Influence of the photoperiod on redox regulation and stress responses in Arabidopsis thaliana L. (Heynh.) plants under long- and short-day conditions
- First Online:
- Cite this article as:
- Becker, B., Holtgrefe, S., Jung, S. et al. Planta (2006) 224: 380. doi:10.1007/s00425-006-0222-3
- 357 Downloads
Arabidopsisthaliana L. (Heynh.) plants were grown in low light (150 μmol photons m−2 s−1 and 20°C) either in short days (7.5 h photoperiod) or long days (16 h photoperiod), and then transferred into high light and low temperature (350–800 μmol photons m−2 s−1 at 12°C). Plants grown in short days responded with a rapid increase in NADP-malate dehydrogenase (EC 220.127.116.11) activation state. However, persisting overreduction revealed a new level of regulation of the malate valve. Activity measurements and Northern-blot analyses indicated that NADP-malate dehydrogenase transcript and protein levels increased within a few hours. Using macroarrays, additional changes in gene expression were identified. Transcript levels for several enzymes of glutathione metabolism and of some photosynthetic genes increased. The cellular glutathione level increased, but its redox state remained unchanged. A different situation was observed in plants grown in long-day conditions. Neither NADP-malate dehydrogenase nor glutathione content changed, but the expression of several antioxidative enzymes increased strongly. We conclude that the endogenous systems that measure day length interact with redox regulation, and override the interpretation of the signals, i.e. they redirect redox-mediated acclimation signals to allow for more efficient light usage and redox poising in short days to systems for the prevention of oxidative damages when grown under long-day conditions.
KeywordsGlutathione Light acclimation Malate valve Oxidative stress Photoperiod Redox regulation
In their natural environment, plants are subjected to conditions where several parameters such as light or temperature can change dramatically for periods which can last from seconds to days. Plants possess a set of mechanisms, such as Mehler reaction, photorespiration or malate valve (Scheibe 2004), to prevent damages in the short-term, but when environmental alterations persist, a response at the transcriptional level is induced (Scheibe et al. 2005). This leads to changes in the ultrastructure of the leaves and in protein and pigment composition (Anderson und Osmond 1987; Das 2004). The chloroplasts of leaves acclimated to low light are characterized by a high content of grana thylakoids, relatively high amounts of photosystems and light harvesting complexes (LHCII) and only low quantities of stromal proteins. Such low-light plants respond to a persisting increase in the light intensity with a very characteristic alteration of gene expression (Rossel et al. 2002; Murchie et al. 2005). Consequently, light saturation of photosynthesis is reached at much higher light intensities, the Chl a/b ratio decreases due to a loss of LHCII, and the content of ATPase and of Calvin-cycle enzymes increases. The relative amount of stroma thylakoids also increased, while the amount of grana thylakoids is decreased (Anderson and Osmond 1987). It was further shown that acclimation of plants to high light intensities delays the onset of photoinhibition and accelerates the recovery from photoinhibitory treatments (Aro et al. 1994). However, in the case of Arabidopsis, acclimation towards different light intensities is not linearly dependent upon the light intensity, but follows a complex pattern with separate low- and high-light responses (Bailey et al. 2001).
Changes in pigment and protein composition are often caused by alterations in gene expression. There is good evidence that in the case of light acclimation the initial signals do not arise from photoreceptors, but originate from the photosynthetic machinery inside the chloroplasts. Studies on mutants with lesions in cytosolic photoreceptors or in chloroplast-to-nucleus communication indicate that neither phytochromes nor cryptochromes are essential for the perception and transduction of light-induced photosynthetic signals (Walters et al. 1999; Fey et al. 2005). From the unaltered redox state of the glutathione pool and the lack of reactive oxygen species (ROS) formation, it was further concluded by Fey et al. (2005) that the redox state of the thylakoid membrane itself is the main origin of redox signals. Additional evidences for a role of photosynthetic electron transport arise from plants with a decreased amount of ferredoxin (Fd) 1, the major Fd isoform during photosynthetic electron distribution. When grown under elevated light intensities, the mutant plants suffer permanently under elevated redox states of PQ and other intersystem carriers. Published (Holtgrefe et al. 2003) and unpublished results (data not shown) indicate an extreme form of light acclimation in the mutant plants. For Dunaliella, it was shown that the expression of the LHC-isoform LHCB I is initially regulated by the thylakoid membrane pH gradient, and later by the redox state of the PQ pool. Assembly and binding of the responsible transcription factors to the Lhcb1 promoter was shown to be thiol-dependent (Chen et al. 2004).
Although significant evidence points to a central role of chloroplasts during acclimation, the pattern of gene expression is not uniform. Even in genetically identical individuals of the same species (i. e. A. thaliana var. Col-0), exogenous and endogenous factors (e.g. photoperiod, sugar accumulation or the developmental state of leaves and plants) modulate both, pattern and intensity of light acclimation (Scheibe et al. 2005; Walters 2005). It is long known that the day length is decisive for flower induction in many plant species. Arabidopsis is a facultative long-day (LD) plant, i.e. flowering is induced earlier in long days or under permanent illumination, while it is delayed under short-day (SD) conditions (Koornneef et al. 1998). As long as the plants are sufficiently supplied with sucrose, flowering will even occur in complete darkness (Madueno et al. 1996). Flower induction follows an endogenous program which is modulated by exogenous factors such as temperature, day-length or light intensity. Phytochrome, phototropin and cryptochrome are probably involved in the initial signal perception. In addition, several mutants with endogenous deficiencies in genes involved in flower induction are known (Koornneef and Peeters 1997). For the perception of the photoperiod, the free Ca2+ in the cytosol plays a role (Dodd et al. 2005). In addition, there are some evidences for interrelations between glutathione content (Ogawa et al. 2004), antioxidative capacities (Kurepa et al. 1998) and flowering.
Although flower induction itself has been intensively studied, the duration of the light period regulates many fundamental processes of plant growth, even before flowering is induced. However, the basic physiological consequences of the duration of the light period on plants in the vegetative state have not been studied yet. In this paper, we provide evidence that the duration of the light period is interconnected with light acclimation and the induction of antioxidative capacities in Arabidopsis plants in the vegetative state.
Materials and methods
Plant material and growth conditions
Seeds of Arabidopsis thaliana L. (cv. Columbia Col-0) were placed on soil and kept in darkness for 48 h at 4°C in order to synchronize germination. After 2 weeks, the seedlings were transferred into larger pots (one per seedling) containing a commercially available soil mixture (9% sand, 36% pumice, 36% compost, 19% peat, additionally supplemented with crystalline fertilizer).
As light sources, SON-T AGRO 400 lamps (Philips, Eindhoven, The Netherlands) were used. During the initial growth period, the light intensity was around 150 μmol quanta m−2 s−1, measured at leaf height, and the relative humidity was 65%. During the light period, the temperature was 22°C, and the temperature in the dark was 18°C. At the time points given under “Results”, the plants were transferred into a growth cabinet with the respective conditions (350 or 800 μmol photons m−2 s−1 at 12°C).
Leaf samples were taken by quickly removing all leaves from the rosette of the plant and transferring them into liquid nitrogen in less than 10 s. Leaves from 2 to 4 different plants were combined per sample. The leaves were homogenized under liquid nitrogen, and stored at −80°C until use. All measurements were repeated with at least 3 different samples.
MDH measurements in leaf samples
The samples used for the estimation of the in vivo NADP-MDH activity were obtained from individual leaves. The leaves were cut from the plant and immediately transferred into liquid nitrogen. Activation state and maximum activity (capacity) of NADP-MDH, and the NAD-MDH activity were determined as described by Scheibe and Stitt (1988).
Antioxidative enzyme activities
For the determination of the enzyme activities, leaves were immediately frozen in liquid nitrogen, stored at –80°C, and ground in liquid nitrogen. The following steps were performed at 4°C. For measurements of APx and Cat activity, 1 ml extraction buffer A, consisting of 50 mM sodium phosphate, pH 7.0, 1% (w/v) PVPP and 1 mM ascorbic acid was added. The samples were allowed to thaw on ice. After centrifugation for 10 min at 20,800 g and 4°C, the supernatants were used for activity measurements. For the determination of GR activity, 1 ml of extraction buffer B, consisting of 50 mM sodium phosphate, pH 7.0, 1% (w/v) PVPP, 3 mM EDTA, pH 8.0 and 10 mM 2-mercaptoethanol was added to the frozen powder. After thawing on ice, the samples were centrifuged for 10 min at 20,800 g and 4°C. The supernatants were applied to a NAP-10 column (Amersham Pharmacia Biotech), which had been equilibrated with buffer C, consisting of 50 mM sodium phosphate, pH 7.0, 3 mM EDTA, pH 8.0 and 10 mM 2-mercaptoethanol. Elution of the protein fraction was performed with buffer C, and the desalted sample was used for the enzyme activity assays. The activities of Cat, GR and APx were determined according to Del Longo et al. (1993).
Determination of the glutathione content
The amounts of GSH and of GSSG were determined with a modified method of Schulte et al. (2002). Frozen powder from Arabidopsis leaves was weighed and then deproteinized with 2% (v/v) perchloric acid for 30 min on ice. Afterwards, the perchloric acid was removed by adding saturated K2CO3 until a neutral pH was reached. The subsequent derivatization of the samples with monobromobimane (mBBr; Sigma) and stabilization with acetic acid were performed as described in Schupp and Rennenberg (1988). The glutathione-mBBr derivatives were separated with a Supelcosil LC 18 column (Supelco-Sigma) using a Shimadzu SL 10 system. The excitation wavelength was 380 nm, and fluorescence emission was recorded at 480 nm. For the quantitation of GSH and GSSG, a set of standards with GSH or GSSG amounts between 0.03 and 0.8 nmol was used. The quantification of the glutathione-mBBr derivatives within the samples was performed with the Shimadzu Class-VP software. The recovery rates of exogenously added GSH to Arabidopsis samples ranged between 94 and 106%.
Determination of chlorophyll and protein content
Measurements of chlorophyll fluorescence
For the measurement of chlorophyll fluorescence, leaf disks were cut and analyzed immediately. The leaf disks were illuminated first at 220 μmol photons m−2 s−1 for 10 min, followed by illumination at 880 μmol photons m−2 s−1 for 10 min. Chl fluorescence was measured with a PAM fluorimeter (Walz, Effeltrich, Germany). The quantum yield of PSII (φII) was determined from steady-state values as described in Backhausen et al. (1998). For each measurement, leaf disks from three different plants were used.
Gene expression analysis with a macroarray
For the large-scale analysis of the expression of genes of interest, a cDNA macroarray was used containing 96 gene specific PCR-products (188–923 bp in length) in duplicates on nylon membranes (data not shown). The synthesis and quantification of digoxygenin-labeled cDNA, the array hybridization and data evaluation were performed as described by Kandlbinder et al. (2004), but the normalisation of signals was carried out with reference to the actin spots. The alterations in gene expression are shown as the ratio of the spot intensities of treated samples (6 or 78 h) to the spot intensities of control samples, taken immediately before treatment (0 h) −1. A value of zero indicates constant expression, positive values indicate an increase, and negative values represent a decrease of gene expression.
The Northern-blot analysis was essentially conducted using standard techniques (Sambrook et al. 1989). DNA probes for NADP-MDH were amplified by PCR from total cDNA with specific primers. The PCR products were separated on agarose gels, and purified using the Nucleo-Spin Extraction Kit (Macherey & Nagel). The purified DNA probes were radioactively labelled according to the manufacturer`s instructions (Ready-to-go DNA Labeling Kit, Amersham). Non-incorporated nucleotides were removed by gel filtration with NAP-5 columns (Amersham), as described in von Schaewen et al. (1995). Finally, the probes were denatured. RNA-blot analysis was conducted using standard techniques (Sambrook et al. 1989). Total RNA was isolated as described in von Schaewen et al. (1995), and separated on denaturating formaldehyde-agarose gels. After electrophoresis, the RNA was blotted on nylon membranes. Hybridization with the radioactively labelled probes, and washing of the membranes followed the methods described in Church and Gilbert (1984). The Northern blots were exposed for 3–5 days at −80°C to X-OMAT/XAR-5 films (Kodak), and were then analyzed using the Gelix One Gel Scan software.
Non-competitive RT-PCR was performed essentially as described by Ahn (2002). The cDNA was synthesized from 3 μg total RNA using oligo(dT) as primers for 50 min at 42°C according to the manufacturer`s instructions (Superscript II Reverse Transcriptase, Invitrogen). 1 μl of cDNA was used in 20 μl PCR reactions. The PCR settings were: 1 cycle 94°C for 4 min, and for the optimised number of cycles for each gene product 1 min 94°C, 1 min 40–67°C and 2 min 72°C and a final extension at 72°C for 5 min. Following separation of the PCR products on 2% agarose gels, the bands were quantified densitometrically using the Gelix One Gel Scan software. Each band was normalized against the intensity obtained with the same cDNA using the UBQ10-specific primers. Gene code numbers, primers, melting temperatures and number of cycles are given in the supplementary material.
(1) Spotting list for the array with EMBL and MatDB numbers; (2) Numbers and primer sequences; (3) Primer list, cycle numbers, buffers and melting temperatures for RT-PCR.
Establishment of conditions for the induction of NADP-MDH expression
Changes in malate valve, chlorophyll fluorescence and antioxidative system
For GR activity, such difference between both light regimes was less obvious. In SD as well as in LD, the initial activity was around 0.7 μmol cm−2 min−1, and increased during the treatment (Fig. 4b). The average final activities were lower in the LD plants, but the values had a very high variability. Interesting differences were detected in the glutathione content of the leaves. In Fig. 4d, the time courses of the GSH contents are shown. In the SD plants, the amount of GSH increased from 8 nmol cm−2 by 50% within a few hours, and the final concentration was twice as high as it was initially. In contrast, the GSH content in the leaves of the LD plants remained constant at 14 nmol cm−2. The content of GSSG always paralleled the GSH content (data not shown). Thus, the redox state of the glutathione system remained between 86 and 92% reduced under all conditions.
In order to obtain more direct physiological evidences for the contrasting responses of SD and LD plants, the time-course of the quantum yield of PSII (φII) was followed during the treatment. Leaf disks were illuminated at 220 and 880 μmol photons m−2 s−1, until stable values were obtained. Under both light intensities, the quantum yield of the SD plants increased steadily by 15–20%. In the LD plants, only a temporary increase was observed after 6 h, but afterwards the φII dropped again to the initial values (at 880 μmol photons m−2 s−1) or even below (at 220 μmol photons m−2 s−1).
The quantitative changes of the transcripts analyzed with the array that were apparent after 6 h and after 78 h are summarized in Fig. 5. The largest changes were observed after 6 h of treatment for both, LD and SD plants. In the SD plants, the transcript levels of 16 genes had increased by a factor of two or more, while in the LD plants, ten transcripts had increased to a similar degree. In the SD-grown plants, cytosolic G6PDH, two enzymes of glutathione metabolism (cyt. GR and chl. GPx), three SOD isoforms and three thioredoxins (chloroplast Trx f and m, and cytosolic Trx h3) were among the transcripts with the strongest increase of expression after 6 h of treatment. In addition, the transcript levels of three LHC genes had increased. Although several transcript amounts decreased, none of them fell below 50% of the initial value. After 78 h, only chloroplast FBPase and two of the thioredoxins (cyt. Trx h3 and Trx h4) remained high. The transcript levels of all other genes had returned more or less to their initial values.
In the LD-grown plants, the transcript levels of ten genes had increased by more than 100%, but the pattern was different. Among the transcripts with the largest increase after 6 h were the antioxidative enzymes APx3, chloroplastidic MDHAR and FeSOD1. In addition, the transcript levels of nine genes decreased by more than 100%. Among them were three LHCs. After 78 h, ten transcript levels were increased above 100%, but the pattern showed some differences. Only the APx3 transcript was still high, and the levels of some LHC transcripts were now increased. Only LHCB1.5 was also increased in SD after 6 h, but the other LHC transcripts were different between LD and SD.
Changes of NADP-MDH expression in SD plants
The nuclear-encoded NADP-MDH is one of the chloroplast enzymes, which is subject to light/dark modulation by the ferredoxin-thioredoxin system. The activation state of the NADP-MDH is influenced by the NADP/NADPH ratio (Scheibe and Jacquot 1983), so the NADP-MDH activation state becomes low when little NADPH is available. Because of this fine-tuning, the enzyme converts oxaloacetate into malate only when the NADPH/NADP ratio is high. The operation of the malate valve is important to stabilize the chloroplast NADPH/ATP ratio, especially under rapidly changing conditions (Scheibe 2004). Since malate is exported into the cytosol, the malate valve acts also as an export system for excess reducing equivalents (Backhausen et al. 2000). The mitochondria can use the produced malate as source for cytosolic ATP supply. The importance of mitochondria as an effective sink for reductants produced by the chloroplasts was recently confirmed in the Chlamydomonas stm6 mutant which is deficient in the mitochondrial transcription termination factor MOC1 (Schönfeld et al. 2004).
There have been only few earlier reports that not only the activation state of the existing enzyme, but also the amount of NADP-MDH is regulated and responds to changes in the environment. In tobacco plants, the NADP-MDH content showed a strong age dependency, with the total activity being the highest in young, still growing leaves (Faske et al. 1997). When tobacco plants were grown under different CO2 concentrations, which caused alterations in the stromal redox state due to the different ratios of ATP and NADPH consumption, it was shown that under elevated CO2 (which leads to a higher NADPH demand per ATP and causes a decreased stromal redox state), the expression of NADP-MDH was lowered (Backhausen and Scheibe 1999). Savitch et al. (2000) observed that the total NADP-MDH activity is nearly ten-fold higher on a Chl basis in cold-grown winter wheat, as compared to control plants grown at 20°C. In cold-acclimated Arabidopsis plants grown in a SD, the total NADP-MDH, calculated on Chl base, activity was also higher than under control conditions (Savitch et al. 2001). However, the increase was only threefold, which is less than in wheat, but similar to the changes of capacity obtained in this work with the SD plants.
It is concluded from the data shown in Fig. 3 that in the SD plants, the increase in NADP-MDH activity is mainly caused by an increased transcription rate already a few hours after the transfer into reducing conditions. It is also evident that this up-regulation is not a single event, but part of the overall process of high-light acclimation. It seems that increased metabolic flexibility, i. e. a higher capacity for the safe removal of excess electrons and a higher glutathione content are required to increase the light-use efficiency. Increased malate-valve capacity ensures that e. g. upon a sudden change in the light intensity, more electrons can be exported as malate, and thus in turn can be consumed by mitochondria.
It is assumed that the alteration in the chloroplast redox state, as caused by the transfer into high light and low temperature, is responsible for the release of redox signals which in turn activates the transcription of several nuclear-encoded genes, among them NADP-MDH. It was shown recently that excitation imbalances between PS I and PS II generate redox signals in the thylakoid membranes which induce acclimatory changes in the structure of the photosynthetic apparatus and also controls the expression of nuclear-encoded genes (Fey et al. 2005). Increased redox states of electron carriers between PS I and PS II, as well as a more reduced stromal redox state (evident from the higher activation state of NADP-MDH, Fig. 3a) are also expected in the plants used in our study. The experiments with photoreceptor mutants carried out under similar conditions (Walters et al. 1999; Fey et al. 2005) make it unlikely that phytochromes or cryptochromes are directly involved in the chloroplast-to-nucleus signalling.
Gene expression patterns and antioxidative enzymes
It is interesting to note that increased levels of NADP-MDH transcript and protein, and several other acclimatory changes were only observed when the plants were grown in SD conditions. In LD, many parameters (e.g. NADP-MDH activity, glutathione content) were already at higher values before the treatment started, and the response towards a transfer into reducing conditions was poor or completely absent. The differences in the quantum yields of PS II indicate, that in fact metabolism is altered, and not only changes in gene expression or protein levels occurred. However, the expression changes of NADP-MDH, the glutathione content and of some antioxidative enzyme activities are not the only parameters that were influenced by the duration of the light period. The expression patterns of selected genes (Figs. 6, 7) did not reveal many responses common to SD and LD plants, indicating that the specific effect of the light period on gene expression and its responsiveness to changing light intensity and temperature is a more general phenomenon and not restricted to NADP-MDH and glutathione.
From the array and RT-PCR data, a clear tendency became apparent, showing that no further increase, but often a decrease was detected when the expression level of a gene was already high at the beginning of the experiments. However, an increased expression in the SD plants only rarely reached the initial expression level of the LD plants. The expression changes observed with both methods yield identical tendencies in most cases, but the extent of the change was often higher when analyzed with the array. Only in two cases (Trx f2 and Trx h5), contrasting results were obtained. Trx h5, which showed a large increase on the array, gave only very poor signals in the RT-PCR, and did not show an expression increase at all. Probably the expression level of the mRNA was too low for this approach. However, in the case of members of large gene families (e. g. Trx or LHC), we cannot exclude the possibility that a cross reaction with another Trx isoform gave a false-positive signal on the array.
The differing expression patterns between SD and LD that arise from the enzyme activity measurements can also be found in the array data. The activity of antioxidative enzymes such as APx or Cat responded much more in LD than in SD, but the corresponding data from the gene expression analysis point to a more complex pattern. Regarding only the activity data, it is neither possible to distinguish between different isoenzymes nor to assess whether transcription or translation are affected. From the comparison of activity data with the gene expression data it seems that the activity changes are the result of complex up- and downregulation of the various compartment-specific genes, which follow different patterns in SD and LD. Different modes of transcriptional regulation may also play a role.
Redox-signal interpretation in LD and SD plants
From our studies it became apparent that the redox-mediated changes in the activity of NADP-MDH, in some antioxidative enzymes, and in the glutathione content were only visible when the Arabidopsis plants were grown under SD conditions. In addition, both the initial gene expression levels and the responses towards a transfer into reducing conditions differed significantly, depending on the length of the light period. When the light period was extended above 8 h, noisy data without a clear pattern were obtained for all parameters, independent of plant or leaf age (data not shown). Only under extreme LD conditions, with a light period above 15 h, a clear picture emerged and reproducible data could be obtained (Figs. 5b, c, 6). It is likely that the noisy data obtained from plants grown with light periods between 8 and 15 h arose from a mixture of LD and SD responses in individual plants. This raises two questions: The first is, how a population of genetically identical plants can show such an individual variation. This question cannot be answered from our data. The second question is how the duration of the light period can interfere with redox signals and modify the response of the plants in such a way. The distinct plant habitus (Fig. 1) is related to different growth and flowering strategies in SD and LD conditions. SD plants have an extended vegetative period and invest much of the absorbed light energy into production of leaf biomass. LD plants start to flower as soon as possible in order to reach reproduction as soon as possible. Unfortunately, only few data are available for an influence of day length on plants that are still in the vegetative state, since the majority of published data deals with the switch that shifts plants from the vegetative state towards flowering. Our data show that SD and LD plants differ in the interpretation of chloroplast-generated redox signals long before flowering (or even senescence) starts.
Correlation between redox state and flowering induction
In the last years, it has become clear that different pathways within the plants collect informations about exogenous parameters such as nutrition, light or temperature. They control the expression of so-called ‘floral integrators’, which include LEAFY, FT and SOC1 (Lohmann and Weigel 2002). When the expression of these genes reaches a critical level, flowering is induced. In Arabidopsis, there are two positively interacting pathways. LD conditions activate the photoperiod pathway, whose output is mediated by the CONSTANS gene (Schultz and Kay 2003). The daylength-dependent expression of CONSTANS is regulated via FKF1, a flavin-containing F-Box protein which probably acts as photoperiodic blue-light sensor (Imaizumi et al. 2003). A high expression level of CONSTANS induces FT, leading to flowering under LD conditions. Transgenic plants which overexpress COL9, a protein that represses CONSTANS expression, always flower late in LD (Cheng et al. 2005). In SD, flowering depends on another pathway that is regulated by gibberellins. The activity of both pathways is negatively regulated by the repressor FLC, which integrates the exposure to low temperatures with the other events (Schmid et al. 2003).
There are some evidences for an interrelationship between light period, antioxidative capacities, flower induction and life span. In the work of Yanagida et al. (2004) it was shown that GSH biosynthesis is associated with bolting in Eustoma grandiflora. The higher rate of GSH biosynthesis, which is caused by an induction of γ-ECS activity on a post-transcriptional level (May et al. 1998), was suggested to promote flowering induction.
Since in annual plants like Arabidopsis, flowering also leads to senescence and finally to the end of the individual life span, it should be considered that redox-active components might also prolong the life span. The viability of many non-plant organisms has been correlated with their resistance against oxidative stress (Martin et al. 1996). A positive correlation between antioxidative capacities and life span was observed in yeast, Caenorhabditis elegans, Drosophila melanogaster and in mammals (Fleming et al. 1992; Kennedy et al. 1995; Jazwinski 1996; Martin et al. 1996). The idea of a connection between antioxidative status and life span is supported by the finding that most long-living mutants of C. elegans are highly resistant against oxidative and thermal stress, and against UV irradiation. The degree of resistance against such stresses was directly proportional to the increased life span (Murakami and Johnson 1996).
The authors thank H. Rennenberg and M. Eiblmeyer (Universitaet Freiburg, Germany) for their help with the method of glutathione determination. Further thanks are due to S. Klocke for her help with performing the experiments, and R. Brockmann for the frustrating job of measuring the NADP-MDH activities in Arabidopsis plants grown in mixed SD and LD conditions. We finally thank H. Wolf-Wibbelmann and S. Steinbach for excellently growing the plant material. This work was financially supported by a grant from the Deutsche Forschungsgemeinschaft (FOR 387, TP1 and TP3).