type is slightly up-shifted to 38°C in the mutant. The Q band emission in the presence of DCMU (10 mM) occurs at 15°C in the wild type and at 3°C in DpsbTc. BL indicates the baseline signal generated from dark-adapted, unexcited samples. B, Q band emission of wild-type (a = 5 mM Ioxynil) and DpsbTc-1 (b = 5 mM, c = 10 mM, and d = 20 mM Ioxynil) thylakoid samples. The Q band emission temperature is down-shifted to 210°C at all Ioxynil concentrations compared with 3°C to 5°C in the wild type. Note that curves b, c, and d depict a persistent B band emission in the presence of Ioxynil. BL, Baseline. C, The B band oscillation pattern shows a 1/5 oscillation type in DpsbTc-1 compared with the typical 2/6 type for the wild type. All SD values are below 0.5%. Comparable results were obtained with DpsbTc-2 plants. (Fig. 6B, top). The composition of the bands representing the LHCII-PSI supercomplex was also demonstrated by immunoblot analysis of the native gel using antibodies raised against LHCII and PsaA/B (Fig. 6B, bottom). These data indicate a role of the PSII protein PsbTc in the oxidation of reduced plastoquinone in the dark.

The Q B Binding Site Is Required for Oxidation of PQH 2 in the Dark
Reduction of plastoquinone takes place via several nonphotochemical pathways, and reoxidation often depends on oxygen (for review, see Rumeau et al., 2007). A prominent functional effect of the mutation seems to be an alteration in the role of the Q B binding site important for efficient forward electron flow in the light and presumably reoxidation of the plastoquinone pool, which is nonphotochemically reduced in darkness. Were this true, specific inhibition of the Q B site with DCMU or Ioxynil in the wild type should prevent plastoquinol oxidation even in the presence of oxygen in the dark. In order to check this assumption, wildtype leaves were first treated with far-red light for 1 h to allow complete oxidation of plastoquinol and de-phosphorylation of LHCII (Fig. 6C). Subsequently, leaves were incubated in the dark overnight under aerobic and anaerobic conditions. In the presence of oxygen, LHCII remained in a dephosphorylated state, indicating that plastoquinone remains oxidized. In contrast, in the absence of oxygen, phosphorylation of both LHCII and CP29 was increased, indicative of a reduced plastoquinone pool. DCMU or Ioxynil was added 1 h after far-red light treatment, and leaves were maintained in strict darkness. Indeed, using both inhibitors, LHCII and CP29 were highly phosphorylated after 12 h of dark incubation even under aerobic conditions (Fig. 6C). Since binding of Ioxynil does not change the Q B site conformation, it is unlikely that phosphorylation of antenna proteins under aerobic conditions is due to structural changes within PSII, allowing easier access of the corresponding kinase(s) to the target. Unlike in DpsbTc-1, phosphorylation of the PSII proteins CP43, D1, and D2 in the wild type was not affected by treatment with inhibitors. Therefore, we conclude that alterations in the redox regulation of LHC phosphorylation resulted from changes in the Q B binding site, which is directly and specifically involved in the oxidation of PQH 2 in an oxygendependent manner. Accordingly, our data indicate that the presence of PsbTc is directly or indirectly required to keep the Q B site in proper conformation for the reoxidation of PQH 2 .

DISCUSSION
Effect of PsbTc Deletion on Assembly, Stability, and Structure of PSII Studies on PsbTc deletion mutants were reported from Synechocystis 6803 (Iwai et al., 2004) and two Thermosynechococcus species (Henmi et al., 2008) representing eubacterial organisms and from Chlamydomonas reinhardtii (Monod et al., 1994;Ohnishi and Takahashi, 2001;Ohnishi et al., 2007). Data obtained with Thermosynechococcus suggested that PsbTc is involved in the assembly of dimeric PSII complexes. However, since deletion of psbTc in the Chlamydomonas mutant resulted only in a weakly impaired PSII, it was concluded that this protein is not essential for the organization and function of PSII. Solubilization studies of proteins labeled in vivo indicate that PsbTc in tobacco is not required for the assembly but rather for the stability of the dimeric PSII core (Fig. 2, B and C).
The absence of PSII-LHCII assemblies in DpsbTc-1 to -3 is intriguing, since PsbTc has been localized to the region interconnecting PSII monomers next to the central 2-fold axis (Ferreira et al., 2004;Loll et al., 2005). From its outlined function and position, PsbTc could be involved in the attachment of the outer antenna, possibly through interaction with adjacent proteins, like CP47; therefore, it exerts a long-distance effect.
The proposed positions of PsbTc, PsbL, and PsbM close to the central axes of the PSII dimer previously  Fig. S1C). B, Analysis of protein complexes separated by BN-PAGE in the first dimension (top). Immunoblot analysis of protein complexes separated by BN-PAGE using LHCb1 and PsaA/B antibodies demonstrates the presence of LHCII-PSI supercomplexes (indicated by arrows and boxes) in the wild type under state II conditions and in the mutant after dark adaptation (bottom two panels). C, LHCII phosphorylation after dark incubation of wild-type leaves under aerobic (+O 2 ) and anaerobic (2O 2 ) conditions in the presence and absence of DCMU and Ioxynil (IOX).
suggested that these LMW peptides are involved in dimerization (Ferreira et al., 2004;Nield and Barber, 2006). However, PsbI has been shown to represent a component required for the formation of the dimer presumably by forming a "bracket" (Schwenkert et al., 2006), whereas apparently neither PsbTc nor PsbM is essential for PSII dimer assembly. Loss of PsbTc may affect the function and/or conformation of other adjacent proteins, like PsbL. However, DpsbL knockouts are clearly distinct from DpsbTc plants with respect to photoautotrophic growth, assembly of PSII dimers, photosynthetic performance, and levels of PSII proteins, indicating that neither function nor localization of PsbL is impaired in DpsbTc knockout plants (Ohad et al., 2004). We cannot rule out that loss of PsbTc causes the removal of PsbM, since the corresponding mutant also shows a 1/5 B band oscillation type and increased LHCII phosphorylation in the dark (Umate et al., 2007). However, phosphorylation of CP43 and recovery of PSII from photoinhibition are not affected in DpsbM, indicating that major deficiencies are due to DpsbTc inactivation. It is intriguing that specific deletions of the predictably closely located proteins PsbTc, PsbL, and PsbM result in quite diverse phenotypes in higher plants (Ohad et al., 2004;Umate et al., 2007).

Lack of PsbTc Results in Increased D1 Degradation
Increased rates of PSII-mediated plastoquinone reduction exceeding plastoquinol oxidation result in an increased back electron flow and charge recombination, causing damage to PSII core proteins (Keren et al., 1997;Vass et al., 2007). Only under strong illumination were PSII activity and D1 levels reduced in the Chlamydomonas mutant. In contrast to the data obtained with tobacco mutants, photosensitivity and D1 degradation in the presence of CAP were apparently not changed in the Chlamydomonas mutant, suggesting that PsbTc may not be relevant for photoprotection but may be required for an efficient recovery of photodamaged PSII in the algae (Ohnishi and Takahashi, 2001). Furthermore, PsbTc appeared to be involved in the activation of forward Q A electron flow during repair of PSII from photoinhibition in C. reinhardtii (Ohnishi et al., 2007). Photosensitivity is more striking with translation inhibitors in DpsbTc tobacco plants than without. The difference of photoinhibition with and without inhibitors as an indicator of recovery is almost the same in wild-type and mutant plants. This indicates that mutant plants are capable of repair from photoinhibition, although this is delayed compared with that in the wild type, suggesting a similar function of PsbTc in Q A activation during the repair process in tobacco.
Although the function of PsbTc in different organisms is similar, some species-specific features may reflect adaptations to their ecophysiological niches. Yet, it is conceivable that light sensitivity in tobacco might represent a secondary effect of the mutation, since facilitated back electron flow results in increased rates of charge recombination and related oxidative damage in DpsbTc.

Effect of PsbTc Deletion on Q A and Q B Site Properties and PSII Electron Flow
The conformational changes in the organization of PSII induced by the loss of PsbTc affect Q A and Q B site properties and/or interaction, resulting in an increased ratio Q B /Q B 2 and alterations in the PSII charge recombination, as judged from TL measurements. This indicates that there are more oxidized Q B sites than semiquinones in the mutant, causing an impaired back electron flow and reduction of the plastoquinone pool in dark-adapted PSII populations. However, it still remains controversial which plastoquinone-binding site is largely or directly affected by the absence of PsbTc, because both peaks of Q and B bands in transformant plants were shifted compared with those of wild-type plants. Due to the closer vicinity of PsbTc to Q A , it is conceivable that conformational changes at the Q A site may affect its interaction with the Q B site. Therefore, the effect on the Q B site could represent a secondary effect of the mutation, as has also been implied for the long-distance effect on the assembly of PSII-LHCII supercomplexes and the stability of PSII dimers in DpsbTc.
The increase in the B band emission temperature and the down-shift of the Q band also imply an increase in the redox potential gap between the Q A / Q B sites and a decrease of the pheophytin/Q A redox potential gap, thus lowering the forward electron flow and promoting charge recombination by back electron flow from Q A . This is supported by a decreased PSIIdependent electron transport capability, measured as oxygen evolution rate, in DpsbTc to about 72% of wildtype levels.

PsbTc Deletion Alters the Redox State of the Plastoquinone Pool and Phosphorylation of PSII Proteins
It is well established that the association of LHCII, CP26, and CP29 with PSII and partially with PSI (state I and state II, respectively) is regulated via reversible phosphorylation of the chlorophyll a/b-binding apoproteins (Kargul and Barber, 2008). Activation of the respective kinase(s) is regulated by the interaction of reduced plastoquinone with the oxidation site of the cytochrome b 6 f complex (Vener et al., 1997). The orthologous protein kinases, STT7 and STN7 in Chlamydomonas and Arabidopsis (Arabidopsis thaliana), respectively, are involved in the regulation of phosphorylation of LHCII, CP26, and CP29, although the primary targets of the kinases still remain unknown (Rochaix, 2007).
In this study, it could be shown that accumulation of reduced plastoquinone leads to phosphorylation of LHCII in darkness in DpsbTc mutant plants. Moreover, BN-PAGE analysis revealed that LHCII is not only phosphorylated but also attached to PSI in dark-adapted mutant plants. Apparently, no light-induced signals cause migration of the LHCII antenna to PSI in the mutant. Therefore, the data uncover a new level of control of LHCII phosphorylation in darkness.
Although the related protein kinase, STN8, triggers phosphorylation of all three PSII core proteins D1, D2, and CP43 (for review, see Vener, 2006), phosphorylation of D1/D2 proteins is considerably more responsive than that of CP43 to redox regulation in tobacco wild-type plants ( Fig. 6A; Umate et al., 2007), indicating that additional kinases could be involved in PSII core protein phosphorylation. Phosphorylation of outer antenna proteins and presumably of PSII core proteins is affected not only by activation/deactivation of the respective protein kinase(s) but also by the exposure of segments of substrates containing phosphorylation sites at thylakoid membrane surfaces and, thus, on their accessibility to the active site of the respective protein kinase(s) (Zer et al., 1999Vink et al., 2000). Phosphorylation of the core proteins CP43, D1, and D2 is almost completely abolished in DpsbTc under all light conditions. The same has been described for DpsbI in tobacco (Schwenkert et al., 2006). As both mutants lack stable PSII-LHCII supercomplexes, it is likely that accessibility for the kinase(s) that phosphorylate PSII core proteins is blocked by structural changes of the PSII dimer in the mutants. In contrast, alterations of phosphorylation patterns of PSII antenna proteins are likely due to physiological consequences rather than structural changes. The effect of PsbTc knockouts on the phosphorylation of PSII core and antenna proteins is unknown from other organisms, and it would be challenging to check whether the lack of PsbTc in Chlamydomonas induces similar effects.
Nonphotochemical reduction and reoxidation of plastoquinone takes place via several pathways (for review, see Rumeau et al., 2007). The plastid terminal plastoquinone oxidase has been suggested to be involved in reoxidation of reduced plastoquinone. Alternatively, reoxidation of the PQH 2 pool may be mediated by charge recombination with Q B and probably subsequently with cytochrome b 559 (Bondarava et al., 2003). Impaired reoxidation of plastoquinol could account for LHCII phosphorylation in darkness in DpsbTc plants. Finally, we suggest that PsbTc stabilizes the Q B binding site in vivo that is essential for oxidation of reduced plastoquinone in darkness in an oxygen-dependent manner, possibly to keep the PSII acceptor side oxidized. Our data indicate that the PSIIdependent oxidation of PQH 2 in darkness predominantly controls the redox state of the plastoquinone pool. Therefore, this effect is prevalent when compared with other activities known to oxidize plastoquinol, like the plastid terminal plastoquinone oxidase localized in stroma lamellae (Rumeau et al., 2007), or with a nonenzymatic PQH 2 oxidation. We postulate that PSII could represent the hitherto unidentified major terminal oxidase involved in chlororespiration in leaves (Nixon, 2000;Bennoun, 2002). A protective role of PSII-mediated oxygen reduction and plastoquinol oxidation in the light-driven electron transport, to avoid overreduction of plastoquinone and harmful charge recombination leading to damage of PSII, remains to be shown.

Preparation of Thylakoid Membranes, SDS-PAGE, and Immunoblot Analysis
Thylakoid membrane proteins were isolated, separated by SDS-PAGE, and subjected to immunoblot analysis as described (Umate et al., 2007). Quantitative immunoblot analysis was performed using AIDA software (Raytest).

BN-PAGE
BN-PAGE was performed as described earlier  1.5% digitonin. Lysates were separated on a 4% to 12% polyacrylamide gel. Proteins were either transferred onto polyvinylidene difluoride membranes or gel lanes were excised and run on a second dimension using 15% SDS-PAGE. In vivo labeling was performed using [ 35 S]Met (Schwenkert et al., 2006).

Thylakoid Protein Phosphorylation Assay
State I conditions were adjusted by either keeping plants in darkness for 24 h or treatment with light at 728 nm (PSI light) of 20 mE m 22 s 21 for 30 min. For state II conditions, leaves were illuminated with light at 650 nm (PSII light) of 45 mE m 22 s 21 for 30 min. All buffers used for thylakoid preparations contained 10 mM NaF. Phosphorylated proteins were detected using antiphosphothreonine antiserum (New England Biolabs). Anaerobic conditions were generated by placing leaves in a humid screw-cap vial flushed for 10 min with pure argon at room temperature.

PSII Quantum Yield and Fluorescence Quenching Analysis
A pulse amplitude-modulated fluorimeter (PAM-101; Waltz) was used to study chlorophyll a fluorescence kinetics. Leaves dark adapted for not less than 5 min were used for measurements. Red actinic light (650 nm, 5 and 45 mE m 22 s 21 ) was used for measurement of fluorescence parameters. The maximum quantum yield of PSII was determined as (F m 2 F o )/F m = F v /F m , and the effective quantum yield of PSII, F PSII , was expressed as F m # 2 F s #/F m # (Genty et al., 1989). qP and NPQ were determined by repetitive saturation pulses. The quenching coefficients, NPQ and qP, were calculated as (F m 2 F m #)/F m # and (F m # 2 F)/(F m # 2 F o ), respectively (van Kooten and Snel, 1990).

PSI Activity
PSI activity was investigated on leaves as absorption changes of PSI at 830 nm induced by far-red light (DA max ) and actinic light (DA) using the PSI attachment of PAM-101 (Walz). The redox condition of PSI at the indicated light intensities was expressed as the ratio DA/A max (Klughammer and Schreiber, 1994).

Oxygen Evolution Measurements
PSII electron transport activity was determined using a Clark-type oxygen electrode as described (Umate et al., 2007). In order to measure exclusively PSII-dependent electron transport, the PSII-specific electron acceptor p-benzoquinone as well as the uncouplers NH 4 Cl (5 mM) and gramicidin (3 mM) were used under saturating light conditions. PSII activity of wild-type thylakoids was 198 6 19 mmol oxygen mg 21 chlorophyll h 21 .

TL Measurements
TL measurements of thylakoids were measured using a home-built apparatus as described (Schwenkert et al., 2006). The Q band emission resulting from Q A 2 /S 2 charge recombination was measured by blocking electron flow to Q B with the addition of DCMU (10 mM) or Ioxynil (5, 10, and 20 mM) during dark adaptation. Concentrations of Ioxynil higher than 20 mM were avoided due to high fluorescence quenching induced by this herbicide. A computer-based simulation program allowing the prediction of S-state ratio and the occupancy ratio Q B /Q B 2 was employed. The program simulates predicted oscillation profiles and checks for the correlation between simulated and measured values (Umate et al., 2007).

PSII Photoinactivation and Recovery Kinetics
The sensitivity of PSII to oxidative stress was determined using leaf discs of wild-type and DpsbTc plants exposed to 1,200 mE m 22 s 21 heterochromatic light. Photoinactivation of PSII was measured as changes in the F v /F m parameter as a function of exposure time. To estimate the contribution of the PSII recovery process during treatment with high light, leaf discs were infiltrated with a solution of CAP (200 mg mL 21 ) in darkness for 30 min prior to exposure to high light. To assess photoinhibition kinetics and capacity to recover PSII activity, leaf discs were exposed in the absence of CAP pretreatment to high light (1,200 mE m 22 s 21 ) until an F v /F m of 0.17 was reached in wild-type and DpsbTc discs and then allowed to recover PSII activity by lowering the light intensity to 3 mE m 22 s 21 for up to 6 h, measuring the F v /F m level every 1 h.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Studies on the assembly, photoinhibition, and phosphorylation patterns of the independent transformants DpsbTc-2 and DpsbTc-3.
Supplemental Table S1. Computer-based simulation for the prediction of the S-state ratios and the occupancy ratio Q B /Q B 2 .