Interactions between β-carotene (β-C) and Chl a turnover were investigated in relation to photoinhibition and D1 protein turnover in mature leaves of Arabidopsis (Arabidopsis thaliana) by 14CO2 pulse–chase labeling. Following a 2 h treatment of leaves with water, lincomycin (Linco; an inhibitor of chloroplast protein synthesis) or norflurazon (NF; an inhibitor of carotenoid biosynthesis at phytoene desaturation) in the dark, 14CO2 was applied to the leaves for 30 min under control light (CL; 130 μmol photons m–2 s–1) conditions, followed by exposure to either CL or high light (HL; 1,100 μmol photons m–2 s–1) in ambient CO2 for up to 6 h. Under both light conditions, 14C incorporation was strongly decreased for Chl a and moderately suppressed for β-C in Linco-treated leaves, showing a marked decline of PSII efficiency (Fv/Fm) and β-C content compared with water-treated leaves. Partial inhibition of carotenoid biosynthesis by NF caused no or only a minor decrease in Fv/Fm and Chl a turnover under both conditions, while the β-C content significantly declined and high 14C labeling was found for phytoene, the substrate of phytoene desaturase. Together, the results suggest coordinated turnover of Chl a and D1, but somewhat different regulation for β-C turnover, in Arabidopsis leaves. Inhibition of carotenoid biosynthesis by NF may initially enhance metabolic flux in the pathway upstream of phytoene, presumably compensating for short supply of β-C. Our observations are also in line with the notion that HL-induced accumulation of xanthophylls may involve a precursor pool which is distinct from that for β-C turnover.
Light is essential for photosynthesis but it can also cause oxidative damage to the photosynthetic apparatus when the light energy absorbed by Chl a molecules cannot be utilized for photochemical reactions. A range of photoprotective mechanisms exist in and around pigment–protein complexes of photosynthetic membranes (thylakoids) to minimize photooxidative damage resulting from formation of triplet-state Chl and singlet O2 (Demmig-Adams 1990, Noctor and Foyer 1998, Müller et al. 2001, Maeda and DellaPenna 2007). Nevertheless, PSII, the site of water oxidation and plastoquinone reduction, undergoes light-induced inactivation even under low irradiance, which manifests itself as irreversible damage to the reaction center polypeptide D1 (Tyystjärvi and Aro 1996, Keren et al. 1997). Damaged D1 protein of photoinactivated PSII is replaced by a new copy through an elaborate repair cycle operating between stacked grana and non-stacked stroma regions of thylakoids (Melis 1999, Baena-González and Aro 2002). The D1 and D2 polypeptides of the PSII reaction center bind six Chl a, two pheophytin, two plastoquinone and two β-carotene (β-C) molecules (Kobayashi et al. 1990). Whereas Chl a, pheophytin and plastoquinone are engaged in the PSII electron transfer as redox-active cofactors, β-C is thought to play a photoprotective role (Tracewell et al. 2001, Telfer 2005). It has been suggested that translation elongation and membrane insertion of a new D1 protein during the D1 repair cycle requires assembly partners (Zhang et al. 1999) as well as ligation of Chl a (Kim et al. 1994, He and Vermaas 1998) and β-C (Trebst and Depka 1997). Whenever the repair and turnover of D1 cannot keep pace with the rate of photodamage, e.g. upon an increase in irradiance or under stress conditions, the overall efficiency of PSII declines (‘photoinhibition’).
In parallel with the continuous D1 turnover in the PSII reaction center in the light, interconversions between epoxidized and de-epoxidized xanthophylls (xanthophyll cycle) take place in thylakoids, catalyzed by two enzymes, violaxanthin (V) de-epoxidase and zeaxanthin (Z) epoxidase (Jahns et al. 2009). Light-induced acidification of thylakoid lumen activates V de-epoxidase which converts V molecules released from light-harvesting antenna complexes into antheraxanthin (A) and Z. The reactions are reversed by the activity of Z epoxidase, which becomes evident when V de-epoxidase is inactive in the dark or under non-stressful light intensities. Accumulation of Z via operation of the xanthophyll cycle enhances non-photochemical quenching of singlet-excited Chl (Demmig-Adams 1990, Müller et al. 2001) and provides antioxidative protection (Havaux and Niyogi 1999) under strong illumination. Consistent with these photoprotective functions, sunlit leaves typically contain larger amounts of the xanthophyll cycle pigments (V + A + Z) than leaves in shaded environments (e.g. Demmig-Adams and Adams 1992, Matsubara et al. 2009).
De novo synthesis of V + A + Z (especially Z) as well as β-C and/or lutein (Lut) has been observed in leaves in response to high light exposure and associated with up-regulation of carotenoid biosynthesis to enhance photoprotection (e.g. García-Plazaola et al. 2002, Förster et al. 2009). Recently, we have demonstrated continuous turnover of β-C and Chl a in Arabidopsis (Arabidopsis thaliana) leaves by 14CO2 pulse–chase labeling (Beisel et al. 2010). Based on rapid 14C incorporation into these PSII core complex pigments, but not in light-harvesting antenna pigments (xanthophylls and Chl b), turnover of β-C and Chl a observed under both stressful and non-stressful light conditions was regarded as part of the D1 repair cycle. Since Z is synthesized by β-ring hydroxylation of β-C, a direct link between β-C turnover during the D1 repair cycle and stress-induced Z accumulation has been proposed for Chlamydomonas reinhardtii (Depka et al. 1998). An important implication of these previous observations is that carotenoid metabolism in chloroplasts can sense and respond to environmental changes to meet photoprotective demand for different carotenoids. In support of this notion, increased transcript abundance of β-ring hydroxylase genes, the products of which predominantly catalyze hydroxylation of β-C to Z (Kim et al. 2009), has been reported for Arabidopsis leaves shortly (∼1 h) after high light exposure (Rossel et al. 2002, Cuttriss et al. 2007).
If degradation and synthesis of β-C and Chl a play a part in the D1 repair cycle in leaves, turnover of these pigments may be regulated in conjunction with concurrent turnover of D1 protein. In order to gain insights into interactions between turnover of β-C, Chl a and D1, we have conducted 14CO2 pulse–chase labeling experiments in Arabidopsis leaves following treatments with inhibitors of D1 or β-C synthesis. Lincomycin (Linco), an inhibitor of chloroplast protein synthesis, arrests translation of the plastid-encoded psbA gene product (D1), and thus repair of photoinactivated PSII, resulting in reduced PSII efficiency in leaves (Tyystjärvi and Aro 1996, Lee et al. 2001). Synthesis of β-C can be inhibited by norflurazon (NF) which impairs the catalytic activity of phytoene (Phy) desaturase upstream of β-C. Carotenoid (β-C) depletion by incubation with NF is known to diminish D1 as well as photosynthetic activity in green algae (Sandmann et al. 1993, Trebst and Depka 1997). Also in higher plants, NF treatment leads to impaired PSII efficiency, decreased contents of D1 and other Chl-binding protein complexes of PSII, and, ultimately, bleaching (Markgraf and Oelmüller 1991, Corona et al. 1996, Xu et al. 2000, Welsch et al. 2003). After pre-treatment with these inhibitors and a chase period of up to 6 h under different light conditions, pigments were extracted from leaves and the radioactivity of each 14C-labeled pigment was measured. The results are discussed in terms of (i) whether the turnover processes of β-C, Chl a and D1 are coupled with each other in mature Arabidopsis leaves; and (ii) how carotenoid synthesis responds to NF-induced restriction of metabolic flux in the pathway.
Maximal PSII efficiency
After detaching leaves from Arabidopsis plants grown under 130 μmol photons m–2 s–1 (control light, CL), petioles were immediately put in water (Control), 3 mM Linco or 70 μM NF solution and the leaves were kept for 2 h in the dark with airflow forcing transpiration. The leaves were then floated on water or diluted inhibitor solutions (1 mM Linco or 10 μM NF) with the adaxial surface facing the air, and placed under 130 μmol photons m–2 s–1 (CL→CL treatment) or 1,100 μmol photons m–2 s–1 (CL→HL treatment) starting at 0 h (Fig. 1). Measurements of Chl a fluorescence were performed at different time points during the light treatments. The maximal PSII efficiency (Fv/Fm) remained high in water-treated leaves and NF-treated leaves under CL→CL, whereas the corresponding leaves in CL→HL showed a significant decrease (Fig. 1A, C). While the initial (up to 0.5 h) rapid decrease under CL→HL was similar in water-treated leaves and NF-treated leaves, the subsequent slow decrease in Fv/Fm was more pronounced in NF-treated leaves. The leaves treated with Linco exhibited a significant and almost linear reduction of Fv/Fm from 0.84 to 0.62 over 6 h under CL→CL (Fig. 1B), whereas Fv/Fm declined rapidly to <0.6 in the first 30 min and to almost zero after 6 h under CL→HL conditions.
Chl and carotenoid composition
Chl and carotenoid composition was analyzed in leaves during the Fv/Fm measurements under CL→CL (Fig. 2) and CL→HL conditions (Fig. 3). In both conditions, neither Chl a nor Chl b content changed significantly in leaves of the three treatments throughout the 6 h experiment (Figs. 2A, D, G, 3A, D, G).
Under CL→CL, V was generally the major xanthophyll cycle pigment in leaves (Fig. 2B, E, H); only traces of A and Z could be detected. Except for a small and transient decrease in V in water-treated leaves shortly after the dark to CL transfer at 0 h, the levels of the xanthophyll cycle pigments remained practically unchanged during the experiment and the values were comparable between the leaves treated with water, Linco and NF. The initial levels of Lut, neoxanthin (Neo) and β-C were also similar in the three treatments (Fig. 2C, F, I). However, while Lut and Neo remained nearly constant, the β-C contents significantly decreased by 6 h in Linco-treated leaves and by 3 h in NF-treated leaves.
The carotenoid composition changed substantially in the leaves of all three treatments under CL→HL conditions (Fig. 3). In the first 30 min, water-treated leaves quickly accumulated Z and some A at the expense of V via operation of the xanthophyll cycle (Fig. 3B); the extent of de-epoxidation was 70% (i.e. 70% of V + A + Z in the form of Z or A) at 0.5 h. The highest level of de-epoxidation (90%) was reached at 3 h, concomitant with a significant increase (+50% compared with 0 h) in the total amount of V + A + Z. Thereafter, no further de-epoxidation or increase in V + A + Z was observed. Linco-treated leaves showed similar but slower changes compared with water-treated leaves (Fig. 3E); de-epoxidation progressed more gradually after the first 30 min and V + A + Z initially decreased and then increased by 30% between 3 and 6 h. The operation of the xanthophyll cycle was also evident in NF-treated leaves under CL→HL (Fig. 3H). The time course of de-epoxidation in NF-treated leaves was very similar to the pattern in water-treated leaves. The increase in V + A + Z was also found in these leaves between 0.5 and 3 h, but it was much smaller than in water-treated leaves.
Regardless of the inhibitor treatments, Neo remained unchanged under CL→HL conditions (Fig. 3C, F, I). The variations in Lut were mostly not significant, with the exception of the somewhat higher values (+15% compared with 0 h) in water-treated leaves at 3 and 6 h (Fig. 3C). This increase in Lut coincided with the increase in V + A + Z in these leaves (Fig. 3B). As seen under CL→CL, the β-C contents were stable in water-treated leaves under CL→HL (Fig. 3C) whereas both Linco-treated leaves and NF-treated leaves showed a significant decrease in β-C (Fig. 3F, I). The levels of β-C decreased more rapidly and/or strongly in CL→HL than in CL→CL, with the total reduction under CL→HL after 6 h being –32 and –27% in Linco-treated leaves and NF-treated leaves, respectively, compared with –6 and –12% in the corresponding leaves under CL→CL.
Pulse–chase labeling experiments were conducted with 14CO2 in order to infer synthesis and degradation of photosynthetic pigments in Arabidopsis leaves in the presence of Linco or NF. Following the treatment with either water or the inhibitors (3 mM Linco or 70 μM NF) for 2 h in the dark, leaves were subjected to a 30 min pulse application of 14CO2 under CL, with their petioles incubated in the corresponding solutions (water, 1 mM Linco or 10 μM NF). Then, leaves were placed under CL→CL or CL→HL conditions in ambient air for different durations (chase) while floating on water, Linco (1 mM) or NF (10 μM) solution. Incorporation of 14C into photosynthetic pigments was examined by radio-HPLC analysis (Beisel et al. 2010), and the radioactivity of each pigment was expressed relative to the Chl a content.
For all treatments, rapid and strong incorporation of 14C was detected in Chl a (Fig. 4A, B) but not in Chl b (data not shown). The radiosignal of Chl a did not differ significantly between CL→CL and CL→HL conditions, even when the leaves were treated with Linco or NF. The 14C levels of Chl a were comparably high in water-treated leaves and NF-treated leaves throughout the chase; only the water-treated leaves in CL→CL showed a tendency towards increasing radioactivity for Chl a at the end of the experiment. In contrast, much less radioactivity (∼40 and 50% in CL→CL and CL→HL, respectively) was measured in the Linco-treated leaves.
Also β-C showed similarly high 14C signals under both light conditions (Fig. 4C, D), with no or little labeling detected for xanthophylls (data not shown). The 14C signal of β-C reached maximal levels after a 30 min chase, followed by a decrease or no substantial change until 6 h. In general, less 14C was incorporated in β-C in the inhibitor-treated leaves compared with water-treated leaves. Despite the treatment with NF, β-C was clearly labeled with 14C in the Arabidopsis leaves; under CL→HL a significant reduction in the 14C-labeled β-C was observed with Linco but not with NF (Fig. 4D).
However, a prominent peak emerged in the radiograms of NF-treated leaves at the position of Phy (Fig. 5), the substrate of Phy desaturase, confirming the effect of our NF treatment. While only marginal levels of 14C radioactivity were measured for Phy in water-treated leaves and Linco-treated leaves, strikingly high 14C labeling of Phy was found in NF-treated leaves under CL→CL as well as CL→HL conditions (Fig. 4E, F). The radiosignal of Phy in NF-treated leaves (>100 Net Bq μg–1 Chl a) was higher than the highest radiosignal of β-C measured in these experiments (≤80 Net Bq μg–1 Chl a). The 14C ratio of Phy : Chl a was about 1 : 2 in NF-treated leaves, with the values approaching 1 : 1 in CL→HL after 6 h. For comparison, the 14C ratio of β-C : Chl a was between 1 : 3 and 1 : 5 in water-treated leaves, between 1 : 3 and 2 : 3 in Linco-treated leaves and between 1 : 4 and 1 : 5 in NF-treated leaves. When the 14C signals of β-C and Phy were added together, NF-treated leaves had a 47% higher signal, and Linco-treated leaves had a 49% lower signal, compared with water-treated leaves already at 0 h. Of the three treatments, Linco-treated leaves had the lowest total 14C radioactivity in the sum of Chl a, β-C and Phy at any time point, which was about 60% lower than in water-treated leaves.
Effects of lincomycin on pigment turnover
The D1 protein of the PSII reaction center undergoes a continuous repair cycle at all light intensities (Tyystjärvi and Aro 1996), going through photoinactivation, PSII disassembly, D1 degradation, insertion of a newly synthesized D1 and PSII reassembly (Baena-González and Aro 2002). The antibiotic Linco inhibits protein synthesis in chloroplasts, thereby disrupting, amongst others, the D1 repair cycle and leading to accumulation of photoinactivated PSII (Tyystjärvi et al. 1992, Lee et al. 2001). This explains the reduced maximal PSII efficiency in Linco-treated leaves even under CL→CL (–25% after 6 h; Fig. 1B), revealing the extent of PSII photoinactivation in leaves of CL-grown Arabidopsis plants under non-stressful CL conditions. Obviously, the repair cycle could compensate for this level of D1 damage in water-treated leaves under CL→CL (Fig. 1A). However, transfer to HL dramatically accelerated the rate of PSII photoinactivation in all three treatments (Fig. 1A), concomitant with the operation of the xanthophyll cycle and associated photoprotective mechanisms (Fig. 3B, E, H). Under CL→HL conditions, the rate of photoinactivation exceeded the repair capacity of the CL-acclimated plants, resulting in a decline in Fv/Fm, which manifested itself most strikingly in Linco-treated leaves (Fig. 1B).
Much less 14C was incorporated in both Chl a and β-C in Linco-treated leaves compared with water-treated leaves (Fig. 4), indicating reduced de novo synthesis of these pigments after the Linco treatment. The reduction of 14C incorporation in Linco-treated leaves was more pronounced for Chl a than for β-C, suggesting closer interactions between D1 protein and Chl a turnover. Since Chl a molecules and some of the Chl precursors can become harmful photosensitizers unless they are bound to proteins (Meskauskiene et al. 2001), continuous synthesis and degradation of Chl a may need to be coordinated with the D1 repair cycle. While some secondary effects of Linco cannot be ruled out (Fiekers et al. 1979), the low 14C signal of Chl a measured already at 0 h (Fig. 4A, B) suggests quick down-regulation of Chl a synthesis upon inhibition of D1 protein synthesis. Whether this Linco-induced down-regulation of Chl a turnover is mediated by retrograde signaling is not known. Considering the immediacy of down-regulation, however, the earliest events do not seem to depend on reactions starting from gene transcription.
The total Chl a content did not change in Linco-treated leaves under CL→CL conditions (Fig. 2D), and there was only a small, statistically non-significant decrease in Chl a under CL→HL at 3 h (Fig. 3D) despite severe PSII photoinhibition (Fig. 1B). Given that no more than a few percent of Chl a molecules are bound in PSII reaction centers in leaves, changes in Chl a content during photoinactivation and repair are probably difficult to detect by measuring the total Chl a amount. Judging from the low but steady levels of Chl a radiosignal in spite of the acute PSII photoinhibition in Linco-treated leaves after 6 h in CL→HL (Figs. 4B, 1B), both synthesis and degradation of Chl a seem to slow down when the D1 protein turnover is inhibited. While Chl recycling (Vavilin and Vermaas 2007) may maintain the steady-state 14C level of Chl a, operation of Chl a recycling in photoinactivated PSII is unlikely unless there is concurrent operation of D1 repair. Parallel down-regulation of Chl a and D1 turnover suggests possible enzymatic control of Chl a degradation during the PSII repair cycle, although this does not preclude photooxidation of Chl a under excess light. In addition to thermal dissipation in antenna complexes, strong quenching in photoinactivated PSII (Matsubara and Chow 2004) could reduce formation of Chl a triplet states and the resulting production of singlet O2 to restrict oxidative degradation of Chl a.
The effects of Linco on β-C turnover were similar to those observed for Chl a, albeit not as large (Fig. 4C, D). Thus, coupling between β-C and D1 protein turnover may not be as tight as between Chl a and D1. This is in line with our previous finding of contrasting HL-acclimatory responses of Chl a and β-C turnover in Arabidopsis leaves; acclimation to HL significantly enhanced Chl a turnover, as is known for D1 turnover (Tyystjärvi et al. 1992), while reducing that of β-C (Beisel et al. 2010). Unlike Chl a, free β-C molecules can accumulate in membranes to confer photoprotection, as can be seen in leaves of field-grown plants (Verhoeven et al. 1999, Matsubara et al. 2003). Thus, synthesis and degradation of β-C may be subject to regulatory mechanisms different from those of the D1 repair cycle.
In contrast to the small change in Chl a, the marked decrease of the β-C content in Linco-treated leaves in CL→HL at 3 h (–15%) and 6 h (–30%) (Fig. 3F) suggests enhanced β-C degradation, probably through photooxidation (Tracewell et al. 2001) after pronounced photoinhibition (Fig. 1B). One may ask whether β-C turnover in non-stressful conditions, like in water-treated leaves under CL→CL (Fig. 1A), also reflects bleaching of this pigment in PSII or is controlled by carotenoid cleavage enzymes (Auldridge et al. 2006). Hydroxylation of β-C released from photoinactivated PSII during the repair cycle, giving rise to an increase in Z under HL stress as proposed for Chlamydomonas (Depka et al. 1998), is an alternative scenario to explain the observed β-C decrease concomitant with an increase in Z and V + A + Z (Fig. 3E, F) in Linco-treated leaves. However, the decreasing radiosignal of β-C in Linco-treated leaves and water-treated leaves after 6 h under CL→HL (Fig. 4D) did not result in 14C labeling of Z (data not shown), suggesting that hydroxylation of β-C to Z during β-C turnover may not play a predominant role in Arabidopsis (Beisel et al. 2010).
Effects of norflurazon on pigment turnover
The herbicide NF inhibits carotenoid biosynthesis at the step of Phy desaturation upstream of β-C. In contrast to obvious bleaching of leaves after long-term application of NF during plant cultivation (e.g. Dalla Vecchia et al. 2001), our treatment by feeding detached leaves with a 70 μM NF solution for 2 h did not entirely stop 14C incorporation into β-C (Fig. 4C, D); longer (overnight) treatment with NF did not improve the inhibitory effect (data not shown). The amount of NF taken up by leaves during these treatments was apparently not enough to saturate the cofactor- (plastoquinone; Norris et al. 1995) binding site of Phy desaturase to bring about full inhibition of the enzyme (Breitenbach et al. 2001). Nevertheless, the high 14C labeling of Phy in NF-treated leaves (Fig. 4E, F), together with a decrease in β-C content under both CL→CL and CL→HL conditions (Figs. 2I, 3I), demonstrates some effects of the inhibitor.
Inhibition of β-C synthesis by NF resulted in loss of PSII activity and D1 protein in Chlamydomonas under high light, which has been interpreted as evidence for interrelationships between β-C and D1 protein turnover: photobleaching of β-C is thought to trigger D1 degradation, while the subsequent D1 protein assembly in functional PSII seems to require β-C synthesis (Trebst and Depka 1997). In our experiments with NF-treated Arabidopsis leaves under CL→CL, partial inhibition of β-C synthesis, causing approximately 15% reduction in β-C after 3 h (Fig. 2I), had no obvious impact on Chl a turnover (Fig. 4A) or photoinhibition (Fig. 1C), supporting loose coupling of β-C turnover with Chl a and D1 turnover, as was discussed above for the Linco experiments. This is also in line with the recent finding in Synechocystis sp. PCC 6803 mutant cells, which are completely depleted of carotenoid pigments (Sozer et al. 2010); these mutant cells contain readily detectable levels of D1 and PSII core assembly intermediates (lacking CP43 or both CP43 and CP47) but no dimeric PSII and only a trace of monomeric PSII, indicating a role for β-C in the PSII core complex assembly.
On the other hand, it is difficult to reconcile the NF-induced significant loss of β-C (Fig. 2I) with zero PSII photoinhibition under CL→CL (Fig. 1C) if β-C is needed for the D1 repair cycle even if not directly at the step of D1 protein synthesis and insertion. The following possibilities can be considered to explain these observations: (i) Phy can replace β-C in PSII; (ii) PSII reaction centers lacking β-C are rapidly degraded and do not accumulate; or (iii) not all newly synthesized β-C molecules are needed for D1 repair. Disappearance of PSII by deletion of the Phy desaturase gene in Synechocystis sp. PCC 6803 (Bautista et al. 2005), accumulating Phy as the only carotenoid pigment, rejects (i). As NF-treated leaves showed photoinhibition (Fig. 1C) when the decrease in β-C was more pronounced under CL→HL (Fig. 3I), rapid degradation of photoinactivated PSII reaction centers in the absence of β-C molecules, as stated in (ii), is rather unlikely. The presence of a β-C pool not required for D1 repair and PSII activity, as assumed in (iii), could explain the lack of a short-term effect of NF treatment on Chl a turnover and Fv/Fm under CL→CL conditions, although 15% of the β-C pool in non-stressed leaves appears too large to be dispensable for PSII activity.
Notably, we found about a 50% increase of 14C radioactivity in the sum of β-C and Phy (especially the latter) in NF-treated leaves already at 0 h (Fig. 4C–F). These results are similar to the previous report in leaves of Capsicum annuum showing marked accumulation of Phy and increased total carotenoid levels (twice as high as the levels in control leaves when Phy is included) after 48 h of NF treatment (Simkin et al. 2003). Enhanced labeling of β-C + Phy found in Arabidopsis leaves shortly after NF treatment may suggest rapid up-regulation of carbon flux down the carotenoid biosynthetic pathway, perhaps via a metabolic feedback mechanism (Cazzonelli and Pogson 2010). Activation of the Phy desaturase gene promoter, presumably by end-product regulation, has been documented in tobacco seedlings grown on an NF-containing medium (Corona et al. 1996). Our data in Fig. 4 also point to compensatory up-regulation, but upstream of Phy, which may precede the transcriptional activation of the Phy desaturase gene. It has been shown in Arabidopsis seedlings that gene transcription of Phy synthase, the first committed and rate-limiting enzyme of carotenoid biosynthesis providing the substrate for Phy desaturase, does not increase in response to NF (Welsch et al. 2003). If the situation is the same in mature leaves, the strong accumulation of Phy in NF-treated leaves may involve post-transcriptional enhancement of Phy synthase or up-regulation further upstream of the biosynthetic pathway.
Alternatively, a strong 14C signal of Phy in NF-treated leaves may reflect high stability of Phy molecules (Simkin et al. 2003) compared with β-C which undergoes continuous degradation in leaves during illumination (Beisel et al. 2010). Based on the extent of Phy accumulation in NF-treated Capsicum leaves, Simkin et al. (2003) have estimated a de novo carotenoid synthesis rate as high as 50% d–1 of the total carotenoid content in leaves. If the differences in 14C signal levels of β-C + Phy between water-treated leaves and NF-treated leaves at 0 h were due to such rapid degradation of 14C-labeled β-C upon synthesis, as opposed to high stability (or zero degradation) of 14C-labeled Phy, about 60–70% of the newly synthesized β-C must have been degraded instantly during the 30 min 14CO2 application prior to 0 h. Although such futile synthesis with immediate degradation cannot be excluded, we think it is more likely that up-regulation of biosynthesis also played a role in increasing carotenoid labeling in our NF-treated Arabidopsis leaves.
The enhanced 14C labeling of Phy already at 0 h was followed by a further increase at 6 h in the NF-treated leaves under the CL→HL conditions (Fig. 4F). This second increase was accompanied by a significant decrease in both β-C content (Fig. 3I) and Fv/Fm (Fig. 1C). Neither a β-C decrease alone, as in NF-treated leaves in CL, nor a β-C decrease combined with severe photoinhibition, as in Linco-treated leaves in CL→HL, led to a further increase in the radioactivity of Phy (or β-C) at the end of the measurement (6 h; Fig. 4C–F). In the case of Linco-treated leaves, the 14C signal of β-C even declined. Whatever the mechanisms for the rapid (within 1 h) and slow (∼6 h) increase of 14C-labeled β-C + Phy in NF-treated leaves, the total β-C content did decrease in these samples (Figs. 2I, 3I), suggesting that β-C degradation could not be stopped (or strongly down-regulated) to counterbalance short supply.
The substantial loss of β-C was accompanied by a much reduced increase in V + A + Z in NF-treated leaves compared with the control under CL→HL conditions (Fig. 3H and B, respectively). Thus, inhibition of Phy desaturase by NF can largely suppress the HL-induced increase of V + A + Z in Arabidopsis leaves, as has also been observed in duckweed (García-Plazaola et al. 2002). The residual increase in V + A + Z found in NF-treated leaves may be attributable to incomplete effects of the inhibitor in our experiment, or hydroxylation of β-C molecules released from photoinactivated PSII during D1 repair (Depka et al. 1998). However, we were unable to detect 14C-labeled Z (data not shown), even in the water-treated or Linco-treated samples in which Z as well as V + A + Z levels were significantly increased under CL→HL (Fig. 3B, E). Altering the length and timing of 14CO2 application by, for example, an additional second pulse at a later time or a longer chase of >24 h did not lead to obvious 14C labeling of Z (or other xanthophylls) in the Arabidopsis leaves. Based on these observations and considering the presumably minor role of the β-C hydroxylation during operation of the D1 repair cycle in Arabidopsis plants, it can be hypothesized that HL-induced synthesis and accumulation of V + A + Z may utilize a precursor pool which is not immediately linked to photosynthetic CO2 fixation. If so, chloroplasts may have two distinct pools and fluxes of carotenoid precursors in mature Arabidopsis leaves, one for continuous synthesis of β-C in the light (rapidly deriving carbon from photosynthesis) and another for stress-induced synthesis of V + A + Z and perhaps also Lut (Fig. 3B, C). Further investigations are needed to elucidate the functional and metabolic interactions between pigments and photosynthesis in green leaves.
Both synthesis and degradation in continuous turnover of Chl a seem to be regulated in coordination with D1 protein turnover in mature leaves of Arabidopsis. The β-C turnover, on the other hand, is not tightly coupled with Chl a and D1 turnover, presumably because of indirect involvement of β-C in D1 protein synthesis and insertion as well as its photoprotective function requiring a distinct control. Inhibition of Phy desaturase by NF results in significantly higher initial labeling of β-C + Phy, suggesting up-regulation in metabolic flux down the carotenoid biosynthetic pathway.
Materials and Methods
Plant materials and growth conditions
Arabdiopsis Columbia-0 wild-type plants were grown in soil (ED 73 Einheitserde, Balster Einheitserdewerk) in a growth cabinet with a 12 h/12 h day/night photoperiod under constant relative air humidity of 50% and 22°C/18°C (day/night) air temperature. A photosynthetically active photon flux density of approximately 130 μmol photons m–2 s–1 (CL) was provided by a combination of FLUORA and warm white fluorescent tubes (Osram). At the beginning and at the end of the day period, the light intensity in the growth cabinet was gradually increased or decreased over 1 h to simulate sunrise and sunset. Mature leaves (up to three leaves per plant) of 6- to 7-weeks-old plants were used for all experiments.
Inhibitor and light treatments
At the end of the night period, leaves were excised from plants and the petioles immediately put in water (Control), 3 mM Linco or 70 μM NF solution. Leaves were left to transpire for 2 h in the dark under ventilation. During the subsequent light treatment for experiments of Chl a fluorescence measurements, analysis of photosynthetic pigments and isotope labeling with 14CO2, leaves were floated on water or diluted inhibitor solutions (1 mM Linco or 10 μM NF) with the adaxial surface facing the air. All experiments were conducted under CL (CL→CL condition) or after transfer to HL (1,100 μmol photons m–2 s–1, applied with Master HPI-T Plus lamps; Philips) (CL→HL condition) at an ambient air temperature of 19°C. A preliminary experiment with fluorescence imaging showed no significant spatial variation of Fv/Fm in inhibitor-treated leaves under the CL→HL condition, suggesting homogeneous effects of the inhibitors across the lamina.
Chl a fluorescence measurements
Following different durations of exposure to CL→CL or CL→HL conditions, the detached leaves were placed on a moist tissue and dark-adapted for 15 min by using leaf clips. The maximal PSII efficiency, Fv/Fm (fluorescence nomenclature according to van Kooten and Snel 1990), was determined by measuring Chl a fluorescence in dark-adapted leaves with a Handy PEA (Hansatech).
Analysis of photosynthetic pigments
For pigment analysis, leaf discs (1.54 cm2) were taken from the detached leaves at different times during the CL→CL or CL→HL treatment. The discs were frozen in liquid nitrogen and stored at –20°C for up to 2 weeks until acetone extraction. Pigment extraction and the HPLC analysis were performed according to the method described by Matsubara et al. (2005) using an Allsphere ODS-1 C18 column (5 μm, 250 × 4.6 mm; Alltech) and a corresponding guard column (5 μm, 7.5 × 4.6 mm; Alltech). Pigments were detected by a PAD-996 UV/VIS detector (Waters) and peak areas were integrated at 440 nm with Waters Empower software. The Chl content was calculated per unit of leaf area (μmol m–2) and the contents of different carotenoids were expressed relative to the amount of Chl a (mmol mol–1 Chl a) in each sample.
Isotope labeling with 14CO2 and radio-HPLC analysis of 14C-labeled pigments
14CO2 labeling of leaves was performed in a gas circuit system as previously described (Beisel et al. 2010). Detached leaves were supplied with water or diluted inhibitor solutions (1 mM Linco or 10 μM NF) during administration of 14CO2 released by acidification of aqueous sodium [14C]carbonate (1.85 MBq per leaf; GE Healthcare) under CL conditions at 19°C. Immediately after a 30 min application of 14CO2, some leaves were harvested for analysis (time 0 h) while others were floated on water or diluted inhibitor solutions in ambient air under CL→CL or CL→HL conditions for different chase periods before harvesting for radio-HPLC analysis. The radiolabeled pigments were extracted first with 1.2 ml of acetone, followed by two-phase extraction (ethyl acetate and water) performed twice according to the method of Pogson et al. (1996). The extracts were concentrated to a final volume of 200 μl under a nitrogen gas stream and dim laboratory light. The concentrated extracts were either immediately analyzed by radio-HPLC (100 μl injection volume) or stored at –20°C for <5 h until analysis. Radio-HPLC analysis was carried out with a Prontosil reversed-phase C30 column (3 μm, 250 × 4.6 mm; Bischoff) and a corresponding guard cartridge (3 μm, 10 × 4.0 mm; Bischoff) according to Beisel et al. (2010). The pigments were detected with a UV/VIS detector (Jasco) and radioactivity by a Radioflow detector LB 509 (Berthold Technologies) with a time delay of 20 s between the two detectors (Fig. 5).
Peak integration was performed with RadioStar software (Berthold Technologies) for both UV/VIS chromatograms (286 and 440 nm for Phy and photosynthetic pigments, respectively) and radiograms. Peak areas of the radiogram were normalized to the Chl a content obtained from the corresponding 440 nm chromatogram and expressed as Bq μg–1 Chl a.
An extra peak found in the radiogram of NF-treated leaves (Fig. 5C) was identified as Z-phytoene (Phy) by adding a pure standard of (E/Z)-phytoene (CaroteNature) to a leaf extract and measuring the absorbance at 286 nm (Fig. 5B).
Statistical data analysis
Pigment contents and 14C labeling data were statistically tested by one-way analysis of variance (ANOVA; Dunnett's method). Pigment contents were checked for significant differences in time-course variations within each treatment by comparing data at 0.5, 3 and 6 h with 0 h. Differences in 14C incorporation between inhibitor-treated and water-treated (Control) leaves were statistically tested at each sampling time point.
We thank Diana Hofmann and Stephan Köppchen (Forschungszentrum Jülich) for valuable suggestions and technical assistance with radio-HPLC analysis. Critical comments of Siegfried Jahnke (Forschungszentrum Jülich) and Ralf Welsch (University of Freiburg) on the early version of the manuscript are greatly appreciated. K.G.B. acknowledges the support of her Ph.D. thesis at the Heinrich-Heine-Universität Düsseldorf.
control light (130 μmol photons m–2 s–1)
maximal PSII efficiency in dark-adapted leaves
high light (1,100 μmol photons m–2 s–1)