Deletion of PsbM in Tobacco Alters the QB Site Properties and the Electron Flow within Photosystem II*

Photosystem II, the oxygen-evolving complex of photosynthetic organisms, includes an intriguingly large number of low molecular weight polypeptides, including PsbM. Here we describe the first knock-out of psbM using a transplastomic, reverse genetics approach in a higher plant. Homoplastomic ΔpsbM plants exhibit photoautotrophic growth. Biochemical, biophysical, and immunological analyses demonstrate that PsbM is not required for biogenesis of higher order photosystem II complexes. However, photosystem II is highly light-sensitive, and its activity is significantly decreased in ΔpsbM, whereas kinetics of plastid protein synthesis, reassembly of photosystem II, and recovery of its activity are comparable with the wild type. Unlike wild type, phosphorylation of the reaction center proteins D1 and D2 is severely reduced, whereas the redox-controlled phosphorylation of photosystem II light-harvesting complex is reversely regulated in ΔpsbM plants because of accumulation of reduced plastoquinone in the dark and a limited photosystem II-mediated electron transport in the light. Charge recombination in ΔpsbM measured by thermoluminescence oscillations significantly differs from the 2/6 patterns in the wild type. A simulation program of thermoluminescence oscillations indicates a higher QB/Q –B ratio in dark-adapted mutant thylakoids relative to the wild type. The interaction of the QA/QB sites estimated by shifts in the maximal thermoluminescence emission temperature of the Q band, induced by binding of different herbicides to the QB site, is changed indicating alteration of the activation energy for back electron flow. We conclude that PsbM is primarily involved in the interaction of the redox components important for the electron flow within, outward, and backward to photosystem II.

Photosystem II (PSII), 5 a supramolecular pigment-protein complex of photosynthetic organisms, utilizes absorbed light energy to oxidize water, releasing dioxygen and electrons that serve as the major source of reducing power in photosynthetic activity. The mechanisms of this process have been studied extensively. Based on biochemical (reviewed in Ref. 1), biophysical (2,3), and structural analysis, including electron microscopy (4,5) and x-ray diffraction (6 -9), comprehensive knowledge of the basic steps of the water oxidation mechanism, electron transfer reactions, and organization of the individual components involved in these processes has been delineated. Information is also available on the structure of LHCII and the energy transfer to the inner antenna and the photochemical reaction center (10,11), as well as on the components involved in the biogenesis of PSII (12)(13)(14). Most if not all genes encoding PSII subunits have been identified in cyanobacteria and higher plants, and their function, expression, and regulation have been studied extensively (15)(16)(17).
PSII is the most intricate assembly of thylakoid membrane systems consisting of more than 30 subunits. It assembles as a dimer together with the minor light-harvesting antenna CP24, CP26, and CP29 for each monomer and is further surrounded by trimeric LHCII protein complexes (5). Based on the similarity of subunit sequences, composition, and activity of PSII, it is generally accepted that the structure of the PSII core in eukaryotes is basically similar to that of cyanobacteria, for which x-ray diffraction structures between 3.8 and 3.0 Å resolution have been obtained (8,9). However, despite sustained attempts to obtain a higher structural resolution for the PSII of higher plants (10), the heterogeneity of the photochemical center of PSII caused by light-induced changes has so far prevented the formation of crystals allowing a higher resolution of the complex.
One of the most intriguing features of the PSII core is the presence of 16 bitopic, intrinsic, or peripheral low molecular weight proteins. Knowledge about their roles in the overall pho-tosynthetic process is still fragmentary. In eukaryotic PSII, 11 of them are encoded by plastid chromosomes, notably PsbE, -F, -H, -I, -J, -K, -L, -M, -N, -Tc, and -Z (reviewed in Refs. 5 and 16 -19). The fact that most low molecular weight subunits (LMWs) of PSII have been highly conserved throughout the evolution implies that they perform essential functions, as indeed this has been established for PsbI, PsbT, the ␣ and ␤ subunits of the two-chain cytochrome b 559 , PsbE and PsbF, respectively, as well as for PsbL and PsbJ, which fulfill crucial structural and functional roles (20 -28). However, only limited information is available on the function and exact position of the other LMWs within the PSII complex. Moreover, studies on distinct LMWs from various organisms have indicated different roles (reviewed in Refs. 16,17) or control of the same function, albeit to a different degree, as is the case for PsbJ in Synechocystis sp. PCC 6803 and tobacco (23). Apparently, the management of PSII electron flow in terms of energy dissipation and avoiding generation of oxygen radicals has to cope with different requirements in various organisms. Such differences may depend on the eco-physiological conditions under which different organisms thrive and the divergence of the outer antenna. It is therefore conceivable that the properties of the LMWs of PSII vary in different organisms.
The study of all chloroplast-encoded LMWs in one organism offers a better chance to understand the roles for each one of them in the same assembly. Therefore, we have inactivated plastid genes encoding LMWs of PSII using a transplastomic approach in tobacco, and we reported on the roles of six of them, including PsbE, PsbF, PsbL, PsbJ, PsbI, and PsbZ (23)(24)(25)(26)(27)(28).
The PsbM polypeptide has been detected in PSII complexes isolated from Chlamydomonas reinhardtii (29), Synechocystis sp. PCC 6803 (30), and Synechococcus vulcanus (31), and its presence in Synechocystis sp. PCC 6803 (32) and pea (33) has been confirmed recently using proteomics approaches. However, mutation studies of PsbM have not yet been reported from any organism. Moreover, the exact position of PsbM within the PSII assembly as well as its function remains to be established (8,17).
The analyses of the first PsbM knock-out presented here demonstrate that the biogenesis of PSII is not significantly altered in the absence of PsbM. However, properties of the Q B site and its interaction with Q A and charge recombination within PSII are specifically impaired in ⌬psbM resulting in a shift of thermoluminescence (TL) B band oscillations, a decreased rate of oxygen evolution and forward electron flow, and thus an increased light-induced photoinactivation as well as dephosphorylation of LHCII. Because of a high plastoquinol (PQH 2 ) content in dark-adapted mutant plants, levels of LHCII phosphorylation are significantly elevated as compared with the wild type. Phosphorylation of the reaction center proteins D1/D2 is faintly detectable presumably because of conformational changes induced by loss of PsbM proteins.

EXPERIMENTAL PROCEDURES
Clone Construction to Inactivate the psbM Gene in Tobacco Chloroplasts-The recombinant plasmid B20 (tobacco plastome clone bank) (34) containing a 17,235-bp insertion (nucleotide positions 26,426 in the plastid chromosome, accession number Z00044) in the vector pBR322 was digested with BamHI, and the resulting 2,096-bp fragment containing psbM was subcloned into the singular BamHI restriction site of pBluescript II KS Ϫ (Stratagene Inc., La Jolla, CA). Inactivation of the psbM gene (nucleotide position 30,861 (N) to 30,757 (C) bp, accession number Z00044) was achieved by insertion of the aadA cassette, including a terminator signal (35) at a unique BsgI restriction site in the N-terminal part of the gene (nucleotide position 30,844). Transformation of Nicotiana tabacum cv. Petit Havanna, the selection procedure, and in vitro propagation of transformants were carried out essentially as described (25). Isolation of plastid chromosomes by orthogonal pulsefield gel electrophoresis was carried out as described (36). Tobacco lines carrying the aadA cassette in a neutral insertion site and referred to as RV plants were used as wild type (WT) control plants (25).
Preparation and Handling of Thylakoid Membranes-Thylakoid membranes isolated from 4-week-old plants grown under greenhouse conditions were chosen for immunoblot analysis, phosphorylation experiments, and analysis of their composition by the blue native gel (BN) method as described (25). For separation of photosynthetic chlorophyll-protein complexes by sucrose gradient centrifugation, 4-week-old in vitro grown (10 -20 E m Ϫ2 s Ϫ1 light intensity; 12-h photoperiod) plants were used.
Blue native-PAGE (BN-PAGE) was performed as described earlier with modifications (25). The appearing spots were sequenced by mass spectrometry and assigned accordingly (37).
For TL measurements, thylakoids were prepared by grinding a few leaves in a buffer containing 20 mM Tris-HCl, pH 7.4, 5 mM MgCl 2 , 20 mM NaCl, and 100 mM sorbitol. The material homogenized at 0°C was filtered through nylon micromeshes and used immediately for measurements.
Chlorophyll a Fluorescence Induction Kinetics-Chlorophyll a fluorescence induction kinetics was measured using a pulse amplitude-modulated fluorimeter (PAM-101, Waltz, Effeltrich, Germany) (38). Prior to measurements, leaves were dark-adapted for 5 min. The potential maximum quantum yield of PSII was measured as (F m Ϫ F o )/F m ϭ F v /F m . Red actinic light (650 nm, 20 and 250 E m Ϫ2 s Ϫ1 ) was used for measurements of fluorescence quenching. Photochemical (qP) and nonphotochemical (NPQ) quenching were determined by repetitive saturation pulses. The quenching coefficients, NPQ and qP, were calculated as (F m Ϫ F m Ј)/F m Ј and (F m Ј Ϫ F)/(F m Ј Ϫ F o ), respectively (38).
State Transition and Thylakoid Protein Phosphorylation-State transition in intact leaves was calculated using the PAM-101 fluorimeter as (F m Ј Ϫ F m Љ)/F m Ј (39). Protein phosphoryla-tion was carried out using isolated thylakoids as described (25,40). All buffers used during thylakoid preparation contained 10 mM NaF. Detection of the phosphorylation level of thylakoid membrane proteins was carried out by immunoblotting using anti-phosphothreonine antibodies (New England Biolabs) as described earlier (25).
Low Temperature Fluorescence Measurements-Thylakoid suspensions (20 g of chlorophyll/ml) of WT and mutant leaves were frozen by immersing a liquid nitrogen-cooled glass rod (4 mm diameter) into the thylakoid suspension and rapidly returning it to the Dewar vessel of the sample holder of the fluorimeter filled with liquid nitrogen (Fluoromax-3, Horiba Jobin-Yvon, France). Fluorescence emission spectra were recorded using 430 nm excitation and 1.5 nm slits for both excitation and emission monochromators.
Photosystem I (PSI) Redox State-The redox state of PSI was measured on leaves using the PSI attachment of PAM101 (Walz, Effeltrich, Germany). The oxidation status of PSI at the light intensities indicated was expressed as the ratio ⌬A/⌬A max (41).
Thermoluminescence Measurements-TL of thylakoid suspensions was measured using a home-built apparatus as described (42). Samples (400 l) of 40 g of chlorophyll/ml were placed on the TL stage, dark-adapted at 20°C for 3 min, and rapidly frozen to Ϫ20°C by a stream of liquid nitrogen. The sample was then excited by saturating flashes delivered by a xenon arc discharge lamp (0.05 microfarad capacitor, charged at 1000 V, 3 s at 70% light emission; EG and G), then heated at a rate of 0.6°C s Ϫ1 , and photons were counted. For measurements of the intensity of the B band emission oscillations (Q B Ϫ / S 2 ,S 3 charge recombination) as a function of the numbers of single turnover excitations, a train of flashes (1-6 flashes, about 300-ms interval between flashes) was given at 0°C followed by rapid freezing.
For detecting the TL signal resulting from Q A Ϫ /S 2 recombination (Q band), the herbicides DCMU or ioxynil (Serva, Heidelberg, Germany), both binding to the Q B site and thus preventing Q A Ϫ oxidation, were added at concentrations as indicated before dark adaptation. Concentrations of ioxynil higher than 20 M were avoided because of the high fluorescence quenching induced by this herbicide. Glycerol (25% v/v) was added to the samples to avoid distortion of the linear heating rate during the ice-melting stage. The kinetics of the Q B Ϫ decay in darkness was measured following excitation of the dark-adapted sample by a single flash at 20°C, followed by further incubation in darkness for the indicated periods, and followed by rapid freezing to Ϫ20°C and starting the heating of the sample and photon-counting process.
A computer-based simulation program allowing the prediction of the S-states ratio and the occupancy ratio Q B /Q B Ϫ was employed (supplemental S1). The program simulates predicted oscillation profiles and checks for the correlation between simulated and measured values. The free parameters in the simulation are as follows: 1) the S-states and Q B /Q B Ϫ occupancy levels in the dark adapted state, which can vary between 0 and 1 in 0.1 steps; 2) the misses, i.e. the fraction of reaction centers that is not excited by a flash, which may vary between 0 and 0.5; and 3) the double hits, i.e. the fraction of reaction centers that are excited twice by a single flash, which may vary between 0 and 0.5. The value ⌺ 1-6 ͉intensity measured Ϫ intensity simulated ͉, where 1-6 represents the number of flashes, was used to estimate the correlation between measured and simulated results. The program used is given in the supplemental S1. Minimization of the cost function was performed by a simulated annealing algorithm (43).
Oxygen Evolution Measurements-Photosynthetic electron flow was determined using thylakoids isolated as described for TL measurements. The PSII specific electron acceptor p-benzoquinone was used under saturating light conditions using a Clark-type oxygen electrode for measuring oxygen evolution.
Measurements of Photoinhibition and Recovery Process-The sensitivity of PSII to oxidative stress has been determined with leaf disks (10 mm diameter, 5 disks per sample) of WT and ⌬psbM plants exposed to 500 E m Ϫ2 s Ϫ1 heterochromatic light. The 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 disks were infiltrated with a solution of D-threo-chloramphenicol (200 g ml Ϫ1 ) in darkness for 30 min prior to the exposure to high light. As a control, leaf disks were incubated in water. To assess the inhibition kinetics and the capacity to recover PSII activity, leaf disks were exposed to 1,500 E m Ϫ2 s Ϫ1 until an F v /F m of 0.17 was reached in both WT and ⌬psbM. The recovery was followed in low light (3 E m Ϫ2 s Ϫ1 ) for up to 6 h measuring the F v /F m level every 1 h.

RESULTS
Inactivation of psbM-Nine independent transformants with an identical phenotype were initially obtained, and three lines were used for further studies. Their homoplastomic status was confirmed by sequencing of the insertion site and by PCR analysis using isolated plastid chromosomes as template (Fig. 1A). Northern analysis with a strand-specific probe containing the coding region of the psbM gene demonstrated that not even traces of psbM transcripts were detectable confirming the homoplastomic state of the mutant (Fig. 1B).
Levels and Compositions of Thylakoid Membrane Complexes in ⌬psbM Resemble Those of the WT-Inactivation of psbM in tobacco caused a quite distinct phenotype. Different from several other LMWs mutants of the chloroplast, such as ⌬psbE, ⌬psbF, ⌬psbL, and ⌬psbJ, ⌬psbM plants are capable of photoautotrophic growth on soil. However, mutant leaves appeared bleached if the light intensity exceeded ϳ200 E m Ϫ2 s Ϫ1 , thus indicating increased light sensitivity of the photosynthetic apparatus. To elucidate the function of PsbM, homoplastomic ⌬psbM mutants in tobacco were analyzed by biochemical and biophysical approaches.
The relative amounts and sizes of pigment-containing thylakoid membrane complexes in sucrose gradients of ⌬psbM did not differ significantly from those of the WT (Fig. 2A). Only PSII-LHCII supercomplexes were faintly diminished, and consequently trimeric LHCII antennae complexes showed a slight increase in ⌬psbM. Moreover, a significant increase in the LHCII-CP24-CP29 complex was observed in the mutant compared with WT. These results demonstrate that PSII can be assembled in the absence of PsbM and that its deletion does not alter significantly the assembly of thylakoid membrane complexes in vivo, or during the solubilization process. This finding was substantiated in two ways, by BN-PAGE and sucrose gradient centrifugation (Fig. 2, A and B) followed by denaturing SDS-PAGE (supplemental S2). In the first electrophoretic dimension, the separation profile of the native thylakoid membrane complexes in ⌬psbM was close to that of WT, but again a very slight decrease in the amounts of PSII-LHCII supercomplexes was noticed (Fig. 2B). Comparably, immunoblot analyses showed no difference in the presence of the relative amounts of the intrinsic PSII subunits D1, D2, CP43, CP47, PsbE, PsbI, and PsbW, the extrinsic proteins PsbO, PsbP, PsbQ, and PsbR, involved in water oxidation, as well as the antenna proteins CP29, CP26, CP24, and LHCII (supplemental S3).
Levels of PSII relative to PSI were estimated by quantitative immunoblot analysis using D1 and PsaD antibodies and the AIDA software. We repeatedly used dilution series of WT and mutant thylakoids based on chlorophyll amounts to precisely quantify levels of D1 and PsaD (Fig. 2C). Based on this analysis, we calculated that D1 accumulates to 103.4 Ϯ 3.6% and PsaD to 100.7 Ϯ 5.5% in the mutant relative to the WT. Therefore, PSII and PSI levels are comparable in mutant and WT.
Levels of other thylakoid complexes, i.e. PSI, cytochrome b 6 f complex, ATP synthase, and antenna proteins remained unal-tered as well in the mutant (supplemental S3). Because the water splitting complex could have been affected by the loss of PsbM, salt treatment of thylakoid membranes was performed with increasing Na 2 CO 3 concentrations up to 200 mM demonstrating that PsbO, PsbP, PsbQ, and PsbR remain stably associated to the complex in ⌬psbM comparable with WT (data not shown).
Inactivation of PsbM Affects Photosystem II Function-The measurement of chlorophyll fluorescence kinetics of photoautotrophically grown mutants showed a reduced quantum yield (F v /F m ) of PSII (0.78 Ϯ 0.012 versus 0.81 Ϯ 0.003 in WT). Fluorescence quenching analysis uncovered at low light intensity (20 E m Ϫ2 s Ϫ1 ) that the mutant exhibits qP values comparable with those of WT (0.93 Ϯ 0.009 versus 0.96 Ϯ 0.006 in WT), but at higher light intensities (250 E m Ϫ2 s Ϫ1 ) this value remained much lower (0.34 Ϯ 0.041 versus 0.54 Ϯ 0.052 in WT), which could be due to a lower rate of PSII-dependent electron transport causing an increase in the ratio Q A Ϫ /Q A in the mutant as compared with the WT. Lower NPQ values were also detected in the mutant at both light intensities indicating that mutant PSII is more susceptible to thermal dissipation (Table 1). Taken together, these results indicated that although PSII levels are comparable in mutant and WT, the quantum yield of PSII is significantly lower in ⌬psbM. The results above are also supported by our findings that PSII activity measured as O 2 evolution under saturating light was lowered to about 52% in ⌬psbM as compared with WT thylakoids (Table 1).
Energy Transfer to PSII-The 77 K fluorescence emission of ⌬psbM samples exhibited the same peaks as the WT at 685, 695, and 735 nm attributed to the major pigment-protein core complexes of PSII, CP43, CP47, and the PSI-LHCI complex, respectively (Fig. 3). However, a slight lowering of the PSII-related fluorescence emission bands compared with that of PSI was noted in ⌬psbM. This indicated an exiguous disconnection of the outer antenna of PSII and is consistent with the above finding that only levels of PSII supercomplexes are slightly decreased in the mutant.
The Oxidation State of PSI Is Increased in Light-exposed ⌬psbM Plants-Absorption measurements showed a significantly higher oxidation state of PSI in ⌬psbM as compared with WT at both 20 and 250 E m Ϫ2 s Ϫ1 actinic light intensities (Table 1). These results indicate a lower rate of electron flow from PSII to the PQ pool, relative to that of PQH 2 oxidation via PSI activity, and support the conclusion that PSII activity is significantly lower in the mutant as compared with the WT. This conclusion was also supported by measurements of steady state fluorescence levels (F s Ј) elicited by 650 nm actinic light at an intensity of 34 E m Ϫ2 s Ϫ1 on background of far-red light (12 watts m Ϫ2 ). The ratio F s Ј ⌬psbM/F s Ј WT was 0.59 Ϯ 0.06, demonstrating that the electron flow to PQ is much lower than that from the reduced quinol to the electron sink via PSI.
State Transition and Phosphorylation of LHCII and D1/D2 Are Affected in ⌬psbM-Phosphorylation/dephosphorylation of LHCII resulting in transition to state 2 or state 1 was induced by exposing leaves to PSII or PSI light for 15 min, respectively (39). The percentage of state transition was 8.4 for WT and 4.6 for ⌬psbM, corresponding to 55% of the WT activity in ⌬psbM. It has been demonstrated that activation of the protein kinase phosphorylating LHCII is related to the process of plastoquinol oxidation by the cytochrome b 6 f complex (44). Thus, both light-dependent reduction of the PQ pool or addition of duroquinol in darkness induce phosphorylation of LHCII in the WT (Fig. 4A) (40). However, unlike WT, LHCII in ⌬psbM was unexpectedly highly phosphorylated in darkness even in absence of duroquinol. LHCII phosphorylation decreases significantly in light-exposed mutant samples in the absence or presence of DCMU but still increases upon incubation with duroquinol in darkness (Fig. 4A).
A phosphorylation pattern identical to that of isolated thylakoids was obtained when measuring LHCII phosphorylation in vivo. As in isolated thylakoids, LHCII was significantly phosphorylated in dark-adapted mutant leaves. Exposure of such leaves to red light and far-red light of low intensity preferentially exciting PSI caused dephosphorylation of LHCII indicating that the PQ pool in the mutant is reduced in darkness (Fig. 4B). Interestingly, different from WT the phosphorylation level of D1/D2 was strongly reduced under all experimental conditions, whereas that of CP43 was unchanged in the mutant (Fig. 4).

Thermoluminescence Emission Reveal an Increased Ratio for Q B /Q B
Ϫ in Dark-adapted Mutant Plants-Although PSII complexes in ⌬psbM accumulate at levels comparable with the WT, the data presented so far indicated a significant impairment of PSII activity. To check disturbances of the forward/back electron flow within the PSII complex, thermoluminescence emission was recorded (45). Excitation by a single turnover flash of dark-adapted thylakoids advances the S-state cycle by one step, and the quinone or semiquinones bound to the Q B site are fur-

FIGURE 2. Separation of thylakoid protein complexes of WT and ⌬psbM by sucrose gradient centrifugation, BN-PAGE, and quantification of PSII amounts by immunoblot analysis.
A, thylakoid membrane complexes were separated on a continuous sucrose gradient (0.1-1 M) and labeled accordingly (25). B, complexes are labeled according to their identification by mass spectrometry (37). Note slightly reduced levels of PSII supercomplexes in ⌬psbM and the higher accumulation of LHCII-containing complexes. C, representative immunoblot analysis of ⌬psbM and of a dilution series of WT thylakoids for precise quantification of D1 and PsaD proteins. 100% corresponds to 8 g of chlorophyll. ther reduced (46). The luminescence is emitted because of charge recombination between Q B Ϫ and the S 2 -or S 3 -states of the water splitting complex. The temperature at which this luminescence is maximal is related to the redox potential of the Q B :Q B Ϫ and the S-state charge recombining pair and is designated as the B TL emission band. In tobacco, this band occurs at ϳ35°C, as in vascular plants in general (24,47,48).
Because the emission temperature of the B band at 35°C was found to be the same in both mutant and WT, the activation energy for back electron flow and recombination process is not changed in ⌬psbM. The fraction of recombining pairs, Q B Ϫ / S 2 ,S 3 and the B band emission intensities oscillate with a period of four flashes showing a maximum emission at the 2nd and 6th flash in the WT (24). However, ⌬psbM showed various oscillation patterns. In each case the first flash resulted in a higher B band emission than that induced by the second flash. A 1/5 oscillation pattern predominantly appeared in young and fastexpanding leaf material (Fig. 5), whereas binary and zigzag oscillations were rarely and almost exclusively observed in mature and slow growing mutant leaves (supplemental S4). These data clearly reflect a change in the recombination process within PSII in ⌬psbM.
The difference between the patterns of the TL signal oscillation intensity as a function of the number of exciting flashes could be due to deviation from the predicted S 0 /S 1 ratio in the dark adapted state (1/3) and their respective Q B /Q B Ϫ occupancies (1/1) (45). To test this possibility, a simulation program allowing the prediction of the occupancy of the S-states by semiquinone or oxidized quinone, respectively, was employed using the oscillation pattern measured as a function of the number of exciting flashes (supplemental S1).
The results indicate that the quinone occupancy of the Q B site in ⌬psbM significantly differs from that of the WT, showing an average of 2.5 higher ratios for Q B /Q B Ϫ in the various mutant oscillation patterns. No coherent behavior of the S 0 /S 1 ratio in the best fits reported by the simulation program could be observed (supplemental S4). The data presented so far indicate that the back electron flow from Q B Ϫ to the S-states of the    MARCH 30, 2007 • VOLUME 282 • NUMBER 13 oxygen evolving complex in darkness can vary significantly in ⌬psbM.

Effect of PsbM Deletion on PSII Functions
When electron flow from Q A to Q B is inhibited by ligands binding to the Q B site, such as urea or phenolic type herbicides (DCMU and ioxynil, respectively), light excitation results only in the reduction of the Q A quinone. In this case, the related TL band temperature is downshifted, compatible with a lower energy required for back electron transfer to the manganese cluster. The resulting emission band is referred to as Q band (49 -51). Binding of the herbicides to the Q B pocket may slightly modify its structure, which in turn affects the redox potential gap between the Q A and Q B electron acceptors. This was shown to be the case for binding of DCMU that alters the Q A :Q A Ϫ /Q B Ϫ :Q B potential (52,53). The emission temperature of the Q band in ⌬psbM was significantly downshifted in the presence of both DCMU (10°C versus 15°C in the WT) and ioxynil (Ϫ6 to 0°C versus 3°C in the WT) (Fig. 6, A and B). Furthermore, the intensity of the Q band emission increases with increasing ioxynil concentrations (5, 10 and 20 M, respectively), whereas the B band emission persisted in the presence of low ioxynil concentrations and decreased significantly at 20 M (Fig. 6A). These results indicate the presence of functionally distinct populations of PSII centers with altered Q B sites in mutant thylakoids.
The variable and different B band oscillation patterns and the increased Q B /Q B Ϫ ratio in the dark-adapted state of ⌬psbM could be due to a slower back electron flow during the dark adaptation process. This could be caused by an insufficient dark adaptation time of the sample at 20°C prior to light excitation and thus residual Q B Ϫ /S 1 , Q B Ϫ /S 2 , or even Q B Ϫ /S 3 populations may still be present. This in turn could affect the final ratio of the S 1 /S 0 states and their occupancy by semiquinol or oxidized quinone. To check this possibility, the time course of the B band decay was measured at 20°C in darkness. The results demonstrate a complete decay of the B band emission after 3 min of dark adaptation (Fig. 6C). The dark adaptation in all experiments was carried out at 20°C for 3 min prior to application of single turnover flashes and thus was sufficient to allow complete decay of the S 2,3 -states. Therefore, loss of Q B Ϫ /S 2 or Q B Ϫ /S 3 recombination after dark adaptation cannot account for the observed departure from the normal TL oscillation pattern in thylakoids of ⌬psbM.
Photoinhibition and PSII Recovery-The alteration of PSII electron flow and activation energy of recombining pairs within PSII could increase the light sensitivity in ⌬psbM because of singlet oxygen formation accompanied by a loss of variable fluorescence and degradation of the D1 protein (reviewed in Refs. 54 and 55). Therefore, repair of PSII activity requires de novo D1 protein synthesis. To compare the photosensitivity, WT and ⌬psbM leaves were exposed to 500 E m Ϫ2 s Ϫ1 after treatment with chloramphenicol, an inhibitor of chloroplast translation activity. After 180 min of illumination, ϳ84% of PSII quantum yield was lost in ⌬psbM as compared with only 55% in WT leaves. However, exposure to the same light intensity in the absence of chloramphenicol resulted in 70% loss of PSII activity in the mutant as compared with only 27% in WT (Fig. 7A).  (46). C, time course of the B band decay in ⌬psbM thylakoids in darkness. Excitation was given at 25°C, and the sample temperature was maintained for times as indicated prior to fast cooling (10 s) to Ϫ20°C before starting the measurements.
The PSII quantum yield measured in high light-exposed leaves is the result of a balance between the rate of PSII protein degradation (56) and the rate of de novo protein synthesis, reassembly, and photoactivation of the complex. To estimate the ability of the mutant to repair photoinactivated PSII, WT and ⌬psbM plants were irradiated with light of higher intensity (1,500 E m Ϫ2 s Ϫ1 ) to induce significant photoinhibition followed by incubation of the leaves at low light intensity to allow recovery of PSII activity. Before the photoinhibitory treatment, values of F v /F m were typical for WT and mutant leaves used in this experiment. Light exposure was continued until PSII was inactivated to a similar degree (F v /F m ϭ 0.17). PSII activity recovered with the same kinetics and reached almost the initial value of F v /F m within 6 h in both mutant and WT (Fig. 7B), indicating that the capacity of de novo synthesis of the degraded protein(s), reassembly of the complex, and its photoinactivation are not notably impaired by the mutation.

DISCUSSION
PsbM Is Not Required to Maintain the Assembly or Recovery of PSII-PsbM is a conserved hydrophobic LMW subunit of the PSII assembly. The sequence of higher plants PsbM shares ϳ54% identity with that of Synechocystis sp. PCC 6803 (16). The interface between the PSII core monomers forming the dimer that houses PsbM in the cyanobacterial PSII structure as resolved by x-ray diffraction at 3.5 (8) and 3.0 Å (9) also contains the helix PsbTc, and both are close to helix PsbL. PsbM, PsbTc, and PsbL form a protein domain of six helices at the interface of the two PSII monomers. As such, inactivation of psbM results in the loss of two transmembrane helices in the dimeric axis of PSII. Based on the above, two prominent functions could be ascribed to PsbM as follows: (i) involvement in the interaction/ binding of the two monomers, and (ii) a possible role in the controlled association/dissociation of the PSII dimer during the biogenesis and/or the photoinhibition/repair process. If so, loss of PsbM could weaken the dimer interconnection and possibly impair PSII repair during photoinhibitory illumination. However, deletion of PsbM does not prevent formation of dimers as well as assembly of active PSII supercomplexes and photoautotrophic growth under appropriate greenhouse conditions. This implies that PsbM is dispensable for the biogenesis of PSII. The results presented here clearly demonstrate that none of the proposed functions according to the localization within the complex can be assigned to PsbM in tobacco.
LHCII Dark Phosphorylation in ⌬psbM-Our data indicate that loss of PsbM does not affect the accessibility of the corresponding kinase to LHCII nor its redox-regulated phosphorylation in the mutant. Normally, the PQ pool is relatively reduced under state 2 conditions (57). The unexpectedly high level of LHCII phosphorylation in dark-adapted mutant leaves could be due to persistence of a reduced PQ pool in the dark. This might have resulted from chlororespiration and/or NAD(P)H dehydrogenase activity (58,59). On the other hand, an impaired PSII-mediated re-oxidation of PQH 2 in dark-adapted mutant plants could be responsible for the elevated PQ reduction level and consequently phosphorylation of LHCII (see below).
Because application of a weak far-red light was already sufficient to induce dephosphorylation of LHCII, the accessibility of the phosphatase seems to be unaltered. Alternatively, the presence of increased LHCII trimer levels in the mutant may cause aggregation of the trimers during illumination and, as a result, hinder the accessibility of the phosphorylation sites to the protein kinase (40,60).
Loss of PsbM Impairs Electron Transport Within, Outward, and Backward to PSII-Inactivation of the intrinsic PsbM protein caused a reduced PSII activity increasing P700 oxidation and thus maintaining mutant plants in state 1 as revealed by fluorometric analysis and the phosphorylation status of LHCII in the light. This effect could be due to an increase in the redox potential gap between the Q A and Q B quinones. Furthermore, ⌬psbM exhibits a departure from the 2/6 oscillation pattern of the TL B band emission (25).
Unlike WT, mutant plants display quite different patterns of TL oscillations with the number of exciting flashes (supplemental S4). The computational simulation of the dark-adapted PSII states leading to these oscillation patterns indicates that the three different types observed in mutant plants can be generated by a higher Q B /Q B Ϫ ratio in dark-adapted thylakoids ( Fig.  5 and see supplemental S4 and S5).
A residual B band emission persisted even in the presence of 20 M ioxynil in mutant thylakoids, whereas the emission of that band was already abolished by 5 M ioxynil in the WT. Moreover, the emission temperature of the Q band induced by DCMU was downshifted by 5°C in ⌬psbM as compared with the WT, implying a structural change in the Q B site upon binding of DCMU inducing an effect on the activation energy of the recombination process from Q A Ϫ . These results indicate the presence of a residual PSII population in which the Q B sites after ioxynil treatment and dark adaptation are still occupied by PQ that can be reduced following light excitation to form a Q B Ϫ . Possibly, the Q B site properties are progressively altered with the development of mutant leaves. This process varied among mutant leaves as indicated by the TL oscillations and the alteration of the Q B /Q B Ϫ ratio as compared with the WT. Ioxynil binding at the Q B site in ⌬psbM also suggests conformational changes in the structure of the Q B binding pocket and its interaction with the Q A site. This was shown by changes because of DCMU binding at the Q B site and possible alteration of the activation energy for the back electron flow from Q B semiquinone (52,53). Thus, we propose that the Q B site may also have variability in interacting or exchanging bound semiquinone with PQH 2 from the plastoquinol pool. The back electron flow during dark adaptation via deactivation of S 2/3 in the S 2/3 Q B centers may need reduction of the Q B quinone in darkness by a reduced plastoquinol to reach the predicted 1/3 ratio of S 0 /S 1 and 1/1 ratio of Q B /Q B Ϫ (45, 49). Thus it is possible that the alteration of the Q B site interferes with this process resulting in more oxidized Q B sites than semiquinone sites in the dark-adapted PSII populations of the mutant. This again would cause a shift to the 1/5 B band oscillations observed in the mutant. The elevated PQ reduction level and phosphorylation of LHCII in dark-adapted mutant plants may be a consequence of the impaired PSII-mediated reoxidation of PQH 2 involving the Q B pocket. A similar function has also been proposed for the LMW subunit PsbI (25). Unlike ⌬psbI, dimeric PSII supercomplexes accumulate in ⌬psbM. Therefore, loss of PSII dimerization in ⌬psbI does not necessarily account for the observed effect on the 1/5 oscillation pattern and phosphorylation of LHCII in darkness.
To validate the ability of the simulation program to detect changes in the back electron flow, we have simulated the oscillation pattern of the ⌬psbI mutant as well. The oscillation patterns with the number of exciting flashes also fit with the measured values as reported for this mutant (25) when using the parameters in the simulation program as used in this work (supplemental S4). To obtain a good fit of the measured and simulated data, it is necessary to assume the presence of a residual amount of S 2 following dark adaptation. However, an increase in the ratio Q B /Q B Ϫ of the dark-adapted thylakoids is suggested also for the ⌬psbI mutant, in which the Q B /Q A sites interaction is altered as well. Considering the binding properties at the Q B site, these results reflect the most striking functional differences between ⌬psbM and WT.
The results of this work are in agreement with observations that the structure of the PSII core complex exhibits a certain degree of flexibility of the Q B site conformation. This is indicated by the different effect of herbicide binding at this site on the midpoint potential of Q A (53). Furthermore, alteration of the D-de loop of the D2 protein affects the properties of the Q B -binding site of the D1 protein (61). It is thus conceivable that the structures of the quinone-binding sites harbored by the D1/D2 heterodimer and their interactions with surrounding protein helices are not rigid and are subject to fluctuations required for optimizing electron flow outward and within the PSII. Hence, removal of the PsbM helix pair from the interface region could cause a local conformational change that may affect the adjacent regions of the quinone-binding site(s) as well as D1/D2 phosphorylation. The phosphorylation level of CP43 is comparable in mutant and WT implying that its confirmation is unaltered in ⌬psbM allowing access of the corresponding kinase. The LMWs of PSII may play an important role in stabilizing the plasticity of these quinone-binding sites. In their absence, changes may occur in the PSII core that may be detrimental to its activity and light sensitivity.
We conclude that the lower oxygen evolution rate, the reduced photosynthetic quantum yield, the increased light sensitivity, the shift of the Q A midpoint potential, an alteration of the Q B -binding site, and the modified oscillation pattern of the TL signals are related to the loss of PsbM. Our data suggest that the PsbM protein in tobacco plays an important role in ensuring an efficient and functional PSII-mediated electron transport rather than in maintaining the assembly, structure, and stability of the PSII complex.