Dehydroascorbate Reductase Affects Non-photochemical Quenching and Photosynthetic Performance*

Despite the efficiency of the reactions that convert solar energy into chemical energy, the capacity of photosynthesis to use absorbed light energy is limited. Excess absorbed light energy can be dangerous because it can result in the production of triplet state chlorophyll (³Chl*)² that can transfer energy to ground state O₂ to produce singlet oxygen (¹O₂*) and eventually other reactive oxygen species (ROS), e.g. superoxide anion, hydroxyl radicals, and hydrogen peroxide (H₂O₂) (1, 2). ROS can damage protein subunits and pigments of photosystem II (PSII), resulting in protein degradation, inactivation of reaction centers, and inhibition of the subsequent repair mechanisms of the reaction center (3, 4).

Thermal dissipation of excitation energy is described as non-photochemical quenching (NPQ) of chlorophyll fluorescence. NPQ is composed of pH-dependent fluorescence quenching (qE or feedback de-excitation), quenching associated with state transition (qT), and quenching caused by photoinhibition (qI). Under many conditions, most of NPQ occurs as qE in the PSII antenna pigment bed and is rapidly reversible (5, 6). qT contributes little to NPQ and is unlikely to be a significant factor in photoprotection. qI is quenching that includes damage to PSII reaction centers (7) and is either irreversible or slowly reversible (8). Under normal growth conditions, qE predominates, dissipating excess excitation energy as heat to prevent photoinhibition and damage to the photosynthetic apparatus. Thus, NPQ plays a key role in regulating light harvesting and photosynthetic performance (5).

Asc, the most abundant antioxidant in plants, is important in maintaining photosynthetic function (9–12). Asc detoxifies ROS and protects the photosynthetic apparatus from oxidative damage (1). Asc can serve as a direct electron donor to PSI and PSII under conditions where the primary electron donor system is impaired, e.g. under high light stress (13–15). Photoreduction of monodehydroascorbate (MDHA), produced following the oxidization of Asc, plays an important role in maintaining the electron transport flow when NAPD⁺ is limiting by competing with ferredoxin-NADP⁺ for electrons at the reducing side of PSI (11–12). Asc is also an essential cofactor for violaxanthin de-epoxidase that de-epoxidates violaxanthin to produce zeaxanthin in the xanthophyll cycle, a necessary step for the induction of NPQ following exposure to excess light (16, 17). The importance of Asc in protecting photosynthetic function has been shown with vtc mutants of Arabidopsis. The vtc1 mutant, defective in GDP-mannose pyrophosphorylase, accumulates 30% of the wild-type level of Asc and is hypersensitive to ozone, sulfur dioxide, or UVB light but exhibits only a moderately slower growth rate under normal growth conditions (18–20). vtc1 plants exhibit no significant difference in light saturation curves for CO₂ assimilation or chlorophyll fluorescence under these growth conditions, suggesting that the effect on growth was not because of decreased photosynthetic capacity or decreased photochemical efficiency (20). In contrast, the vtc2 mutant, which contains only 10–25% of wild-type levels of Asc as a result of reduced GDP-l-galactose-hexose-1-phosphate guanyltransferase activity, exhibits reduced NPQ and increased photoinhibition when transferred from low to high light (20–22).

Once used, Asc is oxidized to MDHA that can be recycled to Asc by ferredoxin in the chloroplast or by MDHA reductase in the chloroplast or cytosol (23, 24). In the thylakoid lumen, the MDHA produced by violaxanthin de-epoxidase or following donation of electrons from Asc to PSI or PSII is not available for reduction by ferredoxin or MDHA reductase but rapidly disproportionates to Asc and dehydroascorbate (DHA) when the pH of the lumen is low (15, 24). DHA is then reduced to Asc by dehydroascorbate reductase (DHAR) in a reaction requiring glutathione. DHA undergoes irreversible hydrolysis to 2,3-diketogulonic acid if not rapidly recycled to Asc.

DHAR is expressed in rate-limiting amounts and contributes to the regulation of the symplastic and apoplastic Asc pool size and redox state (25–27). Guard cells in DHAR-overexpressing plants exhibited a reduced level of H₂O₂, decreased responsiveness to H₂O₂ or abscisic acid signaling, and greater stomatal opening (26). Suppression of DHAR expression resulted in higher levels of H₂O₂ in guard cells and enhanced stomatal closure under normal growth conditions or following water stress. Increasing DHAR expression also provided enhanced tolerance to environmental ROS, e.g. ozone, despite the increase in stomatal conductance (27). Plants suppressed in DHAR expression exhibited a reduced rate of CO₂ assimilation with slower growth and reduced biomass accumulation (28). Altering DHAR expression affected the level of ribulose-bisphosphate carboxylase/oxygenase (Rubisco) in aging leaves but not in young leaves (28), suggesting that processes other than Rubisco contribute to the observed changes in photosynthesis.

Although the effect that the Asc pool size has on photosynthetic performance has been investigated (8, 20, 29), little is known about how DHAR expression might affect photosynthetic function. In this work, we examined how changes in the level of DHAR affected NPQ and photoinhibition. Reducing DHAR expression reduced qE and increased qI. Increasing DHAR expression resulted in an increase in the xanthophyll pigment and chlorophyll pool size as well as the rate of CO₂ assimilation, particularly at high light intensities, whereas the levels of ROS and photoinhibition following exposure to high light were reduced. The appropriate induction of NPQ in DHAR-suppressed leaves was restored following feeding with Asc, suggesting that the impaired induction of NPQ was a consequence of reduced foliar Asc. Our findings suggest that the level of DHAR expression affects the appropriate level of induction of NPQ and photoprotection.

EXPERIMENTAL PROCEDURES

Plant Material and Growth Conditions—Transgenic tobacco (Nicotiana tabaccum cv. Xanthi) expressing the His-tagged wheat DHAR from the cauliflower mosaic virus 35S promoter was generated using Agrobacterium tumefaciens as described previously (25). Transgenic tobacco plants suppressed for DHAR were identified following the introduction of the tobacco DHAR cDNA in pBI101. All plants were grown in commercial soil in a greenhouse supplied with charcoal-filtered air. Plants were grown under natural light conditions in a 10-h light and 14-h dark cycle and were watered to saturation twice a day. The average temperature during the day was 25.9 ± 0.6 °C, and the average temperature at night was 20.2 ± 0.5 °C. The average light intensity in the morning (9 a.m.) was 514 ± 206 μmol m^–2 s^–1 and in the afternoon (1 p.m.) was 1191 ± 244 μmol m^–2 s^–1. Leaves at the 2-week-old stage were taken from 6-week-old plants; leaves at the 8-week-old stage were collected from 12-week-old plants.

For high light experiments, 2- or 8-week-old leaves that had been dark-adapted were floated on ice-containing water and exposed to a photon flux density (PFD) of 2000 μmol m^–2 s^–1 (supplied from high output sodium light) for the times indicated. F_o (minimum fluorescence in the dark-adapted state) and F_m were measured immediately before the high light exposure and at time points during recovery. For the analysis, leaves with similar initial F_v/F_m values were used.

For Asc feeding experiments, 2-week-old leaves were floated on 10 mm Asc or water for 2 h in the dark. Following dark adaptation, NPQ was measured following exposure to 1500 μmol m^–2 s^–1.

Gas Exchange and Fluorescence Measurements—Gas exchange and fluorescence measurements were performed using the LI-COR Li-6400 portable photosynthesis system (LI-COR Biosciences, Lincoln, NE) with the LI-6400-40 leaf chamber, a relative humidity of 50%, and ambient level of CO₂. Fluorescence measurements were taken using overnight dark-adapted leaves. At the start of each experiment, the leaf was exposed to 2 min of far-red illumination (1μmol of photons m^–2 s^–1) for the determination of F_o. Saturating pulses (0.8 s) of 7000 μmol of photons m^–2 s^–1) were applied to determine the F_m or F_m′ values. Actinic light, consisting of 90% red light (λ = 630 ± 20 nm) and 10% blue light (λ = 470 ± 20 nm), was provided by light emission diode sources. F_s is the steady state fluorescence yield during actinic illumination. F_o′ (minimum fluorescence in the light-adapted state) was determined in the presence of far-red (λ = 740 nm) light after switching off the actinic light. A total of four to six samples were measured in each experiment. All data presented were calculated from at least three independent measurements. Conventional fluorescence nomenclature was used (30). NPQ was calculated from (F_m – F_m′)/F_m′, the quantum efficiency of PSII (ϕPSII) was calculated from (F_m′ – F_s′)/F_m′, qP was calculated from (F_m′ – F_s)/(F_m′ – F_o′), and the electron transport rate (ETR) was calculated from ϕPSII × f × α_leaf where f is the fraction of absorbed quanta that is used by PSII and is typically assumed to be 0.5 for C3 plants; α_leaf is leaf absorbance, and 0.9 was used for tobacco. NPQ_f and NPQ_s were determined as described previously (31).

H₂O₂ Assay—Foliar H₂O₂ concentrations were determined by the method of Gay and Gebicki (32) based on the peroxide-mediated oxidation of Fe²⁺ followed by the reaction of Fe³⁺ with xylenol orange (o-cresolsulfonphthalein 3′,3″-bis(methylimino)diacetic acid, sodium salt). Protein was removed from the leaf extract by adding an equal volume of 25 mm HCl. Protein-depleted extract was mixed with an equal volume of 2× assay reagent (500 μm ammonium ferrous sulfate, 220 mm HClO₄, 200 μm xylenol orange, 200 mm sorbitol). Absorbance of the Fe³⁺-xylenol orange complex (A₅₆₀) was detected after 45 min. The specificity for H₂O₂ was tested by removing H₂O₂ from the reaction mixture using catalase prior to protein removal. Standard H₂O₂ curves were obtained for each independent experiment by adding various amounts of H₂O₂ to 500 μl of assay reagent.

Chlorophyll and Pigment Measurements—Chlorophylls a and b were measured spectrophotometrically as described previously (33). Leaf samples were ground in liquid nitrogen and extracted with 90% (v/v) acetone. The absorbance at 664 and 647 nm was determined and used to calculated chlorophyll a and b content by the equations: Chl a = 11.93A₆₆₄ – 1.93A₆₄₇ and Chl b = 20.36A₆₄₇ – 5.50A₆₆₄, respectively. Each experiment was repeated two to three times, and representative results are presented.

Xanthophyll pigments were extracted with 100% acetone under dim light and were separated on a Spherisorb ODS-1 column (Alltech) as described by Gilmore and Yamamoto (34) using solvent A-1 (acetonitrile, methanol, 0.1 m Tris-HCl pH 7.5 (72:8:3)) and solvent B (methanol:hexane (4:1)). Pigments were identified and quantified by the retention time and amount of standards using a photodiode array detector.

Western Analysis—Protein extracts were resolved using standard SDS-PAGE, and the protein was transferred to 0.22-μm polyvinylidene difluoride membrane by electro-blotting. Following transfer, the membranes were blocked in 5% milk in TPBS (0.1% Tween 20, 137 mm NaCl, 2.7 mm KCl, 10 mm Na₂HPO₄, 1.4 mm KH₂PO₄, pH 7.4) followed by incubation with antiserum raised against DHAR, PsbS, light-harvesting complex protein (LHCP), heat shock protein 70 (HSP70), or eukaryotic elongation factor 1A (eEF1A) diluted 1:1000 in TPBS with 1% milk for 1.5 h. The blots were then washed twice with TPBS and incubated with goat anti-rabbit horseradish peroxidase-conjugated antibodies (Southern Biotechnology Associates, Inc.) diluted 1:20,000 for 1 h. The blots were washed twice with TPBS, and the signal was detected using chemiluminescence (Amersham Biosciences).

Asc and DHA Measurements—Asc was measured as described previously (9). Leaves were ground in 2.5 m HClO₄ and centrifuged at 13,000 rpm for 10 min. Two volumes of 1.25 m Na₂CO₃ were added to the supernatant, and following centrifugation, 100 μl was added to 895 μl of 100 mm K₂HPO₄/KH₂PO₄, pH 5.6. Asc was determined by the change in absorbance at 265 nm following the addition of 0.25 unit of ascorbate oxidase. The total amount of reduced and oxidized ascorbic acid (i.e. Asc and DHA) was determined by reducing DHA to Asc (in a reaction containing 100 mm K₂HPO₄/KH₂PO₄, pH 6.5, 2 mm GSH, and 0.1 μg of recombinant wheat DHAR protein incubated at 25 °C for 20 min) prior to measuring Asc. The amount of DHA was determined as the difference between these two assays.

RESULTS

Photosynthetic Activity and NPQ Decline Concomitantly during Leaf Aging—To characterize the photosynthetic activity of tobacco leaves during development, a light-response curve, representing the CO₂ assimilation rate as a function of incident PFD, was determined. 2-week-old leaves, i.e. the first fully expanded leaves, exhibited a high rate of CO₂ assimilation with a maximum rate achieved at a PFD of 1500 μmol m^–2 s^–1 (supplemental Fig. 1A). A progressive decline in the CO₂ assimilation rate and in the light saturation point was observed with leaf age (supplemental Fig. 1A). The substomatal CO₂ concentration in 8-week-old leaves (i.e. defined here as the presenescent stage) was higher than in 2-week-old leaves throughout the light-response curve (supplemental Fig. 1C), suggesting that the decline in CO₂ assimilation was because of decreased photosynthetic activity.

Chlorophyll fluorescence was also measured in dark-adapted leaves at the same developmental time points to determine which aspects of photosynthetic functioning change during leaf aging. The reduction state of Q_A, the primary quinone electron acceptor of PSII, was calculated from 1 – qP and plotted as a function of illumination time as an indicator of the efficiency of electron transfer from Q_A to downstream electron acceptors. During leaf aging, the Q_A pool shifted toward a more reduced state, i.e. higher 1 – qP values (supplemental Fig. 2A), indicating reduced electron transfer from Q_A to downstream electron acceptors and inversely correlating with the rate of CO₂ assimilation (supplemental Fig. 1A). A reduction in NPQ was also measured during leaf aging (supplemental Fig. 2B), suggesting a reduced capacity to dissipate excitation energy through high energy state quenching, which can lead to over-reduction of Q_A, production of singlet and triplet state chlorophyll a, and an enhanced production of ROS (3, 35).

Damage to chlorophyll a is indicated by a reduction in the maximum quantum yield of PSII (i.e. F_v/F_m) or an increase in the F_o value of dark-adapted leaves. A decrease in the F_v/F_m ratio (p < 0.01, n = 4) and a significant increase in the F_o value (p < 0.01, n = 4) was observed in 8-week-old leaves compared with younger leaves (supplemental Table 1), indicating that presenescent leaves may have accumulated photodamage in the light-harvesting antennae chlorophyll as well as in the PSII reaction center at this stage. Fast relaxing quenching (i.e. NPQ_f), a measure of qE, declined substantially in 6- and 8-week-old leaves (p < 0.001, n = 4), whereas slow relaxing quenching (i.e. NPQ_s), a measure of qI, increased significantly (p < 0.05, n = 4) (supplemental Table 1), suggesting a reduction in photoprotection during leaf aging.

Changes in DHAR Expression Affect Photosynthetic Function—To investigate whether the level of DHAR expression affects photosynthetic performance, the rate of CO₂ assimilation was measured in DHAR-overexpressing (D_OX) leaves that exhibit enhanced Asc recycling (25), in DHAR-suppressed (D_KD) leaves that exhibit impaired Asc recycling (26), and in control leaves. The rate of CO₂ assimilation in young (i.e. 2-week-old) D_KD leaves was significantly lower than the control (p < 0.005, paired t test) at PFD levels above 500 μmol m^–2 s^–1 (Fig. 1A); this was not because of reduced stomatal conductance (Fig. 1C) as the substomatal CO₂ concentration was higher than in the control (Fig. 1E). It also was not because of a decrease in Rubisco as its level in D_KD leaves is identical to that in the control at this leaf age (28).

Although the rate of CO₂ assimilation was lower in presenescent (i.e. 8-week-old) leaves in all three lines, it remained higher in D_OX leaves (p < 0.01, paired t test) and lower in D_KD leaves than in the control (p = 0.013, paired t test) (Fig. 1B). The reduction in photosynthetic function in presenescent D_KD leaves was not because of its reduced stomatal conductance (Fig. 1D) as the substomatal CO₂ concentration was higher than in the control (Fig. 1F). The minimum light intensity that revealed a difference in CO₂ assimilation between D_KD and control leaves declined from 600 μmol m^–2 s^–1 in 2-week-old leaves to 200 μmol m^–2s^–1 in 8-week-old leaves (Fig. 1B), indicating that a reduction in DHAR expression had a more profound effect on photosynthetic function in older leaves. These data suggest that a higher level of DHAR expression increases the capacity of a leaf to harness light energy for CO₂ fixation, whereas reducing DHAR expression decreases the capacity for CO₂ assimilation, particularly at high PFD intensities.

The ϕPSII was reduced in 2-week-old D_KD leaves (p < 0.001, paired t test) but was similar or slightly higher in D_OX leaves relative to the control (Fig. 2A). Although ϕPSII efficiency declined in 8-week-old leaves in all three lines, it remained lower in D_KD leaves (p < 0.05, paired t test) and was similar or slightly higher in D_OX leaves relative to the control (Fig. 2B). The maximum quantum yield of PSII in dark-adapted D_KD leaves (ϕPSII at time 0) decreased from 0.78 in young leaves (Fig. 2A) to 0.53 in presenescent leaves (Fig. 2B), a reduction not observed in presenescent control or D_OX leaves (Fig. 2B). Because a reduction in the maximum quantum yield of PSII is indicative of damage, these data suggest that a greater level of photoinhibition accumulates during the aging of leaves with reduced DHAR expression.

Young and presenescent D_KD leaves had a substantially lower ETR (p < 0.001, paired t test), whereas ETR in the corresponding D_OX leaves was consistently higher than in the control (p < 0.05, paired t test) (Fig. 2, C and D). The reduction state of Q_A in 2-week-old D_KD leaves was substantially higher than that in the control, whereas no significant difference was observed between D_OX and control leaves (Fig. 2E). In 8-week-old leaves, the Q_A pool was more reduced in all lines, but it was even more reduced in D_KD leaves than in the control (Fig. 2F). These data indicate a reduction in electron transport during leaf aging and that reducing DHAR expression results in a consistent increase in the proportion of closed PSII centers and reduced PSII efficiency.

Measurements of NPQ at moderate to high PFD revealed a significant reduction in 2-week-old D_KD leaves relative to the control (p < 0.005, paired t test) (Fig. 2G), suggesting that reducing DHAR expression reduces feedback de-excitation under saturating light conditions. In 8-week-old leaves, NPQ reached a higher level than the control in D_KD leaves at light intensities below 200 μmol m^–2 s^–1 but reached a markedly lower maximum level at high PFD. This resulted in a unique NPQ curve as a function of light intensity (Fig. 2H), suggesting that a reduction in DHAR expression reduces the maximum capacity of a leaf to induce NPQ.

The Asc Pool Size and Redox State of the Chloroplast Are Affected by Changes in DHAR Expression—Changes in DHAR expression alter the foliar Asc pool size and redox state (25–28). The change in photosynthetic function in D_KD leaves suggested that suppression of DHAR expression may have affected the Asc pool size and redox state specifically in the chloroplasts of D_KD leaves. Therefore, the level of Asc and DHA was measured in the chloroplast fraction of D_KD and D_OX leaves to determine whether the chloroplast Asc pool size and redox state were altered as a consequence of changes in DHAR expression. In the chloroplast fraction of D_KD leaves, the Asc pool size was smaller and the redox state was lower than in wild-type (WT) leaves (Table 1). In contrast, the Asc pool size was larger and the redox state was higher in the chloroplast fraction of D_OX leaves than in WT leaves (Table 1). These changes in the Asc pool size and redox state in D_KD or D_OX leaves correlate with the observed changes in photosynthetic function. The Asc redox state in the chloroplast was more reduced than that of the leaf as a whole, which includes the apoplast (Ref. 28 and see Fig. 7 below), suggesting the importance of maintaining Asc in a reduced state.

The last step in Asc biosynthesis occurs in mitochondria from which it is transported throughout the cell (36–39). Although all members of the gene family encoding DHAR in tobacco have not been identified, DHAR is encoded by a highly conserved, five-member nuclear gene family in Arabidopsis, which includes members targeted to the cytosol and chloroplast (40). DHAR in the chloroplast serves to recycle Asc once it has been oxidized, e.g. during periods of high photosynthetic activity such as high light stress. Because Asc is transported throughout the cell, the observed changes in the chloroplast Asc pool size and redox state of D_KD or D_OX leaves may have resulted from changes in cytosolic DHAR activity, which would alter Asc levels throughout a cell, or from changes in the level of DHAR present in the chloroplast. Although the expression of the cytosolic wheat DHAR (i.e. TaDHAR) in D_OX leaves would not be expected to result in changes in the level of DHAR in the chloroplast, the suppression of DHAR achieved using a tobacco DHAR gene encoding a cytosolic form of tobacco DHAR (i.e. NtDHAR) to induce RNA silencing might be expected to suppress the expression of chloroplast-targeted as well as cytosolic DHAR proteins. To investigate this, the level of NtDHAR present in the chloroplast of D_KD or D_OX leaves was measured using Western analysis. A substantial reduction in the level of NtDHAR present in the chloroplast and cytosolic fractions of 2- and 8-week-old D_KD leaves was observed relative to the WT (Fig. 3A). Little decrease in the level of LHCP or chloroplast-localized HSP70 (cHSP70), which were used as loading controls, was detected in 2-week-old leaves (Fig. 3A). In 8-week-old leaves, the cHSP70 loading control suggested an even greater reduction in the level of NtDHAR present in D_KD chloroplasts (Fig. 3A). Little change in the level of NtDHAR was observed in the chloroplast fraction of 2- or 8-week-old D_OX leaves, and the cytosolic TaDHAR expressed in the cytosolic fraction of D_OX leaves, which is distinguished from the endogenous NtDHAR by its larger molecular weight, was absent from the chloroplast fraction (Fig. 3A), demonstrating a lack of cytosolic contamination of the chloroplast fraction. These data indicate that the suppression of DHAR expression in D_KD leaves substantially reduces the level of DHAR present in the chloroplast, which may contribute to the smaller Asc pool size and lower redox state observed in Table 1. As the level of NtDHAR present in the chloroplast of D_OX leaves was unchanged, the larger Asc pool size and higher redox state in the chloroplast of these leaves are likely because of the expression of cytosolic TaDHAR, previously shown to increase the Asc pool size and redox state in the cytosol and apoplast (27).

A Reduction in DHAR Expression Results in Increased Photoinhibition—Because of the effect that reducing DHAR expression had on NPQ in presenescent D_KD leaves, a kinetics analysis of NPQ in dark-adapted leaves exposed to a saturating level of light (i.e. 1500 μmol m^–2 s^–1) was performed. Initially NPQ increased more rapidly in 8-week-old D_KD leaves than in the control, but it reached a substantially lower steady state level (p < 0.05, paired t test) (Fig. 3C) similar to the induction of NPQ as a function of light intensity (Fig. 2H). The induction of NPQ was also lower in 2-week-old D_KD leaves relative to the control but to a smaller extent (Fig. 3B). Rapid kinetics analysis confirmed the faster initial induction of NPQ in 8-week-old D_KD leaves (Fig. 4B). The more rapid initial development of NPQ in 8-week-old D_KD leaves suggests either a slower relaxation of NPQ during dark adaptation and/or an enhanced initial ability to induce NPQ following exposure to light, whereas the lower maximum NPQ level achieved suggests a limited capacity to fully induce NPQ. The Q_A pool remained more oxidized in presenescent D_OX leaves and more reduced in D_KD leaves than in the control (Fig. 3E) as it did in young leaves of the same lines (Fig. 3D), supporting the notion that photochemical efficiency was also affected by a reduction in DHAR expression.

PsbS is essential for feedback de-excitation, and the level of NPQ achieved is proportional to its level of expression (41). To investigate whether changes in NPQ in D_KD leaves correlated with PsbS expression, its level was determined. The level of PsbS in 2-week-old leaves of the three lines was similar to the levels in 8-week-old control and D_OX leaves (Fig. 4C). In contrast, the level of PsbS was elevated in presenescent D_KD leaves, correlating with the faster initial induction of NPQ in these leaves. The level of LHCP in presenescent D_KD leaves relative to the WT was unchanged, demonstrating the increase in PsbS relative to this control protein in D_KD leaves (Fig. 4C).

To determine whether the effect that changes in DHAR expression had on NPQ was because of changes in qE or qI, fast and slow relaxation analysis was performed. 8-week-old D_KD leaves exhibited a significant increase in NPQ_s (p < 0.05, n = 4) and a significant decrease in NPQ_f (p < 0.05, n = 4) relative to the control (Table 2). These data indicate that more photoinhibition occurred in D_KD leaves during the period of light adaptation because a smaller fraction of total NPQ relaxed as NPQ_f, whereas a larger fraction relaxed as NPQ_s. Moreover the significantly higher F_o value in D_KD leaves relative to the control (p < 0.005, n = 4) (Table 2) correlated with a significant reduction in the maximum quantum efficiency of PSII (Fig. 2B), suggesting that reducing DHAR expression increased damage to antenna chlorophyll. The chlorophyll pool size disproportionately decreased in D_KD leaves by 8 weeks of age and was significantly smaller than that in the control (p < 0.005, n = 3) as was the chlorophyll a/b ratio (p < 0.05, n = 3) (Table 3), suggesting that chlorophyll a was preferentially lost in D_KD leaves during aging. The disproportionate decrease in qE as measured by NPQ_f and increase in qI as measured by NPQ_s in 8-week-old D_KD leaves suggests an accumulation of damage during aging.

As photoinhibition can result from ROS, the level of H₂O₂, a measure of ROS, was determined in D_KD leaves. The level of H₂O₂ was significantly higher in young D_KD leaves than in control or D_OX leaves (p < 0.05 and p < 0.005, respectively; n = 3) (Table 3). An even greater increase in H₂O₂ was observed in presenescent D_KD leaves relative to the control (p < 0.005, n = 3) (Table 3), indicating that the level of ROS disproportionately increases over time in D_KD leaves.

DHAR Activity Alters Xanthophyll Pigment Pool Size—The higher level of PsbS observed in presenescent D_KD leaves (Fig. 4C) suggested a higher potential for NPQ. However, despite an elevated induction of NPQ in D_KD leaves under low PFD (Fig. 2H) or during the initial induction of NPQ following exposure of dark-adapted leaves to light (Fig. 4B), induction of NPQ in D_KD leaves was impaired under moderate to high PFD (Fig. 2H) as a result of a lower qE component (Table 2), suggesting that the leaves had a lower maximum capacity to induce NPQ. In addition to PsbS and a pH gradient across the thylakoid membrane, the induction of feedback de-excitation requires pigments that comprise the xanthophyll cycle (35, 42, 43). Following exposure to light, violaxanthin undergoes de-epoxidation initially to antheraxanthin and finally to zeaxanthin in reactions that require ascorbate (35). Zeaxanthin, together with PsbS, is important in establishing the appropriate level of feedback deexcitation in response to excess light (41). To examine whether the changes in feedback de-excitation that were observed in leaves with altered DHAR expression involved changes to the xanthophyll cycle, those pigments comprising the cycle were measured in control, D_OX, and D_KD leaves.

As expected, the level of antheraxanthin and zeaxanthin was low in dark-adapted leaves but increased following exposure to light (Fig. 5). The pool size of violaxanthin, neoxanthin, lutein, and α-carotene in 2-week-old, dark-adapted D_KD leaves was 90.1, 66.3, 86.5, and 90.8%, respectively, of the control, but the level of antheraxanthin and zeaxanthin was 1.46- and 3-fold higher, respectively (Fig. 5 and supplemental Fig. 3). The larger pool size of antheraxanthin and zeaxanthin (i.e. relative to violaxanthin) in dark-adapted D_KD leaves may suggest slower epoxidation to violaxanthin. Exposure of 2-week-old D_KD leaves to light did not result in an increase in the zeaxanthin pool size, resulting in a zeaxanthin pool size in young, light-adapted D_KD leaves that was only 89% of that in control leaves (Fig. 5). This indicates that D_KD leaves are unable to generate the same amount of zeaxanthin as control leaves in response to light and correlates with the lower level of induction of NPQ in D_KD leaves in response to high PFD (Fig. 2G). In contrast, the level of antheraxanthin increased to a level similar to wild-type levels (Fig. 5). With the exception of an 11% increase in the level of lutein, the amount of the remaining pigments in young, light-adapted D_KD leaves was little changed relative to control leaves (supplemental Fig. 3).

In contrast to D_KD leaves, the pool size of violaxanthin, antheraxanthin, and zeaxanthin was substantially larger in 2-week-old, dark-adapted D_OX leaves than in the control (Fig. 5). Following exposure to light, young D_OX leaves contained a similar level of antheraxanthin and zeaxanthin as in control leaves (101 and 99%, respectively) despite a 29% larger violaxanthin pool size, indicating that the level of DHAR expression in WT leaves is sufficient for optimal conversion of violaxanthin to antheraxanthin and zeaxanthin.

A decrease in most of the xanthophyll pigments was observed during leaf aging (Fig. 5 and supplemental Fig. 3) consistent with a reduced ability to induce NPQ (supplemental Fig. 2B and supplemental Table 1). The pool size of these pigments was larger in 8-week-old, dark-adapted D_OX leaves than in the WT by 19–61% (Fig. 5 and supplemental Fig. 3). Following exposure to light, D_OX leaves contained a substantially larger antheraxanthin and zeaxanthin pool size (154 and 140%, respectively) than in the WT, representing increases that were disproportionately greater than the increase in the violaxanthin pool size (111% of the control) (Fig. 5). This suggests that the wild-type level of DHAR expression may become limiting with leaf age and that increasing DHAR expression maintains a higher level of xanthophyll cycle pigments during leaf aging, correlating with a slower, aging-related loss in qE.

In 8-week-old, dark-adapted D_KD leaves, the pool size of most of the pigments declined 22–45% relative to the WT with the only exception being antheraxanthin, which declined by 8% (Fig. 5). In light-adapted D_KD leaves, the level of violaxanthin was 88.2% of the WT, whereas the levels of antheraxanthin and zeaxanthin were not lower (99.5 and 115%, respectively, of the control) (Fig. 5). These data indicate significant perturbations to the pool size of several pigments in presenescent D_KD leaves, including lutein and neoxanthin, which are known to contribute toward photoprotection (44, 45).

DHAR Is Required for Recovery from High Light Stress—Because the induction characteristics of NPQ were affected by changes in DHAR expression, we examined whether the degree of photoinhibition following high light stress was similarly affected. The degree of photoinhibition was measured by the maximum quantum yield (i.e. F_v/F_m) that serves as a measure of stress-induced damage to PSII reaction centers. 2-week-old, dark-adapted control leaves exposed to high light for 5 h exhibited a substantial reduction in the maximum quantum yield that partially recovered over the next 48 h (Fig. 6A). 2-week-old D_KD leaves exhibited a greater degree of photoinhibition than the control both immediately following exposure to high light and throughout the recovery period (Fig. 6A). In contrast, 2-week-old D_OX leaves showed a response similar to that in the WT, although the reduction in F_v/F_m following high light stress was not as great as in control leaves, and F_v/F_m remained higher throughout the recovery period despite a similar initial F_v/F_m (Fig. 6A). The degree of photoinhibition observed inversely correlated with the Asc pool size and redox state (Fig. 7A). The Asc pool size in young D_OX leaves was larger and the Asc redox state was greater (i.e. more reduced) than in the control prior to and following high light stress, whereas the Asc pool size in young D_KD leaves was smaller and the Asc redox state was lower (i.e. more oxidized) than in the control prior to and following high light stress. The level of DHAR remained lower in the chloroplast and cytosolic fractions of 2-week-old D_KD leaves relative to the control following recovery from high light stress (Fig. 7C).

Similar results were obtained for 8-week-old leaves following exposure to high light for 3 h. D_KD leaves exhibited substantially more photoinhibition despite a similar initial F_v/F_m and had a markedly slower rate of recovery than did the WT (Fig. 6B). As observed in young leaves, 8-week-old D_OX leaves experienced less photoinhibition than the WT (Fig. 6B). The Asc pool size in presenescent D_OX leaves was larger and the Asc redox state was greater than in the control prior to and following high light stress, whereas the Asc pool size in presenescent D_KD leaves was smaller and the Asc redox state was lower than in the control prior to and following high light stress (Fig. 7B). These results indicate that the level of photoinhibition following high light stress inversely correlates with the level of DHAR expression.

DHAR Expression Affects High Light-induced Changes in Xanthophyll Pigments—Xanthophyll cycle pigments were then analyzed in 2- and 8-week-old leaves subjected to high light stress. In 2-week-old leaves of all three lines, the combined violaxanthin (V), antheraxanthin (A), and zeaxanthin (Z) (i.e. V + A + Z) pool size decreased moderately following exposure to high light (Fig. 8A). The V + A + Z pool size was slower to recover in D_KD leaves than in control or D_OX leaves (Fig. 8A). As expected, the de-epoxidation status of the xanthophyll cycle pigments (i.e. (A + Z)/(V + A + Z)) was low before exposure to the high light and increased substantially following exposure to the high light only to decline again during the subsequent recovery period that occurred in the dark. Only small differences in the de-epoxidation status among the three lines were observed (Fig. 8B). A transient decrease in the pool size of lutein and neoxanthin was observed in all three lines (Fig. 8, C and D, respectively). Whereas the pool size of lutein and neoxanthin quickly recovered in control or D_OX leaves, its rate of recovery was substantially slower in D_KD leaves. Exposure to high light stress did not substantially change the level of α-carotene in control or D_OX leaves, but its level declined in D_KD leaves substantially for the 1st h following exposure to the high light stress, and its level was slow to recover to the prestress amount (Fig. 8E). Exposure to high light did not substantially change the level of β-carotene in the three lines initially, but an increase in the level of β-carotene was observed by 24 h of recovery in control or D_OX leaves (Fig. 8F). No similar increase in the level of β-carotene was observed in D_KD leaves (Fig. 8F).

As with 2-week-old leaves, the de-epoxidation status of the xanthophyll cycle pigments in 8-week-old leaves was low before exposure to the high light, increased following high light exposure, and declined again during the subsequent recovery period (Fig. 8H). The level of lutein, neoxanthin, and α-carotene declined in all three lines following exposure to the high light, and in each case, recovery of the pigment pool size was most rapid in D_OX leaves (Fig. 8, I, J, and K, respectively). The level of β-carotene in control or D_OX leaves recovered by 24 h following the stress to a level that was greater than that before exposure to the stress, whereas its recovery in D_KD leaves was substantially lower (Fig. 8L). These results indicate that maintenance of the xanthophyll cycle and related pigments following exposure to a high light stress is affected, directly or indirectly, by the level of DHAR expression.

DHAR Expression Affects ROS Production during High Light Stress—To determine whether ROS production correlated with the changes in photoinhibition observed in leaves exposed to high light (Fig. 6), the level of H₂O₂ was measured in 2- and 8-week-old leaves that had been subjected to high light stress. Prior to the high light stress, the level of H₂O₂ was similar in 2-week-old leaves of all lines (Fig. 9A). As expected, the level of H₂O₂ increased in the three lines immediately following exposure to the high light but was lowest in D_OX leaves and highest in D_KD leaves (Fig. 9B). Following 24 h of recovery, the level of H₂O₂ had recovered to prestress levels (Fig. 9C).

Prior to the high light stress, the level of H₂O₂ in 8-week-old D_KD leaves was significantly higher than in the control (p < 0.01), whereas the level in D_OX leaves was similar to that in the control. The level of H₂O₂ in 8-week-old D_KD leaves further increased following exposure to high light, whereas much smaller increases were observed in 8-week-old D_OX leaves and WT leaves (Fig. 9B). Following 24 h of recovery, the level of H₂O₂ remained significantly higher in D_KD leaves than in the control (p < 0.05) (Fig. 9C). These results indicate that the level of ROS correlates with the level of photoinhibition following exposure to high light stress, and both inversely correlate with the level of DHAR expression. Moreover the data indicate that the ability to detoxify ROS following exposure to high light decreases with age in D_KD leaves.

Appropriate Induction of NPQ Can Be Restored in D_KD Leaves following Direct Asc Feeding—To examine whether the reduction in the Asc pool size was responsible for the defect in NPQ induction in D_KD plants, D_KD leaves were supplied with 10 mm Asc for 2 h in the dark. As a negative control, D_KD leaves were supplied with water for the same period of time. WT leaves were also supplied with 10 mm Asc or water for the same period of time. Following dark adaptation, the induction of NPQ in the treated leaves was measured following exposure to 1500 μmol m^–2 s^–1. 2-week-old leaves were used for the analysis to avoid any secondary effects, such as damage, that may have accumulated in aging D_KD leaves and that might complicate the analysis. In the absence of exogenous Asc, the induction of NPQ was impaired in D_KD leaves relative to WT leaves (Fig. 10), data in good agreement with those in Fig. 3B. Following feeding with Asc, the induction of NPQ in D_KD leaves increased substantially to a level that was similar to that in the WT (Fig. 10). Asc treatment did not substantially alter the induction of NPQ in WT leaves. These results demonstrate that the defect in NPQ induction in D_KD leaves can be complemented by exogenous Asc, suggesting that it is the decrease in the Asc pool size that is responsible for the impaired induction of NPQ in D_KD leaves.

DISCUSSION

In this study, we observed that the level of DHAR expression affects the induction of NPQ and the degree of photoinhibition experienced by a plant. Suppressing DHAR expression had a greater effect on feedback de-excitation, photosynthetic activity, and photoinhibition at high light intensities, indicating that DHAR activity is important under conditions of light saturation. The effect that DHAR had on these processes was accompanied by changes in the Asc pool size and redox state of the chloroplast, the pool size of chlorophyll and xanthophyll pigments, ROS, stomatal conductance, and, in presenescent leaves, Rubisco content. The reduction in photosynthetic activity in D_KD leaves was not because of a reduction in stomatal conductance as the substomatal CO₂ concentration was higher than that in WT leaves. Moreover although the levels of Rubisco and chlorophyll declined at an accelerated rate in D_KD leaves, neither was significantly different from the levels in WT for at least the first 2 weeks of leaf growth (Table 3) (28), and therefore, the reduced photosynthetic activity observed in young D_KD leaves could not be explained by changes in Rubisco content or chlorophyll pool size.

NPQ is induced in response to the exposure and absorption of light that exceeds the capacity of photochemistry to utilize the excitation energy (35). NPQ functions as a sensor of the reduction state of the photosystems to prevent the entry of excess light energy into the photosystems where it would result in their over-reduction and the generation of ROS (46, 47). Given the competitive nature of NPQ and qP, a reduction in photosynthetic activity would be expected to result in a greater induction of NPQ to dissipate a larger fraction of the excitation energy. Despite the reduced level of qP in D_KD leaves, the induction of NPQ was impaired in response to high PFD and failed to reach the level observed in wild-type leaves (Fig. 2G), suggesting a reduced capacity to induce feedback de-excitation. The impaired induction of NPQ, particularly under conditions of saturating light, did correlate with an increase in ROS (Table 3) and photoinhibition (Fig. 6 and Table 2) in D_KD leaves, which in turn correlated with a reduction in the maximum quantum yield of PSII, a measure of damage to the photosynthetic machinery (Fig. 2B). The reduced tolerance to high light stress that correlated with the impaired induction of NPQ could not be explained by differences in the maximum quantum efficiency of PSII prior to the onset of the stress particularly in young leaves (Fig. 6). The failure to induce NPQ to a level that would compensate for the reduced level of photochemistry, however, suggests an impaired response that resulted from the reduction in Asc that accompanied the suppression of DHAR. The observation that the impaired induction of NPQ in D_KD leaves could be complemented by exogenous Asc (Fig. 10) supports the notion that the reduced level of Asc in D_KD leaves is responsible for the impaired induction of NPQ.

To what extent the impairment in NPQ contributes to the photoinhibition observed in D_KD leaves is unknown. Mutants affecting components of NPQ in tobacco, such as zeaxanthin or PsbS, would be necessary to show the extent to which impaired induction of NPQ is responsible for photoinhibition following exposure to high light. Given the multiple roles that Asc plays in photosynthesis, it is likely that changes in the Asc level may affect more than one process. The effect that the reduction in the Asc pool size and redox state in D_KD leaves had on these processes is similar in many respects to those observed in the vtc2 mutant in which Asc biosynthesis is substantially reduced (8). D_KD tobacco (28) and vtc2 Arabidopsis, which contains 10–25% of the Asc level present in wild-type Arabidopsis (47, 48), grow more slowly and accumulate less biomass than the wild type (8). Like D_KD plants (26–28), the vtc2 mutant exhibits a lower Asc redox state in addition to smaller Asc pool size particularly when grown under high light (8, 21). The reduced ϕPSII efficiency observed in D_KD tobacco (Fig. 2) was also observed in vtc2 plants (8). The light-response curve for NPQ in the vtc2 mutant exhibits the same defects as that in young D_KD leaves (Fig. 2G) in that induction of NPQ is normal at low PFD but impaired at higher PFD when light becomes saturating (47). The vtc2 mutant is also deficient in qE, particularly during high light stress (17, 21, 49) as is D_KD tobacco, that increases as a function of leaf age (Table 2). The deficiency in qE in vtc2 plants is thought to be a result of substrate (i.e. Asc) limitation of the violaxanthin de-epoxidase enzyme responsible for conversion of violaxanthin to antheraxanthin and zeaxanthin (47). Moreover like D_KD tobacco, vtc2 plants exhibit impaired induction of NPQ under saturating light conditions, and both exhibit symptoms of photooxidative stress following exposure to high light as indicated by a lower F_v/F_m (Fig. 6 and Refs. 8 and 21, respectively), ϕPSII (Fig. 2 and Ref. 8, respectively), and photosynthetic rate (Fig. 1 and Ref. 8, respectively). A similar decrease in NPQ was observed in vtc1 Arabidopsis at high PFD (20). As observed in D_KD leaves (Fig. 10), Asc feeding to vtc2 mutant leaves restored full induction of NPQ (47). vtc2 plants undergo photobleaching and lipid peroxidation when transferred from low to high light (8, 21) similar to the increase in ROS (Fig. 9) and reduction in F_v/F_m (Fig. 6) in D_KD tobacco, indicating a reduced ability to detoxify ROS produced during high light stress thus resulting in a substantial degree of photoinhibition. Similarly D_KD tobacco and vtc1 and vtc2 mutants exhibit greater sensitivity to ozone (18, 27), supporting the notion of a limited ability to detoxify ROS. vtc2 plants grown under high light had a lower chlorophyll a/b ratio similar to that in D_KD tobacco even when the latter was grown at moderate light levels. No change in the level of LHCP, D1, Cytf, or PsaF was observed in vtc2 plants (8, 21) similar to the lack of change in the level of LHCP in D_KD tobacco (Fig. 4).

Some differences between D_KD plants and the vtc2 mutant, however, were observed. Although young D_KD tobacco exhibited a smaller pool size of zeaxanthin following light adaptation (Fig. 5), vtc2 Arabidopsis exhibited a slightly larger zeaxanthin pool size following adaptation to normal PFD (8) but a smaller pool size following adaptation to higher PFD (21). The level of ascorbate peroxidase activity, which uses Asc to reduce H₂O₂, was slightly higher in young D_KD tobacco and the vtc1 mutant (20, 27) but unchanged in vtc2 plants (8, 21). These minor differences may reflect differences between tobacco and Arabidopsis or the conditions used. More significant differences, however, exist between D_KD tobacco and vtc1 Arabidopsis or between the vtc1 and vtc2 mutants themselves. The level of Asc in the vtc1 mutant is higher than in the vtc2 mutant, accumulating to ∼30% of the wild-type level (18–20). Although it is hypersensitive to ozone and exhibits a slower growth rate under normal growth conditions (18–20), little change was observed in the CO₂ assimilation or chlorophyll fluorescence of the vtc1 mutant, suggesting that photosynthetic capacity or photochemical efficiency is unaffected in this mutant under normal growth conditions. The vtc1 mutant did not exhibit symptoms of photooxidative stress (20), whereas the vtc2 mutant and D_KD tobacco did. The more severe nature of the phenotypes observed in the vtc2 mutant suggests that a greater reduction in Asc than that present in the vtc1 mutant is required before photosynthetic activity, photosynthetic efficiency, and the level of photooxidative stress are affected in Arabidopsis at least under normal growth conditions. Although the level of Asc is not reduced in D_KD tobacco to the same extent as in the vtc mutants, the Asc redox state is affected to a greater extent by suppression of DHAR than it is by a reduction in Asc biosynthesis (8), which may account for the effect on photosynthetic processes when demand for Asc is high. It is also possible that photosynthetic processes in tobacco are more sensitive to changes in the level of Asc than they are in Arabidopsis. In fact, the level of foliar Asc in Arabidopsis is substantially higher (4.29 μmol/g fresh weight) than in tobacco (0.96 μmol/g fresh weight) when both are grown under moderate light, supporting the notion that a greater reduction in the level of Asc in Arabidopsis may be necessary to perturb photosynthetic processes than in tobacco. In fact, taking into account the respective decreases in Asc in D_KD and vtc mutant leaves, the level of Asc in D_KD leaves would be lower than that in vtc mutant leaves despite the greater percent decrease in the Asc pool size in the vtc mutants relative to the WT. The impaired induction of NPQ in D_KD and vtc2 plants suggests that a reduction in cellular Asc, whether through reduced biosynthesis or recycling, affects this protective process and results in increased photooxidative stress.

Although the npq1 mutant in Arabidopsis, which lacks zeaxanthin, exhibits increased sensitivity to photoinhibition, the npq1/vtc2 double mutant shows significantly less tolerance to high light than does either mutant alone, demonstrating the synergy between NPQ and Asc in providing protection against photoinhibition (21, 50). The relatively high level of Asc in Arabidopsis compared with tobacco may partially compensate for the loss of qE in the npq1 mutant and thus mask the full effect of high light stress in this species until the level of Asc is substantially reduced, such as in the vtc2 mutant.

The failure to induce NPQ fully in response to saturating light suggested possible alterations to the xanthophyll cycle. An abnormally large pool size of zeaxanthin was observed in young, dark-adapted D_KD leaves, but a lower pool size was observed in light-adapted leaves relative to control leaves. As zeaxanthin is required for the induction of NPQ, its changes may have contributed to the phenotypes observed in D_KD leaves. As zeaxanthin can also function as an antioxidant (51) and functions synergistically with Asc to provide photoprotection (21), the reduction in the zeaxanthin pool size in light-adapted D_KD leaves, in conjunction with the reduced Asc pool size, may have contributed to the increase in photooxidative stress observed in these plants. Additional perturbations to other pigments such as lutein and neoxanthin, which are known to be involved in photoprotection (44, 45), may also have contributed to the alterations in NPQ and photooxidative stress.

The observation that a reduction in DHAR expression can affect induction of NPQ was supported by the observation that an increase in DHAR expression had the opposite effect. A larger xanthophyll pigment pool was present in D_OX leaves (Fig. 5) as was a reduced oxidative load (Table 3) and a higher tolerance to high light stress (Fig. 6). Although suppression of DHAR in D_KD leaves included suppression of DHAR present in the chloroplast, increasing DHAR in D_OX leaves was limited to the overexpression of cytosolic DHAR. Following the last step of its biosynthesis in the mitochondrion (36, 39), Asc is transported throughout the cell by means of specific Asc and DHA transporters, and Asc entry into the chloroplast occurs through facilitated diffusion (37, 38). The increase in the Asc pool size and redox state in D_OX chloroplasts demonstrated that changes in cytosolic DHAR can affect Asc levels in the chloroplast and is similar to the effect such changes have on apoplastic Asc (27). A reduction in the apoplastic Asc level was also observed in the vtc1 mutant (20), further evidence that changes to the level of cytosolic Asc affects the Asc pool size in other cellular compartments.

By the presenescent stage, D_KD leaves have a markedly reduced capacity to induce NPQ in response to high light (Fig. 2H). The reduced level of Rubisco observed in presenescent D_KD leaves (28) may have contributed to the lower efficiency of photochemical quenching and the greater reduced state of Q_A. However, such conditions would be expected to induce NPQ. The failure of NPQ to be induced to even wild-type levels supports the notion that the reduction in DHAR expression resulted in an impaired response to high light. qE and qI both contribute to the level of NPQ that is induced during exposure to light. The impaired induction of NPQ in presenescent D_KD leaves was largely a result of a reduction in qE as indicated by a lower NPQ_f. A significant increase in qI, as indicated by an increase in NPQ_s, suggested that photoinhibition had occurred during leaf aging. This conclusion is consistent with the observed reduction in F_v/F_m, indicating irreversible damage to PSII centers; higher F_o values suggesting changes in the configuration of the antenna light-harvesting complex that can destabilize chlorophyll binding and disrupt energy transfer from antenna chlorophylls to the PSII core complex, hypersensitivity to high light stress, and a substantial increase in ROS. Impaired induction of NPQ in presenescent D_KD leaves in response to high PFD was largely a consequence of high F_m′ values, indicating that energy transfer to xanthophyll pigments such as zeaxanthin, which is present at near wild-type levels in light-adapted, presenescent D_KD leaves, may be impaired. The higher level of PsbS in presenescent D_KD leaves (Fig. 4C) may represent an attempted compensatory response to the impaired induction of qE and increase in qI. These results are supported by the lack of change in NPQ_f and NPQ_s in young D_KD leaves, indicating little difference in photoinhibition at this developmental stage, a result consistent with the larger Asc pool size present in young leaves and the accumulation of damage over time (26, 38).

During leaf aging, aspects of the observed phenotype in D_KD leaves are similar to those observed in WT leaves of significantly older age, suggesting the possibility that loss of DHAR activity may lead to an acceleration of senescence. Conflicting observations concerning whether senescence is accelerated or delayed have been reported for the vtc mutants (20, 52). However, premature senescence of young D_KD leaves does not appear to be occurring as their chlorophyll content (Table 3) and level of Rubisco (28) are not substantially different from those in WT leaves, and yet they exhibit impaired induction of NPQ (Fig. 2G). Moreover the ability of exogenous Asc to restore the appropriate induction of NPQ does not support premature senescence as a cause of the impaired induction of NPQ in young D_KD leaves. Other aspects of D_KD leaves are also not accounted for by accelerated senescence. For example, the level of LHCP declines during leaf aging but remains unchanged in D_KD leaves relative to WT leaves of similar age (Fig. 4C). Therefore, not all aspects of an accelerated senescence program appear to be operative in D_KD leaves.

Because of the similar phenotypes exhibited by D_KD tobacco and vtc2 Arabidopsis, those processes affected by reduced DHAR expression are likely to be the result of changes to Asc, which would be expected to have pleiotropic effects. In addition to those processes examined in this study, a reduction in Asc has been reported to affect hormone signaling, the distribution of enzymes involved in ROS scavenging, and gene expression (20, 53, 54). Like many cases in which global changes in gene expression have been observed, the contribution that such changes make to the vtc1 phenotype has not been determined, and the observed changes in hormone signaling have not been demonstrated to be responsible for the phenotypes associated with the vtc1 mutation (53, 54). The observation that the vtc2 mutation affects NPQ and photooxidative stress whereas the vtc1 mutation largely does not demonstrates that the extent to which Asc biosynthesis is inhibited determines which processes are affected. However, the observation that Asc treatment of D_KD leaves complemented the defect in NPQ induction suggests that the reduction in Asc pool size was responsible for the impaired induction of NPQ and not a result of a long term adaptation to the reduction in Asc. The difference in the phenotypes observed with the vtc1 and vtc2 mutants also suggests that, like genes involved in Asc biosynthesis, DHAR may ensure the appropriate induction of NPQ through a combination of direct and indirect effects that Asc would be expected to have on photosynthetic functioning, ROS detoxification, and the xanthophyll cycle. Thus, despite the many functions of Asc in photosynthesis and in growth and development, our results suggest that the recycling of Asc by DHAR can be as important to the appropriate induction of NPQ and protection against photooxidative stress as the biosynthesis of Asc.