Restoration of Rapidly Reversible Photoprotective Energy Dissipation in the Absence of PsbS Protein by Enhanced ΔpH*

Variations in the light environment require higher plants to regulate the light harvesting process. Under high light a mechanism known as non-photochemical quenching (NPQ) is triggered to dissipate excess absorbed light energy within the photosystem II (PSII) antenna as heat, preventing photodamage to the reaction center. The major component of NPQ, known as qE, is rapidly reversible in the dark and dependent upon the transmembrane proton gradient (ΔpH), formed as a result of photosynthetic electron transport. Using diaminodurene and phenazine metasulfate, mediators of cyclic electron flow around photosystem I, to enhance ΔpH, it is demonstrated that rapidly reversible qE-type quenching can be observed in intact chloroplasts from Arabidopsis plants lacking the PsbS protein, previously believed to be indispensible for the process. The qE in chloroplasts lacking PsbS significantly quenched the level of fluorescence when all PSII reaction centers were in the open state (Fo state), protected PSII reaction centers from photoinhibition, was modulated by zeaxanthin and was accompanied by the qE-typical absorption spectral changes, known as ΔA535. Titrations of the ΔpH dependence of qE in the absence of PsbS reveal that this protein affects the cooperativity and sensitivity of the photoprotective process to protons. The roles of PsbS and zeaxanthin are discussed in light of their involvement in the control of the proton-antenna association constant, pK, via regulation of the interconnected phenomena of PSII antenna reorganization/aggregation and hydrophobicity.

Sunlight can fluctuate frequently and dramatically in both intensity and spectral quality during the day. The photosynthetic membrane of higher plants has evolved to operate efficiently under such light conditions. In shade, large arrays of antenna complexes efficiently harvest and deliver photon energy to the photochemical reaction centers for use in photosynthesis. In high light the membrane switches to a photoprotective state where the excess absorbed energy, which has the potential to cause damage, is safely dissipated as heat (1,2). The amount of energy dissipation can be monitored by the nonphotochemical quenching of chlorophyll fluorescence (NPQ) 2 (1,2). The major component of NPQ, qE, is rapidly reversible in the dark and is dependent upon the level of the transmembrane proton gradient (⌬pH). ⌬pH acts as a feedback signal indicating the degree of saturation of photosynthetic electron transport (1,2). ⌬pH can be generated by linear electron flow (LEF), which involves the oxidation of water by photosystem II (PSII) and the transfer of electrons to NADP ϩ via the cytochrome b 6 f and photosystem I (PSI) complexes (3). In addition, cyclic electron flow (CEF), which involves the transfer of electrons in a closed loop around cytochrome b 6 f and PSI was reported to contribute to ⌬pH (4 -6). The contribution of CEF to ⌬pH generation is essential for maintaining the correct ATP/ NADPH ratio for CO 2 fixation in the chloroplast stroma (6). Imbalance in the ATP/NADPH ratio can limit the acceptor side of PSI in high light impairing LEF and thus generation of ⌬pH and qE (7).
The ⌬pH is known to act upon several targets to regulate light harvesting. It protonates the light harvesting complexes of PSII (LHCs) (8,9), the LHC-related PsbS protein (10), and also activates the violaxanthin de-epoxidase enzyme that converts the LHC-bound xanthophyll cycle carotenoid from violaxanthin to zeaxanthin (11). Interaction of these three factors is believed to bring about formation of a quencher of chlorophyll singlet excited states (1,2). The relationship between qE and ⌬pH varies depending upon the de-epoxidation state of the xanthophyll cycle carotenoids (12)(13)(14)(15)(16)(17)(18)(19)(20). In the absence of zeaxanthin this relationship is sigmoidal with a pK reported in the range of 4.7-5.0, whereas in the presence of zeaxanthin it is hyperbolic with a pK reported in the range of 5.7-6.5 (12)(13)(14)(15)(16)(17)(18)(19)(20). Thus, because the physiologically typical values of lumen pH encountered in excess light are estimated in the pH 5.5-5.8 range (21), qE is much smaller in the absence of zeaxanthin (22,23). By affecting the sensitivity of qE to ⌬pH, zeaxanthin also modulates the kinetics of qE, enhancing the rate of formation and retarding the rate of qE relaxation (18,24).
The exact site of qE within the PSII antenna and the pigments involved in the quenching mechanism remain under debate. Earlier, it was proposed that PsbS itself is the site of quenching (10). In Arabidopsis mutants lacking PsbS (npq4) qE was completely abolished and only a slowly forming and relaxing NPQ was present (25). The characteristic 535 nm absorption change * The work was supported by Biotechnology and Biological Sciences (⌬A 535 ), which is believed to monitor a structural change within the PSII antenna that accompanies qE, was also reported to be absent (25). The npq4 plants showed enhanced sensitivity to photoinhibition of PSII in excess light and impaired Darwinian fitness in fluctuating light environments (26,27). Conversely, if PsbS was overexpressed then the level of qE was doubled compared with the wild-type (26). When zeaxanthin was mixed with purified PsbS its absorption spectrum was red-shifted mimicking the ⌬A 535 change (28). PsbS was therefore suggested to act as both the sensor of ⌬pH and the binding site of a putative zeaxanthin quencher (10). However, recently new evidence has emerged that argues against a PsbS-zeaxanthin complex being the site of qE. First, both reconstituted and native PsbS were found consistently unable to specifically bind pigments (29). Second, PsbS was still able to perform its function in the absence of zeaxanthin (30,31). Finally, npq4 plants were still shown to possess wild-type levels of photoprotective energy dissipation albeit forming and relaxing 10 times more slowly than in the wild-type (32). The slowly forming NPQ in the npq4 mutant was dependent upon ⌬pH and was also characterized by the ⌬A 535 absorption change, suggesting this signal does not arise from a PsbS-zeaxanthin complex (32). Purified LHCs, on the other hand, readily adopt quenching states upon aggregation in low detergent buffers indicating that they already contain quenching sites (1,13,17). This in vitro quenching in LHCs was found to be modulated by pH and xanthophyll cycle carotenoids and possesses many of the spectroscopic fingerprints associated with qE (1,13,17). It has thus been suggested that PsbS plays an indirect role in qE, somehow supporting the formation of quenching within one, several, or all of the LHCs (17,33,34).
The idea of an indirect role of PsbS in qE is corroborated by its influence on the organization of PSII and LHCs within the thylakoid grana membrane. PSII core dimers are organized within large supercomplexes containing two copies each of the minor LHCs: CP24, CP26, and CP29, and four trimeric LHCII complexes (35). These PSII-LHCII supercomplexes can further associate to form large semi-crystalline domains within the grana (35). PsbS levels were found to influence the amount of these semi-crystalline PSII-LHCII domains and to affect the Mg 2ϩ dependence of grana membrane organization (36,37). Disrupted macro-organization of PSII-LHCII supercomplexes in the grana membranes of Arabidopsis mutants lacking certain LHC proteins has been shown to affect the ⌬pH sensitivity of qE (19). One possibility therefore is that the modified grana membrane organization in the absence of PsbS shifts the ⌬pH threshold for activation of qE to higher values.
To investigate this possibility further we examined ways to increase the levels of ⌬pH in chloroplasts. Artificial mediators of CEF around PSI such as diaminodurene (DAD) and phenazine metasulfate (PMS) were reported to generate high levels of ⌬pH chloroplasts, even in the presence of the LEF inhibitor 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) (38,39). The ability of DAD and PMS to form ⌬pH and thus induce quenching in the presence of DCMU provided the first evidence for a photoprotective form of quenching that was independent of the PSII redox state (38). Later, it was shown that DAD could enhance qE and ⌬pH in pea chloroplasts (40). DAD, in either its neutral or protonated forms, in the absence of ⌬pH was insufficient for observation of quenching, indicating that the compound is not itself a quencher (41,42). The ability of DAD and PMS to modulate CEF relies upon their reduction by the PSI acceptor side. This leads to proton uptake by the mediators from the stroma, diffusion through the membrane, oxidation by the PSI donor side, and proton release into the lumen (42). This artificial cycle around PSI was found to be dibromothymoquinone insensitive and is thus most likely independent of cytochrome b 6 f involvement (43). In the following study it is shown that by enhancing the level of ⌬pH by using DAD and PMS that qE can be observed in intact Arabidopsis chloroplasts from plants lacking the PsbS protein. Moreover, qE in the absence of PsbS is associated with F o quenching, protects PSII from photoinhibition, is modulated by zeaxanthin, and is accompanied by the ⌬A 535 absorption change.

EXPERIMENTAL PROCEDURES
Chloroplast Isolation-Arabidopsis thaliana cv Columbia and npq4 mutant derived from it were grown as previously described (44). Intact chloroplasts were prepared as previously described (30). Chloroplasts devoid of zeaxanthin and antheraxanthin (v) were prepared from spinach leaves dark adapted for 1 h. Chloroplasts enriched in zeaxanthin (z) were prepared from leaves pre-treated for 30 min at 350 mol of photons m Ϫ2 s Ϫ1 under 98% N 2 , 2% O 2 .
Chlorophyll Fluorescence Induction-Chlorophyll fluorescence was measured with a Dual-PAM-100 chlorophyll fluorescence photosynthesis analyzer (Walz, Germany) using the liquid cell adapter. Ascorbate was omitted from all buffers to prevent further deepoxidation taking place during illumination. Control chloroplasts showed a high degree of intactness and qE without addition of exogenous electron acceptors. Where mentioned 400 M DAD or 100 M PMS (in their reduced state) was added to intact chloroplasts to stimulate CEF around PSI in the DAD or PMS-treated chloroplasts. Where mentioned DAD-treated chloroplasts were supplemented with either 50 M DCMU (to inhibit PSII reduction of plastoquinone and thus LEF) or 2 M nigericin (to uncouple the thylakoid membrane). Actinic illumination was provided by arrays of 635 nm LEDs. F o and F o Ј (the fluorescence level with PSII reaction centers open) was measured in the presence of 10 mol of photons m Ϫ2 s Ϫ1 measuring beam. F s is the steady state fluorescence level under actinic illumination. The maximum fluorescence in the darkadapted state (F m ), during the course of actinic illumination (F m Ј), and in the subsequent dark relaxation periods (F m Љ) was determined using a 0.8-s saturating light pulse (4000 mol photons m Ϫ2 s Ϫ1 ). Fluorescence parameters were calculated as fol- Measurement of ⌬pH-⌬pH was determined from the measurement of 9-aminoacridine (9-AA) fluorescence using the Dual-ENADPH and Dual-DNADPH modules for the Dual-PAM-100 chlorophyll fluorescence photosynthesis analyzer (Walz, Germany). Intact chloroplasts were treated as described above in the presence of 1 M 9-AA. Excitation was provided by 365 nm LEDs and fluorescence emission was detected between 420 and 580 nm. Lumen pH and ⌬pH were estimated from the values of 9-AA quenching, using the equation: where Q is the level of 9-AA quenching, V out is the sample volume (1.4 ml), and V in is the lumen volume (56 l/mg of chlorophyll) calculated for the sample chlorophyll concentration of 35 g of chlorophyll/ml using the relationship V in ϭ 50 liters/mol assuming the internal volume of thylakoids is 56 l/mg of chlorophyll (45).
⌬pH versus qE Titrations-Titrations of the ⌬pH dependence of qE were carried out by varying the light intensity of 635 (for actinic illumination) or 710 nm (for far-red illumination) LEDs from 0 to 350 mol of photons m Ϫ2 s Ϫ1 to produce different amplitudes of ⌬pH and the amount of qE that was formed following 5 min of illumination at each light intensity was recorded. For titrations, the amount of qE was expressed on the linear quenching scale where qE ϭ (F m Љ Ϫ F m Ј)/F m Љ. The titration data were fitted to a curve defined by the equation qE ϭ qE max ⌬pH n /(⌬pH n ϩ ⌬pH 0 n ), where qE max is the theoretical maximum qE, ⌬pH is the level of ⌬pH at a particular light intensity, ⌬pH 0 is the level of ⌬pH at which qE ϭ 0.5 qE max , and n is the sigmoidicity parameter (Hill coefficient) (18). pK was then estimated by converting ⌬pH 0 to estimated lumen pH by subtracting ⌬pH 0 from the starting pH (8.0).
pH Versus qE Titrations-As an alternative to estimates of lumen pH based on 9-AA quenching acid titrations were also performed. qE in DAD-treated chloroplasts was formed during 5 min illumination with 350 mol of photons m Ϫ2 s Ϫ1 635 nm LED light. After 5 min an aliquot of concentrated HCl was added to the solution to change the bulk pH to a given value between 3.0 and 8.0 and the light was then immediately turned off. The amount of quenching at pH 8.0 was subtracted from the amount of quenching preserved at each pH value to give the level of qE. The titration data were fitted to a curve defined by the equation qE ϭ qE max pH n /(pH n ϩ pH 0 n ), where qE max is the theoretical maximum qE, pH is the pH of the bulk medium, pH 0 (pK) is the level of pH at which qE ϭ 0.5 qE max , and n is the sigmoidicity parameter (Hill coefficient) (18).
⌬A 535 Absorption Changes-Absorption changes in the 410 -565 nm region were measured as previously described (44).
Pigment Analysis-Pigment composition was determined as previously described (46). Chlorophyll concentration was determined according to the method of Porra et al. (47).

RESULTS
The maximum level of ⌬pH, measured by 9-AA quenching, in control npq4 chloroplasts prepared from light-treated leaves (npq4(z)) ( Table 1) was 37 Ϯ 2.4%. This level of 9-AA quenching corresponds to a ⌬pH of 2.4 and an estimated lumen pH of 5.6, which is broadly consistent with estimates of the normal physiological maximum (21,48). The addition of DAD to npq4(z) chloroplasts enhanced the level of 9-AA quenching to 58.0 Ϯ 3.4% (Fig. 1A), corresponding to a ⌬pH of 4.1 and an estimated lumen pH of 3.9. A similar increase in 9-AA quenching compared with the control was also observed in DAD-treated wild-type chloroplasts prepared from light-treated leaves (WT(z)) (Table 1) (Fig. 1A). DAD enhanced the level of NPQ in control WT(z) chloroplasts by a factor of 2 (Fig. 1, A and  B). In contrast, in control npq4(z) chloroplasts virtually no NPQ was observed within 5 min of illumination. Yet remarkably, in the presence of DAD NPQ was enhanced to almost the same level as in DAD-treated WT(z) chloroplasts (Fig. 1, A and B). More than 90% of the enhanced NPQ in both the DAD-treated WT(z) and npq4(z) chloroplasts was rapidly reversible upon collapse of the ⌬pH when the light was turned off (Fig. 1, A and  B), suggesting it was of the qE-type. In addition, the quenching of the F o fluorescence level increased from 0 and 18.0 Ϯ 3.0% in the control npq4(z) and WT(z) chloroplasts, respectively, to 28 Ϯ 2.3 and 30.0 Ϯ 2.0% in those treated with DAD, suggesting the additional quenching occurred within the PSII antenna.
The absence of qE in npq4 plants makes them potentially more vulnerable to rapid exposure to high light. This is due to photoinhibition, the permanent closure of PSII reaction centers as a result of photodamage caused by excess light energy in the absence of NPQ (26,32). The enhanced NPQ mediated by DAD in npq4(z) and WT(z) chloroplasts was capable of substantially lowering the PSII excitation pressure during illumination as indicated by photochemical quenching (qP) (Fig. 2). Levels of qP in DAD-treated npq4(z) chloroplasts were nearly double those of the control npq4(z) sample during illumination. Upon dark recovery the level of qP in the control npq4(z) chloroplasts was reduced relative to the pre-illumination level, consistent with photoinhibition (32), a feature not observed in control WT(z) chloroplasts (Fig. 2). In contrast, in the DAD-treated npq4(z) chloroplasts the level of qP following dark recovery was virtually identical to the pre-illumination level indicating that the observed NPQ was photoprotective (Fig. 2).
We performed several control experiments to check that the enhanced NPQ in DAD-treated chloroplasts was (a) related to the enhanced ⌬pH and (b) that the enhanced ⌬pH was due to enhanced CEF around PSI. First we tested whether DADtreated chloroplasts showed any NPQ in the presence of the uncoupler nigericin. Both npq4(z) and WT(z) chloroplasts showed virtually no NPQ in the presence of nigericin indicating that the quenching relied on the ⌬pH (Fig. 3A). Second we tested the effect of inhibiting LEF with DCMU on the enhanced NPQ in DAD-treated chloroplasts. In the absence of LEF, the level of 9-AA quenching was reduced to 55.0 Ϯ 2.6% in WT(z) chloroplasts and 53.0 Ϯ 3.4% in npq4(z) chloroplasts. The level

Sample
Neo

qE-type Photoprotective Energy Dissipation without PsbS
of NPQ was also reduced in both the DCMU/DAD-treated WT(z) and npq4(z) chloroplasts (Fig. 3A) compared with those treated with DAD alone. The decrease in NPQ was larger in npq4(z) than WT(z) chloroplasts (Fig. 3A). However, higher levels of NPQ could still be sustained by ⌬pH generated by CEF alone compared with the control (cf. Figs. 1B and 3A). Third, we tested whether PMS, another artificial CEF modulator could induce high ⌬pH and NPQ. Addition of PMS to WT(z) and npq4(z) chloroplasts mimicked the effect of DAD on ⌬pH increasing 9-AA quenching to 59.0 Ϯ 3.3 and 56.0 Ϯ 2.9%, respectively. The enhanced ⌬pH in the presence of PMS supported high levels of NPQ in both WT(z) and npq4(z) chloroplasts (Fig. 3A). Finally, we tested whether far-red light (which is mainly absorbed by PSI) could generate ⌬pH in the presence of DAD. Nearly 3 times higher levels of 9-AA quenching were observed in DAD-treated npq4(z) chloroplasts than in the untreated control (Fig. 3B). These data suggest that the ability of DAD to promote NPQ in npq4(z) chloroplasts is indeed associated with the enhanced levels of ⌬pH stimulated by enhanced CEF.
To further investigate whether the DAD-mediated NPQ observed in npq4(z) chloroplasts occurred by the same mechanism as qE the kinetics of the ⌬A 535 absorption change were recorded and compared with the WT(z) chloroplasts (Fig. 4A). In the control WT(z) chloroplasts the ⌬A 535 signal increases rapidly upon illumination and reaches saturation within ϳ1 min. ⌬A 535 then rapidly relaxes when the light is switched off (Fig. 4A). Consistent with the absence of qE in control npq4(z) chloroplasts no light-induced ⌬A 535 signal was observed (Fig.  4A). The enhancement of NPQ in DAD-treated WT(z) chloroplasts was accompanied by enhanced levels of ⌬A 535 (Fig. 3A). DAD treatment also stimulated the rapid emergence of ⌬A 535 in the npq4(z) chloroplasts with a similar amplitude as in the DAD-treated WT(z) chloroplasts (Fig. 4A). The light-minusrecovery difference spectra in the Soret region for DAD-treated npq4(z) chloroplasts possessed all of the qE characteristic bands that were present in the control WT(z) chloroplasts, with a maximum at 535 nm and minima at 438, 468, and 495 nm (Fig.  4, B and C) (32,44). These data suggest that the enhanced NPQ in the DAD-treated WT(z) and npq4(z) chloroplasts occurs by the same mechanism as qE.

qE-type Photoprotective Energy Dissipation without PsbS
The differences between WT(z) and npq4(z) chloroplasts were investigated further by determining the qE versus ⌬pH titration curves in control and DAD-treated samples (Fig. 5).
Because DAD enhances ⌬pH it should be possible to extend the known ⌬pH versus qE titration curves (12)(13)(14)(15)(16)(17)(18)(19)(20). In addition, it can be more carefully ascertained whether DAD generates qE in npq4(z) chloroplasts by merely enhancing ⌬pH or rather by modifying the nature of the relationship between qE and ⌬pH. The qE data in the presented titrations is expressed according to the linear quenching calculation as qE ϭ (F m Љ Ϫ F m Ј)/F m Љ, consistent with a number of previous studies (12)(13)(14)(15)(16)(17)(18)(19)(20). Both the control and DAD-treated WT(z) chloroplasts obeyed the same hyperbolic relationship between qE and ⌬pH (Fig. 5). It has previously been observed in wild-type Arabidopsis chloroplasts that the absence of zeaxanthin shifts the ⌬pH dependence of qE to higher values (12)(13)(14)(15)(16)(17)(18)(19)(20). It was interesting to see whether DAD-enhanced qE was sensitive to zeaxanthin so we titrated the level of qE in chloroplasts prepared from dark-adapted wild-type Arabidopsis leaves (WT(v)) ( Table 1). As in WT(z) chloroplasts the relationship between qE and ⌬pH was unchanged in DAD-treated WT(v) chloroplasts, following the same sigmoidal relationship as in the control (Fig. 5). In DADtreated npq4(z) chloroplasts virtually no qE was observed below a ⌬pH of 2.5, consistent with the behavior of control npq4(z) chloroplasts. However, above this value qE rises to levels approaching the maximum wild-type value in DAD-treated npq4(z) chloroplasts (Fig. 5). The titration curve in npq4(v) chloroplasts, prepared from dark-adapted leaves, was shifted to the right compared with that in npq4(z) chloroplasts, thus higher levels of ⌬pH were required to bring about the same level of qE in the former (Fig. 5). Thus in both DAD-treated npq4 and WT chloroplasts zeaxanthin affected the relationship between qE and ⌬pH, providing further evidence that DAD-mediated qE occurs by the same mechanism. The extent of sigmoidicity (i.e. the degree of cooperativity, as given by the Hill co-efficient) and the pK (i.e. the proton-antenna association constant for qE)   JUNE 3, 2011 • VOLUME 286 • NUMBER 22 of each titration curve can be quantified by transforming the data with the Hill equation (see "Experimental Procedures"). The pK of qE in DAD-treated WT(z) was 6.2 with a Hill coefficient of 1.36, compared with a pK of 5.0 and a Hill coefficient of 3.7 in WT(v) chloroplasts (Table 2). In the DAD-treated npq4(z) chloroplasts the pK was 4.6 with a Hill coefficient of 4.3, whereas in the npq4(v) chloroplasts the pK was shifted as low as 4.3 with a Hill co-efficient of 7.22. The titration curves thus reveal for the first time that PsbS, like zeaxanthin, controls the relationship between qE and ⌬pH. Moreover the effects of zeaxanthin and PsbS were found to be additive.

qE-type Photoprotective Energy Dissipation without PsbS
Previous studies have shown that a lumen pH of 5.0 and below has damaging effects on the oxygen-evolving complex of PSII and on plastoquinone oxidation by the cytochrome b 6 f complex (49,50). Any such negative effects of low lumen pH would be predicted to suppress the level of qP during illumination. Contrary to this prediction qP was actually elevated in the presence of the DAD-mediated NPQ in npq4(z) and WT(z) chloroplasts relative to the control samples (Fig. 3). Therefore, we checked whether the 9-AA quenching method provides a reasonable estimate of lumen pH by titrating the relaxation of qE in WT(z) and npq4(z) chloroplasts with acid. It has previously been demonstrated that qE can be sustained by lowering the pH of the bulk medium in spinach thylakoids, i.e. preventing the escape of protons from the lumen (15). Here we allowed qE to form in DAD-treated WT(z) and npq4(z) chloroplasts and then added an aliquot of acid to change the bulk pH of the medium to a chosen value between pH 8.0 and pH 3.7. The actinic illumination was then immediately turned off. When the bulk pH was lowered to 4.0 virtually all of the qE was maintained in the dark in the DAD-treated WT(z) and npq4(z) chloroplasts (Fig. 6, A and B). However, when the bulk pH was lowered to 5.2 the WT(z) chloroplasts still maintained over 50% of the maximum qE in the dark (Fig. 6A), however, in npq4(z) virtually all of the qE was absent (Fig. 6B). By repeating these experiments across a range of pH values qE titration curves

TABLE 2 ⌬pH versus qE titration curve fitting parameters in intact Arabidopsis chloroplasts
The titration parameters for the data presented in Fig. 5 were determined by fitting the data with the equation: qE ϭ qE max ⌬pH n /(⌬pH n ϩ ⌬pH 0 n ), where qE max is the theoretical maximum qE, ⌬pH is the level of ⌬pH at a particular light intensity, ⌬pH 0 is the level of ⌬pH at which qE ϭ 0.5qE max, and n is the sigmoidicity parameter (Hill coefficient). pK was then estimated by converting ⌬pH 0 to estimated lumen pH by subtracting from the bulk pH (8.0).

qE-type Photoprotective Energy Dissipation without PsbS
were determined for WT(z), WT(v), npq4(z), and npq4(v) chloroplasts (Fig. 7A). The pK values determined from these titration curves compared favorably with those calculated from the 9-AA quenching data with a deviation of no more than 0.3 pH units (cf . Tables 2 and 3). This fact is emphasized in Fig. 7A by overlaying the acid titration data for each sample with the respective fits from Fig. 5. The fact that qE in npq4(z) chloroplasts could be maintained in the dark only at pH 4.0 suggests that the level of ⌬pH induced by DAD is indeed as high as 4.0 (Fig. 6B). When the bulk pH of WT or npq4 chloroplasts was lowered to 3.7, less qE was maintained in the dark indicating that the system may be denatured at these values (Fig. 7A). We also investigated the effect of the pH of the bulk medium on the level of qP in the dark in WT(z) and npq4(z) chloroplasts following 5 min illumination (Fig. 7B). Consistent with previously published data (38, 39) we found that the onset of damage to PSII (quantified by qP) began below a bulk pH of 5.0. The fall in qP was particularly sharp between pH 4.0 and 3.5 falling from 0.8 to 0.4 (Fig. 7B). In contrast in samples where only the lumen pH is low during illumination (due to ⌬pH formation) no significant fall in qP levels occurred (Fig. 7B). These data indicate that low bulk pH has more serious consequences for PSII activity compared with low lumen pH alone.

DISCUSSION
DAD-induced CEF around PSI enhanced ⌬pH levels in Arabidopsis chloroplasts from a value of 2.4 in the controls where only natural electron acceptors were present up to 4.1. The ability of far-red light to induce significant levels of ⌬pH in the presence of DAD confirmed the involvement of PSI-driven CEF. However, treatment of chloroplasts with the PSII inhibitor DCMU confirmed that maximal levels of ⌬pH and qE were only observed when LEF and CEF were operating in tandem. CEF-driven ⌬pH has previously been shown to modulate qE in tobacco and Arabidopsis (7,51). Enhanced levels of ⌬pH allowed us to test the hypothesis that PsbS acted as a catalyst of qE by affecting the ⌬pH sensitivity of the process. Wild-type levels of qE were observed for the first time at a ⌬pH level of 4.0 in intact npq4(z) chloroplasts. Acid titration of qE relaxation confirmed that it required a pH of 4.0 to prevent any significant relaxation of qE in npq4, confirming the ability of DAD to induce very high lumen acidification. It is important to note that calculations of lumen pH and ⌬pH in this article require an estimate of the lumen volume. For this end we used the values calculated by Avron and co-workers (52) on lettuce chloroplasts. Indeed, we note that variations in the lumen volume do not affect dramatically the calculated values of ⌬pH. Thus, for example, if 9-AA fluorescence is 50% quenched and a lumen volume of 0.5 l/ml is assumed, the ⌬pH will be 3.3. If the volume is 1.0 l/ml, the actual ⌬pH will be 3.0, only 0.3 pH units lower. Nevertheless, the systematic error in the estimation of the lumen volume could be even higher, therefore the pH calculation can potentially have 1 unit or so of error (as shown in Fig. 5). The cause of this problem could come from the fact that chloroplast intactness and hence thylakoid lumen volume is likely to display significant variability depending upon isolation procedures. Precisely for this reason we undertook an additional set of measurements using the stroma acidification titrations to prevent qE recovery (Figs. 6 and 7) and demonstrated a close consistency in the calculated pH requirement for qE (pK) in both methodologies and are also broadly consistent with estimates made by electrochromic shift (⌬A 515 ) measurements (48). Solid lines are fits of the data presented in Fig. 5 to show close agreement between the two sets of data. B, effect of lowering the pH of the bulk medium on qP levels in the dark following 5 min illumination to induce maximum qE in DAD-treated WT(z) (white circles) and npq4(z) (black circles) chloroplasts. Shown for comparison is the qP level in the dark following 5 min illumination where only the lumen pH level fell below 8.0 (due to ⌬pH formation) during illumination, DAD-treated WT(z) (white squares) and npq4(z) (black squares) chloroplasts.

TABLE 3 pH versus qE titration curve fitting parameters in intact Arabidopsis chloroplasts
The titration parameters for the data presented in Fig. 7 were determined by fitting the data with the equation: qE ϭ qE max pH n /(pH n ϩ pH 0 n ), where qE max is the theoretical maximum qE, pH is the pH of the buffer medium, pH 0 is the pH level at which qE ϭ 0.5 qE max, and n is the sigmoidicity parameter (Hill coefficient).

qE-type Photoprotective Energy Dissipation without PsbS
The qE generated in npq4 chloroplasts in the presence of DAD displayed the same features normally associated with the process in the wild-type. Like qE, it was sensitive to zeaxanthin, accompanied by the qE-related absorption changes in the Soret band (often known as ⌬A535), and quenched the level of fluorescence when all PSII reaction centers were in the open state (F o state) consistent with antenna based quenching. These "qElike" features distinguish the DAD-mediated quenching from PSII donor-side quenching mechanisms induced by low pH in the presence of strong oxidants such as ferredoxin (15,53,54). Our data also provide interesting insights into the inactivation of the PSII OEC by low pH. When the pH of the bulk medium was lower than 5.0 a significant inactivation of PSII occurred (as determined by qP levels) consistent with previous studies (49,50). High ⌬pH alone, however, as seen in DAD-treated chloroplasts did not appear to inactivate PSII. Moreover, the qE in npq4 chloroplasts actually reduced the excitation pressure on PSII protecting it from the light-induced damage that was observed in control npq4 chloroplasts.
An understanding of how different molecules affect the relationship between qE and ⌬pH can provide clues as to their function. Our titrations reveal the additive nature of the effect of zeaxanthin and PsbS on the pK a of protonable residues associated with qE. The apparent pK a of amino acids has been shown to strongly depend upon their environment (55)(56)(57). Hydrogen bonding, steric hindrance, and the di-electric constant of the environment can all affect the pK a of amino acids (55)(56)(57). For instance the pK a of the carboxyl group on aspartate can be as low as 2.4 when exposed to water due to hydrogen bonding, whereas in hydrophobic environments the pK a can be as high as 6.4 (56). The ability of xanthophylls to affect the pK a of lumen-facing amino acids is not surprising given their essential role in shaping LHC tertiary and quaternary structures (58,59). LHC complexes bind 4 different types of xanthophylls, the hydrophobic character of which has been shown to vary due to their differing chemical structures (60). The most hydrophobic xanthophyll zeaxanthin was found to promote LHC aggregation and shift the pK of in vitro quenching to more alkaline values (24,61). There is also biochemical evidence that zeaxanthin can influence the isoelectric point of LHC complexes (62). The more polar xanthophylls, such as violaxanthin, had the opposite effect, inhibiting aggregation and shifting the pK of quenching to more acid values (24,61). Recently, the pK of qE in various xanthophyll biosynthetic mutants was also found to be correlated with xanthophyll hydrophobicity (63). Compared with zeaxanthin the effect of PsbS on the pK a of protonable residues associated with qE is harder to understand.
Biochemical evidence has recently been provided that the PSII antenna is reorganized in response to ⌬pH, altering the interactions between LHC subunits such that a part of the PSII-LHCII supercomplex containing CP24, CP29, and LHCII dissociates (64,65). This reorganization/aggregation of the PSII antenna in high light was shown to depend upon the presence of PsbS and was enhanced by zeaxanthin (64,65). We suggest that the ability of PsbS to affect the pK of qE is associated with its influence on the PSII organization. There is evidence that grana membranes are much more rigidly organized with a higher proportion of ordered PSII-LHCII semi-crystalline arrays (37). The altered organization increases the concentration of Mg 2ϩ cations required for the reassembly of the PSII-LHCII supercomplex in the fluorescent state and restacking of the grana membranes (36). It seems that the reverse is also true, in the absence of PsbS more protons are required to reorganize PSII and LHCII into the dissipative state. We suggest that without PsbS, reorganization of PSII and LHCII occurs only very slowly, explaining the slowly forming NPQ phenotype of control npq4 plants (32). However, the energetic barrier imposed by the more rigid organization of PSII-LHCII in the absence of PsbS can be overcome by higher levels of protonation. Our data are therefore consistent with earlier suggestions that PsbS acts as a catalyst of qE formation (17,33,34). Changes in subunit interactions and/or stabilities of different conformations are predicted by the allosteric model of Monod et al. (66) to alter the cooperativity of substrate (in this case proton) binding, exactly as observed in npq4 chloroplasts. In control chloroplasts where no artificial electron acceptors were added, the maximum ⌬pH we observed was 2.4, consistent with the moderate lumen pH independently estimated in vivo (21). Thus, by promoting reorganization of the PSII-LHCII supercomplex and aggregation of LHCs, PsbS and zeaxanthin work synergistically to shift the pK of qE by nearly 2 pH units allowing almost maximal photoprotection at physiologically attainable levels of lumen pH.
This work provides a plausible explanation for the involvement of PsbS in qE regulation. The mechanism of PsbS action is explained by its synergistic control with zeaxanthin of the proton-antenna association constant, pK, via regulation of the interconnected phenomena of LHC antenna reorganization/ aggregation and hydrophobicity. The findings in this article thus mark another important step toward understanding this vital photoprotective process.