A Novel Mechanism of Nuclear Photosynthesis Gene Regulation by Redox Signals from the Chloroplast during Photosystem Stoichiometry Adjustment*

Photosynthetic organisms acclimate to long term changes in the environmental light quality by an adjustment of their photosystem stoichiometry to main-tain photosynthetic efficiency. By using light sources that predominantly excite either photosystem I (PSI) or photosystem II (PSII), we studied the effects of excitation imbalances between both photosystems on nuclear PSI gene transcription in transgenic tobacco seedlings with promoter:: (cid:1) -glucuronidase gene fusions. Shifts from PSI to PSII light sources (and vice versa) induced changes in the reduction/oxidation state of intersystem redox components, and acclimation of tobacco seedlings to such changes were monitored by changes in chlorophyll a / b ratios and in vivo chlorophyll a fluorescence. The ferredoxin-NADP (cid:2) oxidoreductase gene promoter did not respond to these treatments, those from the genes for subunits PsaD and PsaF of PSI are activated by a reduction signal, and the plastocyanin promoter responded to both reduction and oxidation signals. Additional experiments with photosynthetic electron transport inhibitors 3-(3 (cid:1) ,4 (cid:1) -dichlorophenyl)-1,1 (cid:1) -dimethyl urea and 2,5-dibromo-3-methyl-6-isopropyl- p -benzoquinone demonstrated

Photosynthetic organisms acclimate to long term changes in the environmental light quality by an adjustment of their photosystem stoichiometry to maintain photosynthetic efficiency. By using light sources that predominantly excite either photosystem I (PSI) or photosystem II (PSII), we studied the effects of excitation imbalances between both photosystems on nuclear PSI gene transcription in transgenic tobacco seedlings with promoter::␤-glucuronidase gene fusions. Shifts from PSI to PSII light sources (and vice versa) induced changes in the reduction/oxidation state of intersystem redox components, and acclimation of tobacco seedlings to such changes were monitored by changes in chlorophyll a/b ratios and in vivo chlorophyll a fluorescence. The ferredoxin-NADP ؉oxidoreductase gene promoter did not respond to these treatments, those from the genes for subunits PsaD and PsaF of PSI are activated by a reduction signal, and the plastocyanin promoter responded to both reduction and oxidation signals. Additional experiments with photosynthetic electron transport inhibitors 3-(3,4-dichlorophenyl)-1,1-dimethyl urea and 2,5-dibromo-3-methyl-6-isopropyl-p-benzoquinone demonstrated that the redox state of the plastoquinone pool controls the activity of the plastocyanin promoter, whereas subunit PsaD and PsaF gene transcription is regulated by other photosynthesis-derived signals. Thus, the expression of nuclear-encoded PSI genes is controlled by diverse light quality-dependent redox signals from the plastids during photosystem stoichiometry adjustment.
The photosynthetic apparatus in the chloroplasts of higher plants and algae is comprised of a patchwork of nuclear-and chloroplast-encoded components. Nuclear-encoded genes for structural components of the photosynthetic machinery are either new or a result of a gene transfer from the endosymbiotic ancestor of chloroplasts to the nucleus of the host cell (1)(2)(3)(4). The enormous differences in gene copy number between both compartments require a highly coordinated regulation in their expression during development and acclimation of the organism to environmental cues. This coordination is controlled by the nucleus at many levels (5) but also involves signals from the plastids, which influence the expression of nuclear genes for plastid proteins (6 -9). The exact nature of the plastidderived signal(s) is still elusive. Inhibition of either plastid transcription or translation or photo-oxidative destruction of chloroplasts prevents the transcription of several nuclearencoded photosynthesis genes (10 -13). Other crucial components involved in this interorganelle cross-talk are intermediates and/or components of the tetrapyrrol biosynthesis pathway (14 -18) or the availability of phosphoenolpyruvate (19). One of the central players in this scenario is light, and it seems to regulate the expression of several nuclear genes for plastid proteins via the same cis-active elements as the plastidderived signal(s) (20). Besides cytosolic photoreceptors plants sense changes in light quantity or quality via the modulation of the reduction/oxidation (redox) state of various chloroplast molecules involved in photosynthesis. This redox signaling provides a feedback link between the degree of photosynthetic efficiency and the expression of nuclear photosynthesis genes (21)(22)(23), which helps to acclimate the photosynthetic process to varying environmental conditions. A few examples for this are known. Acclimation to different light intensities and temperatures of the unicellular algae Dunaliella tertiolecta and Dunaliella salina involves changes in the transcription of the nuclear-encoded chlorophyll-binding proteins of the light-harvesting complex (LHCB) genes, which are regulated by the redox state of the PQ 1 pool (24,25). Dark/light shift experiments with transgenic tobacco revealed that photosynthetic electron transport controls also transcription of the pea ferredoxin (FED1) gene as well as FED1 mRNA ribosome loading (26), and the inhibition of photosynthetic electron transport by DCMU also destabilizes the FED1 mRNA (27). Under high light stress, Arabidopsis showed an increase in the transcription of the nuclear APX genes (encoding cytosolic ascorbate peroxidases) probably controlled by the PQ redox state and H 2 O 2 (28,29). Furthermore, an interaction between photosynthetic electron transport and sugar regulation on nuclear photosynthesis gene expression has been postulated. In an Arabidopsis cell culture the observed increase in LHCB and PETE (encoding plastocyanin) transcription after sugar repletion was diminished in the presence of DCMU (30).
Illumination conditions that predominantly excite photosystems I (PSI) or II (PSII) generate an imbalance in excitation energy distribution between PSI and PSII, which results in a decrease of the photosynthetic efficiency. Such light quality gradients that typically appear in canopies of trees or forests, in dense plant populations, and aquatic environments are counterbalanced by plants with an initial short term response called state transitions, and a following long term response that causes a readjustment of PS stoichiometry. This readjustment requires changes in photosynthesis gene expression (31,32). With light sources that predominantly excite PSI or PSII we could demonstrate that the redox state of the PQ pool controls transcription of the chloroplast-encoded PS genes psbA and psaAB in mustard (33,34). However, at least half of the structural compounds of the photosynthetic apparatus are encoded in the nucleus (35,36), and nothing is known about how their expression is coupled to the changed expression in chloroplasts under varying light quality. In this study, we focus on the nuclear-encoded PSI genes PSAD (encoding subunit PsaD), PSAF (encoding subunit PsaF), PETE, and PETH (encoding ferredoxin-NADP ϩ -oxidoreductase) to address the question of whether their expression also responds to changes of plastid redox parameters. These genes encode crucial components of the PSI complex. The subunits PsaD and FNR are located at the stromal side of the complex and provide an essential site for ferredoxin docking (37)(38)(39) and the electron transfer from ferredoxin to the electron acceptor NADP ϩ . The subunits PsaF and plastocyanin are located at the lumenal side. The mobile electron carrier plastocyanin transfers electrons from the cytochrome b 6 f complex to PSI, and PsaF provides a docking site for this protein (40 -42). Recent studies using promoter::␤-glucuronidase (GUS) fusions in transgenic tobacco revealed that these four promoter constructs respond to light and the plastid signal(s) (19,43,44). To test whether promoter activation of these genes is coupled to the photosynthetic electron transport, we grew transgenic tobacco seedlings under the above mentioned PS-specific light sources. We found that the promoters of the genes PSAD, PSAF, and PETE but not PETH respond to light-induced redox signals from the plastid. Furthermore, inhibitor assays revealed that transcription of the genes PSAD, PSAF, and PETE is coupled to photosynthetic electron transport via different redox systems. This demonstrates that redox signals do not only operate within the chloroplast (33) but also affect nuclear-encoded photosystem genes with the aim to coordinate nuclear and chloroplast PSI gene expression during PS stoichiometry adjustment.

EXPERIMENTAL PROCEDURES
Plant Material and Growth Conditions-Plants were grown in a temperature-controlled growth chamber at 22°C. Mustard seedlings (Sinapis alba L.) were grown on vermiculite in plastic boxes under continuous light. They were grown under PSI or PSII light for 7 days or 5 days in PSI light followed by 2 days of PSII light and vice versa as described (33,34). Tobacco wild-type (SamsunNN) and transgenic lines were germinated and grown on 1/2 Murashige and Skoog medium containing 1.35% sucrose and, in case of transgenic lines, 80 g/ml kanamycin in Petri dishes. Generation of the transgenic lines used here have been described earlier (45)(46)(47)(48). Promoter::GUS fusion constructs introduced into tobacco contained the following cis-active regions relative to the transcription start site (ϩ1): Ϫ1802 to ϩ55 (PSAD), Ϫ1074 to ϩ163 (PSAF), Ϫ1126 to ϩ60 (PETE), and Ϫ731 to ϩ231 (PETH). Tobacco seedlings were grown for 18 days under continuous PSI or PSII light or 14 days in PSI light followed by 4 days in PSII light or vice versa. Spectral quality and light quantity of the PSI and PSII light sources have been described (33). White-light control plants were illuminated continuously for 7 days (mustard) or 14 days (tobacco) with 30 W white-stripe lamps (OSRAM) with a photon flux density of 100 microeinsteins.
Inhibitor Treatments-The electron transport inhibitors DCMU and DBMIB (Sigma) have been applied to the plants grown in Petri dishes by spraying 0.5 ml of indicated concentrations on the leaves using a 10-ml fine sprayer. Control seedlings were treated with the solvent without inhibitors. Wild-type seedlings were illuminated with white light until Chl fluorescence measurements were performed. Transgenic tobacco lines grown 14 days under PSI or PSII light sources were treated with inhibitors in the same way directly before the respective light switch to PSII or PSI light. The DBMIB treatment was repeated every 3-4 h during the last 96 h of experimentation. Stock solutions of DCMU were 10 mM in 50% ethanol and of DBMIB were 100 mM in 10% Me 2 SO in ethanol. The different inhibitor concentrations were prepared by dilution in sterile water directly prior use.
Chl Fluorescence Measurements-In vivo Chl a fluorescence parameters were measured with a pulse amplitude-modulated fluorometer (PAM101/103, Heinz Walz, Effeltrich, Germany). Mustard seedlings were positioned solely under the emitter/detector unit, and tobacco seedlings were set together as a dense population on a Petri dish prior to the measurements. After dark acclimation (8 -10 min) the measuring beam was turned on, and minimal fluorescence (Fo) was determined. Then leaves were exposed to a 500-ms flash of saturating white light (6000 microeinsteins) to determine maximal fluorescence (Fm), and the Fv/Fm ratio was calculated. Subsequently leaves were illuminated with 100 mol of photons m Ϫ2 s Ϫ1 of actinic red light of 600 nm (Walz 102-R). Fluorescence was recorded in the saturation pulse mode by the application of saturating flashes every 30 s to determine maximal fluorescence of illuminated leaves (FmЈ) until a stable fluorescence level (Ft) was reached. Actinic light was switched off, and far-red light (Walz 102-FR) was turned on to oxidize the electron transport chain and to determine minimal fluorescence (FoЈ) in the light-acclimated state. The steady-state fluorescence Fs was calculated as Ft Ϫ FoЈ ϭ Fs. The fluorescence quenching parameter qP (photochemical quenching) was calculated as qP ϭ (FmЈ Ϫ Fs)/(FmЈ Ϫ Fo) (49). The effective quantum yield of PSII (PSII) was calculated as PSII ϭ (FmЈ Ϫ Fs)/FmЈ (50).
Chl Determination-Plants were harvested under the respective growth light and kept on ice, and Chl concentrations were determined after grinding the material in liquid nitrogen and extracting the pigments with 80% buffered acetone (51). Chl a and b were calculated according to Porra et al. (52).
GUS Activity Assays-GUS activities in transgenic lines were determined by a standard protocol described by Lü bberstedt et al. (48). 100 mg of fresh plant material was homogenized in an Eppendorf tube, and insoluble particles were removed by high speed centrifugation at 4°C. The GUS activities were determined three times from each extract and expressed as pmol of 4-methylumbelliferone accumulation/min/g fresh weight.

RESULTS
Acclimation to PSI and PSII Light Sources-The acclimation of higher plants to light sources favoring either PSI or PSII can be monitored easily by determination of the Chl a/b ratio, which represents a rough estimate of changes in PS stoichiometry (cf. Ref. 34). Tobacco seedlings were grown under PSI and PSII light (PSI and PSII plants) for 14 days, which resulted in an acclimation (Fig. 1A, 1st acclimation) to the respective illumination conditions as indicated by a low Chl a/b ratio for PSI plants and a high ratio for PSII plants (Fig. 1A). A shift of the seedlings to the other light source forces the plants to acclimate to the new illumination conditions (Fig. 1A, 2nd acclimation). Such shifts were shown to induce state transitions (34) and to influence the redox state of the PQ pool (53). A shift from PSI to PSII light results in a more reduced PQ pool (Fig. 1A,  reduction signal), whereas a shift from PSII to PSI light induces a more oxidized state of PQ (Fig. 1A, oxidation signal). 96 h after the light shift, tobacco seedlings are completely acclimated to the new light conditions. This process occurs fast at the beginning (half-time 24 h) and significantly slower toward the end of the treatment. In vivo Chl a fluorescence measurements with fully acclimated tobacco plants confirm differences in the photosynthetic electron transport efficiency, which result from the different acclimations. Representative Chl fluorescence induction curves of PSI and PSII light-acclimated tobacco obtained by standard pulse amplitude-modulated fluorescence measurements (Fig. 1B) demonstrate a higher level in steady-state fluorescence Fs in PSI plants when compared with PSII plants. A more detailed analysis of photosynthetic parameters is given in Table I. Mustard, for which changes in PS stoichiometry in response to the same PSI or PSII light sources have been described (33,34), were used as a control. Both PSI and PSII plants show no significant changes in the maximal quantum yield after acclimation to the respective other light sources. In addition, the Fv/Fm value of both species is between 0.80 and 0.84 under all conditions tested, which is typical for dark-acclimated, healthy, and unstressed plants. However, the Fs/Fm ratio is high in PSI plants and decreases after acclimation to PSII light; the opposite reaction was observed in PSII plants after acclimation to PSI light. As a consequence the reduction state of the first electron acceptor of PSII, Q A , expressed as 1 Ϫ qP, is significantly higher in PSI and PSII3 PSI plants than in PSII and PSI3 PSII plants. The Fs/Fm ratio is also regarded as a rough measure for the PQ redox state (30), and our data indicate that it is more reduced in PSI and PSII3 PSI-plants than in PSII and PSI3 PSII plants. In contrast, the effective quantum yield, PSII, is higher in PSII plants than in PSI plants and changes accord-ingly after a shift of the seedlings to the respective other light source. Taken together, these data indicate a limited electron transport capacity in PSI and PSII3 PSI plants in comparison with the capacity measured in PSII and PSI3 PSII plants. This is consistent with results from mustard; the PSII/PSI ratio is high after acclimation to PSI light and low after acclimation to PSII light (33,34), causing respective changes in electron transport capacity.
Differential Activation of Nuclear PSI Gene Promoters-To analyze the putative impact of chloroplast redox signals on the expression of the nuclear PSI genes PETH, PSAF, PSAD, and PETE, transgenic tobacco lines harboring the PSI promoter::GUS gene fusions were grown under PSI or PSII light (Fig. 2). The PETH promoter showed the same activity under all light sources tested. In contrast, the PSAF promoter exhibited higher activity in PSII light than in PSI light. 4 days under PSII light stimulates the promoter activity to the level detectable in continuous PSII light; however, no reduction in the activity was observed after a PSII3 I light shift. The PSAD promoter behaves similarly except that it exhibits lower activity under control white light. Similarly, the PETE promoter shows low activity under PSI light, high activity under PSII light, and activation after a PSI3 II light shift. Furthermore, in contrast to the other lines, the GUS activity declined after a PSII3 I light shift. Thus, at least three different response patterns to chloroplast redox signals can be distinguished.
Sucrose has been shown to repress photosynthesis gene expression (54) and to interact with chloroplast redox signals (30). Because we grew seedlings on a medium containing 1.35% sucrose, we tested a possible role of sucrose by transferring seedlings to a sucrose-free medium before shifting them to the other light regime. The GUS activity was determined after 96 h. We found no qualitative changes; however, in general the GUS activities were 20 -30% lower on the sugar-free medium (data not shown). Thus we conclude that sucrose has no effect on the light quality-dependent redox signals under our conditions.
Coupling of Nuclear Gene Transcription to Photosynthetic Electron Transport-The electron transport inhibitors DCMU and DBMIB inhibit electron flow before and behind the PQ pool, respectively, thus resulting in the oxidation or reduction of it (55). We sprayed these inhibitors exogenously onto the seedlings to show a coupling of the expression of PSI genes in the nucleus to the photosynthetic electron transport. Initially, we determined the optimal inhibitor concentrations by adjusting them such that their effects match the observed effects of the light sources on the Chl fluorescence parameters (Fig. 3). Treatment with DCMU (10 and 40 M) resulted in a decrease in the PSII value within 15 min. After 4 and 96 h these values decreased slightly, presumably because of a delayed inhibitory effect caused by the penetration of the substance. Detailed dose-response analyses uncovered that a concentration of 10 M DCMU results in a reduction of electron transport, which is identical to the effect of a PSII3 PSI light shift. In addition, a DBMIB concentration of 25 M caused a similar effect as a PSI3 PSII redox signal; however, the component is labile, and the photosynthetic parameters could only be maintained over the 96-h period when DBMIB was reapplied every 3-4 h. Spraying of control plants with the solvent alone had no effect on the photosynthetic electron flow or the expression of the reporter genes when compared with untreated plants (data not shown). With regard to the redox state of the PQ pool, we used the inhibitors as a reduction (DBMIB) or oxidation (DCMU) signal itself or as antagonists to the respective light-induced redox signals (Fig. 4). The tobacco lines harboring the PETH promoter construct did not respond to any of the inhibitor treatments (data not shown), which is consistent with the results obtained with the light regimes. The other three promoter constructs responded to the inhibitor treatments, however, in a differential way. In all instances, DCMU inhibited the promoter activation by a reduction signal of PSII light on PSI plants. In addition, the DCMU alone down-regulated the promoter activity of PSAD, PSAF, and PETE in PSII plants. Thus expression of these genes is coupled to photosynthetic electron transport in chloroplasts. The promoter activity of PSAD and PSAF is not down-regulated by PSI light in PSII plants, which is in contrast to the results obtained for PETE (compare with Fig. 2). The application of DBMIB in co-action with PSI light resulted in a down-regulation of PSAD and PSAF promoter activity. In contrast, the down-regulation of PETE promoter activity by the PSI light-induced oxidation signal was inhibited to over 50% by the DBMIB treatment. Application of the drug alone on PSI plants, however, resulted in no significant reaction in all three cases. These data indicate that the promoter activity of PETE is controlled by the redox state of the PQ pool, whereas PSAD and PSAF gene expression is redox-controlled; however, a clear initiator component within the chloroplast could not be identified so far. DISCUSSION We demonstrate that chloroplast redox signals participate in light regulation of nuclear PSI genes by using PS-specific light sources originally established for mustard (33). Our studies showed that tobacco seedlings represent a useful model organism for such studies. In addition, this is the first report describing the expression of nuclear PSI genes during photosystem stoichiometry adjustment. We show that redox signals previously found to control gene expression within chloroplasts also control gene expression in the nucleus. The four promoters analyzed in this study showed three different responses to the light sources: no response (PETH), activation (PSAF and  3. Effects of exogenously applied electron transport inhibitors on photosynthetic efficiency of tobacco. 14-day-old white-light-grown wild-type tobacco seedlings were treated with increasing concentrations (given on top) of DCMU or DBMIB, and the inhibitory effects on electron transport were quantified by measuring the Chl fluorescence parameter, PSII (effective quantum yield), at the indicated times after application. As control (C), the plants were treated in the same way with the solvent but without any inhibitor. The values represent means of three independent measurements. Variations were in the same range as shown in Table I. PSAD), and reversible activation/inactivation (PETE). An inhibitor assay with DCMU and DBMIB confirmed the coupling of the promoter response to photosynthetic electron transport. The inhibitor effects were optimized such that they are comparable with the effects of the light sources and that they persist over the whole period of experimentation, i.e. 96 h. The pulse amplitude-modulated fluorometer measurements demonstrate that the inhibitor concentrations mimic the light effects on photosynthetic electron transport. The optimal concentrations (10 M DCMU and 25 M DBMIB) differed from those used by Karpinski et al. (28); they used 4 and 14 M, which were effective after infiltration into Arabidopsis leaf discs. Because we sprayed the inhibitors on the cotyledons, the effective concentrations within the tissue are not known, but they must be considerably lower when compared with the concentration applied by the infiltration technique. 10 M DCMU, for instance, completely inhibits electron transport when applied to isolated chloroplasts (data not shown). DBMIB is a light-labile com-pound, and we found a complete loss of its inhibitory effect after 4 h. In liquid algae cell cultures its active time was reported to be between 6 and 24 h (56), and in Synechocystis PCC6803 cultures a decline of 60% in inhibition of O 2 evolution was reported after 1 h (57). This demonstrates that the optimal concentration depends on the application and organism. We maintained the DBMIB effect by repeated application of the drug, comparable with a recent study with Chlamydobotrys stellata (58).
The PETH promoter construct did not respond to any applied redox signal. This lack of response is not caused by the relatively low expression of the construct, because the same transgenic lines showed clear induction of the reporter gene when dark-grown seedlings were transferred to light (20), demonstrating the sensitivity of the construct for light regulation in general. Responsiveness to chloroplast redox signals could be located at a different level than transcription during PETH gene expression. Further experiments will show if such responses exist. In contrast the PSAD, PSAF, and PETE promoter constructs are activated by PSII light-generated reduction signals that can be blocked by a DCMU-generated oxidation signal. This oxidation signal alone is also able to reduce the activity of all three promoters, demonstrating that they can be regulated by both reduction and oxidation signals. A PSI light-induced oxidation signal reduced the PETE promoter-driven GUS activity but had no influence on the GUS activities driven by the PSAD and PSAF promoters. DBMIB treatment had only an effect in co-action with the PSI lightgenerated oxidation signal but not alone. In the cases of PSAD and PSAF this results in a decrease of the GUS activities as observed after DCMU treatment. This suggests that these two promoters respond to redox signals originating between the PQ pool and PSI or to the electron transport capacity in general. In contrast, because DBMIB prevents the PETE promoter deactivation induced by PSI light, the redox state of the PQ pool seems to regulate this promoter. The different response pattern of the promoters can be caused by different sensitivities to chloroplast redox signals. Activation by a dark/light transition of seedlings is a strong environmental signal and represents an off/on mechanism. PSI and PSII light-induced redox signals are a different type of environmental signal, and responses to them represent a modulation mechanism under persistent illumination. The additive effect of PSI light and DBMIB on PSAF and PSAD promoter activity suggests the existence of a threshold value in these cases that may be not reached when PSI light or DBMIB suppress electron transport alone. The signaling redox components remain unclear; however, the inhibitor experiments demonstrate that PSAF and PSAD promoter activity is clearly connected to the photosynthetic electron transport.
The chloroplast redox signals represent a new facet in the light regulation of nuclear photosynthesis genes. Results with dark-grown pea LIP1 and Arabidopsis COP1-4 mutants imply that the plastid-derived signal(s) is light-independent and not coupled to photosynthesis (13). The chloroplast redox signals investigated in this study are both light-dependent and coupled to photosynthesis, thus representing a new type of plastidderived signal. Physiologically, this light quality-dependent regulation of PSAD, PSAF, and PETE makes sense. The gene products of PSAD and PSAF are structural components of PSI. Thus, chloroplast redox signals adjust the expression of nuclear PSI genes to the overall need of the photosynthetic apparatus and might be a course-regulated process. The fine-tuning of PS stoichiometry adjustment is then achieved by the rapid and direct redox-regulated synthesis of the PsaA and PsaB reaction center subunits (33), which occurs in the same time scale as state transitions (59). In contrast, plastocyanin is not an inte-

FIG. 4. Effects of electron transport inhibitors in interaction with PS-specific light sources on PSI promoter activities.
Growth conditions and inhibitor treatments were as indicated across the top. The identity of gene promoters is given inside the panels. The respective promoter activation of PSII light on PSI plants (compare with Fig. 1, reduction signal) was arbitrarily set to 100%, and changes in promoter utilization induced by switches between light sources and inhibitor treatments were expressed in percent. The values represent the mean of three independent experiments performed with 3-4 parallels. grated component of PSI, and the fine tuning of its expression might occur differently, in particular because this polypeptide operates as an electron transfer component between the two photosystems. The present study provides new insights into chloroplast-nuclear signaling. Our current investigations focus on the impact of redox signals on other nuclear genes, the velocity of signal transduction in comparison to inner chloroplast redox signals, and the way how these signals are transduced.