Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Adjustment of diurnal starch turnover to short days: depletion of sugar during the night leads to a temporary inhibition of carbohydrate utilization, accumulation of sugars and post-translational activation of ADP-glucose pyrophosphorylase in the following light period
Max Planck Institute of Molecular Plant Physiology, Science Park Golm, Am Muehlenberg 1, 14476 Potsdam-Golm, Germany
A larger proportion of the fixed carbon is retained as starch in the leaf in short days, providing a larger store to support metabolism and carbon export during the long night. The mechanisms that facilitate this adjustment of the sink–source balance are unknown. Starchless pgm mutants were analysed to discover responses that are triggered when diurnal starch turnover is disturbed. Sugars accumulated to high levels during the day, and fell to very low levels by the middle of the night. Sugars rose rapidly in the roots and rosette after illumination, and decreased later in the light period. Global transcript profiling revealed only small differences between pgm and Col0 at the end of the day but large differences at the end of the night, when pgm resembled Col0 after a 4–6 h prolongation of the night and many genes required for biosynthesis and growth were repressed [Plant J. 37 (2004) 914]. It is concluded that transient sugar depletion at the end of the night inhibits carbon utilization at the start of the ensuing light period. A second set of experiments investigated the stimulation of starch synthesis in response to short days in wild-type Col0. In short days, sugars were very low in the roots and rosette at the end of the dark period, and after illumination accumulated rapidly in both organs to levels that were higher than in long days. The response resembles pgm, except that carbohydrate accumulated in the leaf as starch instead of sugars. A similar response was found after transfer from long to short days. Inclusion of sugar in the rooting medium attenuated the stimulation of starch synthesis. Post-translational activation of ADP-glucose pyrophosphorylase (AGPase) was increased in pgm, and in Col0 in short days. It is concluded that starch synthesis is stimulated in short day conditions because sugar depletion at the end of the night triggers a temporary inhibition of growth and carbohydrate utilization in the first part of the light period, leading to transient accumulation of sugar and activation of AGPase.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
In the light, leaves carry out photosynthesis and export sucrose to the remainder of the plant, to support respiration, storage and growth. In the night, the plant becomes a net consumer of carbohydrate. Part of the photoassimilate is retained in the leaf, typically as starch, and remobilized at night to support leaf respiration, and continued synthesis and export of sucrose (Geiger and Servaites, 1994; Geiger et al., 2000; Stitt, 1996; Stitt et al., 1987). Diurnal turnover of starch acts as a buffer against fluctuations in the net carbon balance of the plant. Its vital role has been demonstrated by studies of starchless mutants. They grow like wild-type plants in continuous light, but growth is progressively impaired as the duration of the night is increased, and is arrested in short days (Caspar et al., 1985; Huber and Hanson, 1992; Lin et al., 1988).
The mechanisms that stimulate starch synthesis in short days are still obscure. Most previous studies of the regulation of starch synthesis have investigated how starch synthesis is increased in response to high levels of phosphorylated intermediates or sugar. Starch synthesis can be increased via allosteric activation of ADP-glucose pyrophosphorylase (AGPase) by a rising 3PGA/Pi ratio (Preiss, 1988), via post-translational redox-activation of AGPase in response to illumination or high sugar (Hendriks et al., 2003; Tiessen et al., 2002, 2003, see below for more details), and via transcriptional regulation in response to high sucrose (Koch, 1996), low nitrate or low phosphate (Nielsen et al., 1998; Scheible et al., 1997). It is unclear how these mechanisms could contribute to the stimulation of starch synthesis in short days when less carbon is available, at least on a whole plant basis.
Starchless mutants provide a powerful experimental system to unravel the response to a disturbance of diurnal carbon partitioning. Caspar et al. (1985) suggested that starchless mutants grow poorly because they waste carbon. Large amounts of soluble sugars accumulate in their leaves during the day, instead of starch, and are depleted during a period of rapid respiration in the first part of the night. However, this cannot be the main reason for their poor growth. Starch excess (sex) mutants show a similar inhibition of growth in short days, but do not show any enhancement of respiration at the start of the night (Caspar et al., 1991; Zeeman et al., 1998). An alternative explanation would be that inappropriate regulatory responses are triggered by the extreme diurnal changes of sugars in these mutants. Analysis of these responses could provide insights into how wild-type plants regulate photosynthate partitioning in response to changing conditions.
One possibility is that the high levels of sugar at the end of the light period trigger catabolite repression of photosynthesis and other processes (Koch, 1996; Stitt, 1991). Caspar et al. (1985) found altered activities of enzymes involved in carbohydrate metabolism in the starchless mutants in short days, including sucrose phosphate synthase (SPS) and AGPase, and suggested that regulatory imbalances triggered by high sugars at the end of the day might contribute to the inhibition of growth. Starchless mutants contain lower Rubisco activity than wild-type plants, indicating that feedback inhibition of photosynthesis by sugars is increased in low nitrogen (Sun et al., 1999, 2002). However, this response was only seen on low nitrogen, whereas growth of starchless mutants is inhibited more strongly on high nitrogen (Schulze et al., 1991).
A second possibility, which has received less attention, is that the period of low sugar in the second part of the dark period inhibits metabolism and growth. When wild-type tobacco is exposed to a lengthened night, depletion of starch and sugars is followed by a dramatic decrease of NIA (nitrate reductase) transcript and NIA protein in the leaves (Klein et al., 2000; Matt et al., 1998). Tobacco growing in short days has low levels of the NIA transcript, NIA protein and NIA activity, very high levels of nitrate and low levels of amino acids and low levels of protein (Matt et al., 1998). These observations indicate that a transient period of low carbon leads to an inhibition of nitrogen assimilation and use. Recently, Thimm et al. (2004) reported experiments using the 22 K Arabidopsis Affymetrix array, which showed that a 6 h extension of the night leads to decreased levels of hundreds of transcripts that encode proteins involved in nutrient utilization, amino acid, lipid, nucleotide and cell wall biosynthesis, cell wall modification and protein synthesis, and increased levels of transcripts for hundreds of genes involved in amino acid, nucleotide, lipid and cell wall breakdown. Widespread changes also occurred in the expression of many genes involved in hormone synthesis and sensing, and in the regulation of transcription, protein modification and protein breakdown. Similar changes were found at the end of the night in starchless pgm mutants (Thimm et al., 2004). These results indicate that a transient period of acute carbohydrate deficiency triggers a wide-ranging inhibition of biosynthesis and growth. It can be envisaged that when a plant in this state is re-illuminated, carbohydrates will be formed but not immediately used. This would result in a transient accumulation of sugars, which would in turn trigger an increase in starch synthesis.
The following paper examines this hypothesis. The first part of the paper uses two complementary approaches to investigate whether the transient period of low sugar at the end of the night in the starchless pgm mutant triggers a temporary inhibition of carbohydrate utilization in the first part of the following light period. Whole-genome transcript profiles were used to diagnose the physiological state of the pgm mutant, and physiological parameters were measured to investigate whether carbohydrate consumption is inhibited at the start of the light period. The second part of the paper asks whether the pgm mutant serves as a model to understand how wild-type plants adjust their photosynthate allocation in response to a change in the diurnal regime.
Comparison of whole-genome transcript profiles in pgm and Col0
The pgm mutant contains very high levels of sugars at the end of the day and very low levels of sugars at the end of the night (see Introduction, also below for more data). To learn whether pgm resembles a wild-type plant with surplus sugars or a wild-type plant suffering from carbohydrate depletion, global expression profiles of pgm and Col0 growing in a 12 h light/12 h dark diurnal cycle were compared at the end of the day and at the end of the night. In all cases, whole rosettes were harvested. The present article concentrates on a statistical comparison of the datasets. Detailed changes in the expression of individual genes at selected time points have been published in Thimm et al. (2004).
Table 1 summarizes regression coefficients of pair-wise comparisons. The signals were transformed into logarithmic values to obtain a more even distribution of data points across the entire dynamic range of the dataset. Two independent experiments were carried out, at an interval of 3 months. The regression coefficients for four pair-wise comparisons of biological replicates in the two experiments lay between 0.938 and 0.954. When Col0 harvested at the end of the day was compared with pgm harvested at the end of the day, the regression coefficients (0.954, 0.958) resembled those between biological replicates. Lower regression coefficients (0.755, 0.815) were obtained when Col0 harvested at the end of the night was compared with pgm harvested at the end of the night. Similar results were obtained, irrespective of whether all signals were used, or genes called ‘not present’ by the Affymetrix software were excluded. These results show that a transient period of carbohydrate deficiency during the last part of the night has a much larger impact on genome-wide expression than a transient period of high sugars at the end of the day. The dataset obtained for pgm at the end of the night was also compared with datasets for wild-type Col0 after transfer to continual darkness for different periods of time. In both experiments, the regression coefficients improved when pgm at the end of the night was compared with Col0 that had been exposed to a 4 h extension of the normal night (from 0.755 to 0.816 and from 0.815 to 0.876). A further extension of the night led to a decrease of the correlation coefficients.
Table 1. Comparison of regression coefficients between the global gene expression in the pgm mutant at the end of the night, and wild-type Col0 harvested at different times into an extended night. Col0 and pgm were grown in a 12 h light/12 h dark light cycle. Two experiments are shown. The pgm mutant was sampled at the end of the night in both experiments. Col0 was harvested at the end of the night (=0 h), and 2, 4, 6, 8 h into an extended night in experiment 1, and 2 h before the end of the nigh (=−2 h), the end of the night (0 h), and 4, 8, 12 and 48 h into an extended night in experiment 2. All samples included 15 separate plants. All samples were analysed using the Affymetrix ATH1 array, and using the MASC software as in Thimm et al. (2004). The signals were transformed to a log2-scale to obtain a better distribution of data points across the entire dynamic range. Regression coefficients were then calculated between the dataset for the pgm mutant at the end of the night and the datasets for Col0 at different times. This was carried out for a subset of 8953 genes that were called present by the MASC software at all time points in all experiments, and for all genes on the chip
Extended night (h)
R2 (genes called present)
R2 (all genes)
The data from the experiments of Table 1 were subjected to PCA analysis and the first and second components plotted (Figure 1a). The datasets for Col0 at the end of the day group together, and are clearly separated from the datasets for Col0 at the end of the night. An extension of the night led to a progressive displacement, with a similar trajectory for both experiments. The datasets for pgm at the end of the day lie close to those of Col0 at the end of the day, whereas they are clearly separated from the Col0 at the end of the night. At this time they are shifted and lie closer to Col0 after a 4 or 6 h extension of the night.
Figure 1(b) examines expression in a smaller set of genes, assigned to photosynthesis. The genes are taken from the BIN for photosynthesis, defined in the TranscriptScavenger module of MapMan (Thimm et al., 2004). They include many genes reported to be subject to catabolite repression by high sugars (Koch, 1996). For each gene, the signal in pgm was divided by the signal in Col0 at the same time of day. A table of the data is included in the Supplementary material. At the end of the day, expression of this set of genes was only very slightly reduced in pgm compared with Col0. At the end of the night, expression of a subset of the genes was markedly decreased in pgm compared with Col0. This subset included all the genes on the array that encode the small subunit of Rubisco and many of the genes that encode chlorophyll binding proteins (see Supplementary material), which are classical examples of sugar-repressed genes (Koch, 1996). These results show that, at least in these growth conditions, the high sugar levels at the end of the day do not lead to catabolite repression of genes for the photosynthetic apparatus. Instead, a subset is repressed at the end of the night, following a transient period of sugar depletion.
Diurnal changes of sugars and starch in the leaves of Col0 and pgm growing in a 12 h light/12 h dark light regime
Diurnal changes of carbohydrates were investigated in Col0 and pgm rosettes, to learn whether the large changes of expression at the end of the night affected the use of carbohydrates in the first part of the light period. In Col0, there was a small lag at the start of the light period, which was followed by a near-linear manner accumulation of starch (Figure 2a, see below for more data). In the dark starch was remobilized, falling to low levels by the end of the night. Soluble sugars (Figure 2b–c) rose early in the light period, stabilized for the remainder of the light period, fell slightly after darkening and then remained stable until the end of the night. In pgm, sugars accumulated rapidly in the first part of the light period. This accumulation was reversed in the second part of the light period. Similar results were obtained in three independent experiments performed over a time interval of 12 months (data not shown). Sugars fell rapidly after darkening and were very low during the second part of the night, as previously reported (see Introduction).
Rates of export in Col0 and pgm in a 12 h light/12 h dark light regime
Photosynthesis was slightly lower in pgm than Col0 (Figure 3a). However, a constant rate was maintained throughout the light period, showing that the reversal of sugar accumulation in the second part of the light period is not caused by an inhibition of photosynthesis. Respiration in pgm was double that in Col0 during the first 4 h of the night, and lower than in wild-type plants later in the night (see also Caspar et al., 1985).
The rate of carbon export (Figure 3b) was calculated from the difference between the net rate of carbon exchange and the net change of the rosette carbohydrate content in a given time interval. The rosette includes sink leaves, but their size relative to the whole rosette is so small that it does not interfere with the calculation (data not shown). In wild-type Col0, similar rates of export were maintained throughout the light period. Darkening led to a 70–80% inhibition of export (see also Geiger et al., 2000). In pgm, export was inhibited by 50% compared with Col0 during the first part of the light period. During this time interval, pgm retained about half of the fixed carbon in the rosette (compare Figure 3a,b). Later in the light period export increased in pgm, and exceeded the rate in Col0.
Levels of carbohydrates in the roots of Col0 and pgm in 12 h light/12 h dark light regime
The low rate of export during the first part of the light period in pgm could be the result of a direct inhibition of export or a backup of carbohydrates from sink organs. To distinguish these possibilities, we investigated sugar levels in the roots. As exhaustive sampling of roots in soil-grown plants is not possible, plants were grown on vertical agar plates. Carbohydrates were measured in the leaves to confirm that agar-grown plants respond like soil grown plants. In Col0, most of the carbohydrate was starch (data not shown). Carbohydrate accumulation started after a lag and accelerated later in the light period (Figure 4a). In pgm, sugars accumulated rapidly in the leaves at the start of the light period and slowed down later in the light period (Figure 4a). The rate of accumulation of carbohydrates in the first 2 h was twofold higher than in Col0. This confirms the results of Figure 2, using a very different plant growth system.
In Col0 roots, sucrose (Figure 4b) and reducing sugars (Figure 4c) were present at substantial levels at the end of the night, and rose gradually during the day. Starch was <10 μmol hexose equivalents g−1 DW (data not shown). A very different response was found in pgm roots. Sucrose (Figure 4a) and reducing sugars (Figure 4b) were present at low levels at the end of the night (7 and 11 μmol hexose equivalents g−1 DW, respectively). After illumination they rose rapidly to higher levels than in Col0. The increase continued until 4 h into the light period, after which root sugar levels stabilized. In relative terms, the increase of sugars in the roots was much larger than the increase in the shoot.
The data in Figures 2–4 can be used to estimate the whole plant carbon budget in pgm during the first 4 h of the light period. An export rate of about 2C6 equivalents g−1 fresh weight (FW) rosette h−1 would be required to support the observed rate of sugar accumulation in the roots during the first 4 h of the light period (about 280 μmol C6 equivalents g−1 DW root, see Figure 4a,b, the calculation assumes a root water content of 90% and a root/shoot ratio of 3.5). The measured rates of photosynthesis and carbon export in the experiment of Figure 3 during the first part of the light period were 8 and ca. 4 C6 equivalents g−1 FW of rosette h−1. The carbon accumulated in the roots therefore represents about half of the carbon exported from the rosette, and the total carbon accumulated in the roots and rosette represents about 75% of the carbon fixed during the first 4 h of the light period in pgm. This comparison may underestimate the proportion of the current photosynthate that is accumulated, because the growth light intensity was slightly lower in the experiment of Figure 4 (see Materials and methods). These results show that there is a massive inhibition of carbohydrate utilization in pgm during the first hours of the light period.
Short-day conditions stimulate starch synthesis in leaves of Col0
We next investigated whether short days lead to a stimulation of starch synthesis in wild-type Col0, and whether this is also accompanied by a transient accumulation of sugars at the start of the light period. Wild-type Col0 was grown for 2 weeks in an 8 h light/16 h night regime to allow a rosette to be established. Subsets of plants were then transferred to a range of photoperiods for 2 weeks (Figure 5, see below for discussion of the parallel experiment with pgm). In this first experiment, carbohydrates were measured at the beginning (Figure 5a) and end (Figure 5b) of the light period. Considerable amounts of starch were left in the leaves at the end of the night in a 20 h light/4 h dark regime, some in the 16 h light/8 h dark regime, and very little when the day length was shorter. The average rate of starch synthesis (Figure 5c) was estimated by dividing the difference between the starch content at the beginning and end of the day by the length of the light period. Decreasing the day length from 20 to 6 h led to progressive fourfold stimulation of starch synthesis. The rate of photosynthesis declined slightly on an FW basis in plants grown in short days (data not shown), showing that allocation to starch is stimulated by greater than fourfold. Increased starch accumulation in short day regimes was observed in many independent experiments (data not all shown). The extent of the stimulation varied, ranging, for example, 20–100% when plants growing in a 12 h light/12 h dark and a 6 h light/18 h dark cycle were compared.
The results in Figure 5 also reveal that the average rate of starch breakdown at night is decreased in short day conditions. The rate falls from 5.0 to 2.8, 1.8 and 1.1 μmol C6 g−1 FW h−1 as the day length is decreased from 16 to 12, 9 and 6 h, respectively. This was confirmed in many other experiments (see below, and data not shown). Typically, the rate of starch breakdown in a 6 h light/18 h dark cycle was about half that in a 12 h light/12 h dark cycle.
The diurnal changes of leaf carbohydrates in a 6 h light/18 h dark and a 12 h light/12 h dark regime were investigated in more detail in the experiment of Figure 6. Starch synthesis was faster and started without a lag in short days, whereas there was a lag in long days (Figure 6b, see also Figures 3 and 4 and below for more data). Sugars at the end of the night were 50% lower in short days than in long days. After illumination, sugars rose threefold faster in short days than in long days (Figure 6a). As a result, sugar levels in the first part of the light period were higher in short day than in long day-grown plants. This result was confirmed in an independent experiment (data not shown). The stimulation of starch accumulation in short days is therefore accompanied by a transient increase of leaf sugars. This is masked when sugars are compared at the end of the respective light periods (see e.g. Figure 5b), because the leaf sugars continue to rise gradually throughout the whole light period in long days (see Figure 6a). Figure 6 also confirms that the rate of starch breakdown is lower in short day than in long day-grown plants. This may explain why leaf sugars are lower at the end of the night in short day conditions.
Short-day conditions lead to increased accumulation of soluble sugars in leaves of pgm
The experiment of Figure 5 also investigated the impact of day length on carbohydrate accumulation in pgm. The level of sugar in the rosette at the end of the light period rose dramatically as the day length was reduced (Figure 5d,e). The average rate of sugar accumulation per hour was calculated by dividing the net difference in sugar contents at the beginning and end of the day by the length of the light period (Figure 5f). Sugar accumulation in pgm increased dramatically as the day length was decreased. The amount of carbon accumulated as sugar resembled the amount that accumulated as starch in Col0. The finding that carbon accumulation responds to day length in the same way in wild-type Col0 and a starchless mutant provides independent evidence for the conclusion that the stimulation of starch synthesis in short day conditions is a secondary response, which is triggered by decreased carbon export.
The increase in the average rate of sugar accumulation in pgm when the day length is decreased may be partly the result of changes in the timing of sugar accumulation and utilization. When pgm is grown in a 12 h light/12 h dark cycle, sugars accumulate rapidly in the first part of the light period, and remain constant or decrease later in the light period (Figures 2 and 4). Average accumulation rates calculated from the sugar levels at the beginning and end of the light period (Figure 5f) will therefore underestimate the rate of sugar accumulation at the start of the light period. This effect presumably becomes less marked as the light period is shortened. Time-of-day-dependent changes may also contribute to the higher average rate of starch accumulation in short day conditions in wild-type Col0. Whereas starch synthesis typically starts after a lag in long day conditions (see e.g. Figures 2a, 4a and 6b), starch accumulation commences without a noticeable lag in short day conditions (Figure 6b, see also below for more data).
Diurnal changes of sugars at the root tips of Col0 growing in a 6 h light/18 h dark and a 12 h light/12 h dark cycle
To investigate whether the transient accumulation of sugars in the leaf in the first part of the light period in short days (Figure 6a) is the result of a direct inhibition of export or a backup of carbon from sink organs, rosettes, roots and root tips were sampled at different times of the day from Col0 growing on nutrient agar in a 6 h light/18 h dark or a 12 h light/12 h dark cycle. The results for the rosettes resembled those in Figure 6: sugars rose only slowly and starch accumulation started after a lag in long days, whereas there was a rapid transient accumulation of sugars, and starch synthesis started immediately in short days (data not shown). In a 12 h light/12 h dark cycle, sucrose levels in the whole root system (data not shown) and at the root tip (Figure 7) were fairly stable throughout the diurnal cycle. In a 6 h light/18 h dark cycle, sucrose levels in the root (data not shown) and at the root tip (Figure 7) were low at the end of the night, rose rapidly after illumination and decreased in the second part of the light period. This result was confirmed in an independent experiment (data not shown).
A switch from a long to a short day regime leads to transient accumulation of sugars and stimulation of starch synthesis in leaves of Col0
We next investigated whether there is a transient accumulation of sugar and stimulation of starch synthesis at the start of the light period after a transition from a long to a short day regime. Col0 was transferred from a 12 h light/12 h dark regime to a 6 h light/18 h dark regime by extending the night for 6 h. Leaf carbohydrate levels were measured in samples taken just before, during and for 2 days following this shift. At the end of the 12 h night, leaf starch was almost exhausted (Figure 8a). There was still a substantial pool of sucrose (Figure 8b) but reducing sugars were low (Figure 8c). When the night was extended by 6 h, starch fell below detection, and sucrose decreased threefold. This point is termed the ‘end of the first long night’. After illumination, sucrose and reducing sugars increased rapidly for the first 2 h, reaching levels that were much higher than in the first part of the light period in a 12 h light/12 h dark cycle. The rate of starch accumulation in the 2 h was 1.9-fold higher than the average rate in a 12 h light/12 h dark cycle. From 2 h onwards, sucrose and reducing sugars decreased, and the rate of starch synthesis declined to a rate that was 115% of the average rate in a 12 h light/12 h dark cycle.
Overall, about 70% as much starch was accumulated during the first short day as in the long day, although the light period was halved from 12 to 6 h. The amount accumulated was high enough to support starch breakdown through to almost the end of the second long night. Sucrose levels were slightly higher at the end of the second long night than at the end of the first long night. The increase of sucrose and reducing sugars after illumination on the second day after transfer was smaller than on the first day. By the end of the third long night, starch and sugar levels resembled those at the end of the night in a 12 h light/12 h dark cycle.
Glucose-6-phosphate (Figure 8d) dropped to very low levels during the first extended night, and increased rapidly upon re-illumination, but did not reach levels found at the end of the day in the 12 h light/12 h day regime. The following nights, glucose-6-phosphate levels did not fall so far during the night, and did not rise so far in the day. This implies that the higher rate of starch synthesis in short days is not a simple consequence of higher levels of phosphorylated intermediates (see Discussion).
Re-illumination after an extended night leads to accumulation of sugars in Col0 roots
Root carbohydrates were measured to check whether the rapid accumulation of photoassimilates in leaves after transfer to short day conditions is the result of a direct inhibition of export or a backup from sink organs. Control plants growing on nutrient agar in the absence of sugar in a 12 h light/12 h dark regime were harvested at the end of the night and at various times during the day. Another set of plants were subjected to further 6 h darkness at the end of the night, and then illuminated. Plants were harvested at the end of the extended night, and the first 6 h in the light (Figure 9). Sugars levels were lower at the end of the long night than in non-shifted controls in both the roots (Figure 9a) and the rosette (Figure 9b). When the plants were illuminated, sucrose and reducing sugars rose rapidly in the roots (Figure 9a), reaching levels after 2–4 h that were 50–100% higher than at a comparable stage of the light period in the roots of non-shifted controls. There was also a rapid increase of sugars in leaves of the plants that were re-illuminated after a long night (Figure 9b). Starch synthesis commenced immediately in plants that had been subjected to an 18 h night, whereas it commenced with a lag and was slower in non-shifted control plants (Figure 9c).
Root extension growth is rapidly inhibited in an extended night
To provide direct evidence that carbon depletion leads to a rapid inhibition of growth, video imaging was used to monitor the rate of primary root extension of 12 day old seedlings growing on nutrient agar in the absence of sugar. Wild-type Col0 roots grew at almost the same rate throughout the light and the dark period. Growth decreased within 2 h and stopped within 4 h when the plants were subjected to an extended night (data not shown). Root growth stopped during the normal night in the pgm mutant, and this was prevented by including sucrose in the medium (data not shown).
Inclusion of sucrose in the rooting medium decreases starch accumulation in wild-type plants growing in short days
The results so far imply a transient period of sugar depletion at the end of a long night results in a temporary inhibition of carbohydrate use, accumulation of sugars and stimulation of starch synthesis in the first part of the next light period. To provide a further independent evidence for this explanation, we grew Col0 in an 18 h light/6 h dark regime or a 6 h light/18 h dark regime in nutrient agar on a range of sucrose concentrations. Inclusion of sucrose in the medium should prevent the transient period of sugar deficiency at the end of the night in short day conditions and lead in these conditions to a counter-intuitive decrease, rather than an increase, of starch synthesis.
In a preliminary experiment, wild-type Col0 was grown in an 18 h light/6 h dark regime on zero to 5% sucrose, to define a range in which exogenous sucrose could be varied without it leading to a general accumulation of carbohydrates. Starch levels at the end of the light period were low and stable between 0.5 and 2% sucrose, but rose sharply at higher sucrose concentrations (data not shown). Based on these results, Col0 was grown on zero to 2% sucrose in a 6 h light/18 h dark regime or an 18 h dark/6 h light regime. More starch accumulated in short day than long day conditions. Starch was almost completely degraded at the end of the night in both regimes (see Supplementary material for the original data). To estimate the rate of starch synthesis, the difference between the starch content at the beginning and end of the light period was divided by the length of the light period (Figure 10). Increasing the sucrose concentration from zero to 2% led to a progressive decrease in the rate of starch synthesis in short days, but not in long days. This result was confirmed in two further experiments.
A final set of experiments were carried out to identify potential regulatory mechanisms that might contribute to the stimulation of starch synthesis in short days. These experiments were carried out in wild-type Col0 growing in different photoperiods, and in parallel in pgm because this provides an independent biological system, in which the shift to decreased export in the first part of the light period occurs at longer day lengths and in the presence of different levels of the individual carbohydrates.
Expression of members of gene families that encode SUT transporters
Specific real-time RT-PCR primers were used to assay the transcript levels of five members of the SUC family of sucrose transporters in Col0 and pgm growing in a 20 h light/4 h dark, 12 h light/12 h dark and 7 h light/17 h dark regime (Figure 11). The most strongly expressed member was SUC2 (At1g22710), followed by SUC1 (At1g71880), and then SUC3 (At2g02860) and SUC4 (At1g09960). SUC5 (At1g71890) was expressed at only very low levels (not shown).
In wild-type Col0, SUC1 expression increased markedly, and SUC2, SUC3 and SUC4 expression increased slightly as day length decreased. Further, as day length decreased there was a consistent shift towards increased expression at the end of the night, relative to the end of the light period. In pgm, there was a strong trend to increased expression of SUC family members, compared with Col0 (Figure 11). This was especially strong in long days, but was also retained in a 12 h light/12 h dark regime. No SUC family members were repressed in pgm compared with Col0 at the start of the day in a 12 h light/12 h dark cycle, when export is transiently decreased in pgm (see Figure 3b). The only case where expression was lower in pgm than in wild-type Col0 was at the end of the night in short days for SUC4, which is expressed at only a low level compared with the other family members. It therefore appears unlikely that transcriptional repression of SUC transporters is responsible for the inhibition of export during the first part of the light period in the pgm mutant, or in wild-type Col0 in short days.
Expression and overall activity of SPS and AGPase
SPS and AGPase catalyse the first committed steps leading to sucrose and starch, respectively. The activities of these enzymes were measured in optimized conditions in the presence of saturating levels of substrates and allosteric inhibitors. The extraction and assay medium for AGPase also contained dithiothreitol to completely convert AGPase into reduced active form (see below).
Short days led to a 35% decrease of SPS activity on an FW basis (Figure 12a). This decrease is mainly the result of a lower overall protein content in short day conditions (data not shown). Short days also led to a marked decrease of AGPase activity at the end of the day (Figure 12b). In contrast, AGPase activity decreased only marginally at the end of the night. The ratio of AGPase activity/SPS activity at the end of the night rose from 1.7 to 2.2 and 2.8 as the day length was decreased from 20 to 12 and 7 h, respectively (compare Figure 12a,b). This will favour increased partitioning to starch during the first part of the light period in short days. A similar picture emerged in pgm, where SPS activity in pgm resembled that in wild-type Col0 (Figure 12a), and AGPase activity declined at the end of the day but was high at the end of the night. The observation that this marked diurnal rhythm of AGPase activity is already found in a 12 h light/12 h dark regime in pgm but only appears in the 6 h light/18 h cycle in wild-type Col0 indicates that it is linked to a disturbance of diurnal carbon allocation.
Post-translational regulation of AGPase
Redox-dependent post-translational activation of AGPase (see Introduction) involves the reduction of an intermolecular cysteine bridge between the two AGPB subunits of the heterotetrameric holoenzyme (Ballicora et al., 2000; Fu et al., 1998). The reduced monomeric form has a much higher affinity for the substrate ATP and the allosteric activator 3PGA (Hendriks et al., 2003; Tiessen et al., 2002). As a result of the high levels of oxidants in leaves, the reduced active form is rapidly converted into the inactive form after extraction (Hendriks et al., 2003), making it impossible to routinely monitor post-translational activation via activity assays. Post-translational activation can be monitored more conveniently by preparing extracts in trichloroacetic acid to denature the holoenzyme and prevent post extracto formation of dimers, followed by separation in a non-denaturing gel and immunoblotting with specific antibodies against the AGPB subunit (Hendriks et al., 2003; Tiessen et al., 2002) to determine much of the AGPB protein is present as the active 50 kDa monomer, and how much as the inactive 100 kDa dimmer (Figure 13).
We have already reported that AGPase activation is strongly increased throughout the light period in pgm (see Hendriks et al., 2003). When Col0 is grown in a 12 h light/12 h dark cycle, AGPase is totally inactivated at the end of the night, is only marginally activated after 30 min and becomes more activated towards the end of the light period (see also Hendriks et al., 2003). After an extended night, AGPase is already markedly activated within 30 min of illumination, and strongly activated within 2 h. After three days in the short day regime, AGPase activation has decreased but is still higher than in long days. The activation of AGPase is accompanied by a stimulation of starch synthesis (Figure 8a) and a decrease of Glc6P (Figure 8d).
SPS activation was also investigated, by measuring activity in the presence of saturating and limiting levels of substrates and activators. The estimated activation was unaffected by day length, and was also similar in wild-type Col0 and pgm (data not shown).
Growth is inhibited in the pgm mutant because a period of carbon starvation at the end of the night leads to a temporary inhibition of carbon utilization when photosynthesis recommences
The inhibition of growth in starchless mutants is primarily caused by a disturbance of metabolism and growth, which is triggered by a transient period of sugar depletion in the second part of the night. During the first part of the light period, large amounts of carbohydrate accumulate throughout the plant, rather than being used for biosynthesis and growth. Based on our estimates, about 75% of the fixed carbon accumulates as sugars in the roots and shoot. Later in the light period sugars fall in the roots, export from the leaf increases and the accumulation of sugars stops or is even reversed in the leaves. This shows that the rate of carbohydrate utilization now exceeds the rate of photosynthesis. Starch excess mutants also show a rapid increase of leaf sugars after illumination and stabilization or decrease later in the light period (Caspar et al., 1991; Zeeman et al., 1998), indicating that their poor growth in short day conditions can be explained in the same way.
Short days lead to increased starch accumulation in the leaf because sugar utilization is restricted during the first part of the light period
Decreasing day length leads to a stimulation of starch synthesis in Arabidopsis leaves, as already reported for other species (see Introduction). Several lines of evidence show that the stimulation of starch synthesis is a secondary event, which is triggered by decreased utilization of photosynthate at the start of the light period. First, there is an analogous accumulation of sugars in the leaves of the starchless pgm mutant. A similar response is also seen in starch excess mutants (the authors, data not shown). This demonstrates that the stimulation of starch synthesis in wild-type plants cannot be primarily caused by regulation of partitioning between sucrose and starch. Second, the accumulation of starch in wild-type plants in short days is partially reversed by supplying sugars to the roots, revealing that it is triggered by low sugar. Sugars are especially low at the end of the night in short day conditions. Third, illumination of short day-grown plants leads to a marked increase of sugars in the roots and a parallel increase of sugars in the shoots during the first part of the light period. This is accompanied by a stimulation of starch synthesis. This resembles the pgm mutant, where a transient period of sugar depletion in the second part of the night triggers a massive restriction of carbohydrate utilization throughout the plant in the first part of the following light period (see above). This response is found on the first day after transferring Col0 from short to long day conditions and is damped on subsequent days, when a larger pool of starch allows higher sugar levels to be maintained at the end of the night. However, even after several days in short day conditions, sugars fall to lower levels at the end of the night and accumulate more rapidly at the start of the light period than in long days. Incidentally, this explains why starch synthesis commences immediately after illumination in short days, whereas there is typically a lag before starch accumulation starts in long days.
The inhibition of carbohydrate utilization in pgm, and in wild-type Col0 in short days, is accompanied by massive changes in gene expression
The expression of hundreds of genes is altered in the pgm mutant at the end of the night, compared with wild-type Col0 at the same time. This includes many genes that are required for nutrient assimilation, biosynthesis and growth (Thimm et al., 2004). When the night is extended by 4–6 h, global gene expression in wild-type Col0 resembles that in pgm at the end of the normal night (see Figure 1, also Thimm et al., 2004 for a detailed discussion of the genes involved). These results provide a molecular framework to understand why growth is inhibited in the first part of the light period in pgm, and in Col0 after an extension of the night. Further studies will be needed to reveal whether changes that occur in the last part of a normal night are already starting to initiate an inhibition of growth, and whether the changes in expression are already leading to changes in the levels of the encoded proteins, or whether parallel post-transcriptional or post-translational events are responsible for the inhibition of carbon utilization.
Comparisons of plants in different day length regimes could be complicated by circadian regulation. A similar stimulation of starch synthesis after a transition to short days was obtained, irrespective of whether the night was prolonged (Figures 8 and 9) or the preceding day was shortened (data not shown). Further evidence that circadian regulation does not play a primary role in the stimulation of starch synthesis in short days is provided by the finding that diurnal carbohydrate turnover changes in a similar manner when pgm and Col0 are compared in a common photoregime, and when Col0 is compared in short and long day conditions. This conclusion is supported by the similar global expression profiles in pgm at the end of the night and Col0 after an extension of the night (Table 1, Figure 1, see Thimm et al., 2004, for more details).
Post-translational regulation of AGPase is primarily responsible for the stimulation of starch synthesis in the first part of the light period in short day conditions
Higher plant AGPase is a heterotetramer, consisting of two catalytic (AGPB) and two regulatory (AGPS) subunits. The enzyme is subject to post-translational regulation, which involves reversible reduction and oxidation of an intermolecular cysteine bridge between the two AGPB subunits (Ballicora et al., 2000; Fu et al., 1998). The dimeric form is inactive and the reduced monomeric form is active. It has a much higher affinity for ATP and increased sensitivity to allosteric activation by 3PGA (Tiessen et al., 2002). Changes in the activation state will therefore lead to marked changes in AGPase activity in vivo, when it operates in the presence of limiting levels of substrates and activator.
One reason for the increased rate of starch synthesis in short day conditions is that post-translational activation of AGPase is increased. Activation is markedly increased on the first day after transfer from long to short day conditions, and this increase is retained in a damped form after adjustment to the new regime. The observation that glucose-6-phosphate decreases in short day conditions provides indirect evidence that starch synthesis is stimulated via a mechanism that acts directly on AGPase. It is already known that AGPase is activated by high endogenous sugar or exogenous addition of sugars in potato tubers (Tiessen et al., 2002) and by light and sugars in leaves (Hendriks et al., 2003). This prompts the working hypothesis that the higher levels of leaf sugars in the first part of the light period in short day conditions lead to increased post-translational activation of AGPase. The precise mechanism still has to be elucidated. In tubers, sugars act via at least two routes, one involving sensing of hexose and requiring hexokinase, and one involving sensing of sucrose and involving SNF1 (Tiessen et al., 2003).
Changes in the expression or turnover of AGPase and SPS may also contribute to adjustment to short days
Changes in overall enzyme activities reveal that additional mechanisms may also contribute to the adjustment to short days. Short days lead to an increase of overall AGPase activity measured in the presence of dithiothreitol, saturating substrates and activator at the end of the night, whereas SPS activity decreases slightly. The ratio of AGPase/SPS activity is also slightly increased in pgm at the start of the light period. Further studies are required to establish whether the changes in overall AGPase activity are solely because of changes of transcription or whether post-transcriptional regulation is also involved.
None of the SUC family members were repressed at beginning of the day in short days in Col0. There was also no evidence for a decrease of SUC family transcripts in pgm at the beginning of the day in short days or a 12 h light/12 h dark regime. SUC expression nevertheless responds to changes in the sink–source balance in the plant. For most members, expression increased slightly in short day conditions and diurnal regulation was modified resulting in increased transcript at the end of the night relative to the end of the light period. There is an analogous trend when the pgm mutant is compared with Col0.
Short periods of sugar depletion lead to marked changes of gene expression (Thimm et al., 2004) and trigger an inhibition of carbohydrate utilization. Such excursions are likely to be detrimental to orderly growth and development. The results in the present paper provide a framework to understand how plants adjust the allocation of carbon between storage and export in response to daily changes in the carbon supply, in order to avoid or at least minimize such excursions. Even short periods of depletion act to decrease the utilization of sugars, when they subsequently become available. The resulting accumulation of sugars leads increased starch synthesis. Further research is needed to elucidate the signalling pathways that link increased levels of sugars to post-translational activation of AGPase, and to the shift in the overall activities of AGPase and SPS. The preciseness with which the starch turnover is adjusted to changes in day length in a recurring light regime will probably require a very sensitive sensing mechanism to prevent the plant from swinging between excess and depletion. This raises the question whether these mechanisms sense falling levels of sugars and act to slow carbohydrate utilization or increase carbohydrate storage in the following light period, before falling sugars lead to a disturbance of cell function.
Materials and methods
Arabidopis thaliana ecotype Col0 and the pgm mutant in Col0 background were grown on soil as in Thimm et al. (2004). The length of the light period was varied, keeping the total diurnal cycle at 24 h. The different photoperiods are indicated in the figure legends. Extended night experiments on soil were performed with wild-type Col0 grown in a 14 h light/10 h dark day/night cycle at a light intensity of 140 μmol m−2 sec−1. Plants were harvested in the vegetative state with three rosettes per sample, and typically at least five replicate samples. Rosettes were transferred into liquid nitrogen under ambient growth irradiance. Plants were also grown on nutrient agar (0.8%, w/v) supplemented with half-strength Murashige–Skoog medium (Murashige and Skoog, 1962), except that ammonium nitrate was replaced by potassium nitrate [and with various sugar concentrations as indicated in the figures [0, 0.1, 0.5, 1 and 2% (w/v) sucrose, and at a light intensity of 100 μmol m−2 sec−1]. Plates were orientated vertically or horizontally (see Figure legends). Leaves of plants growing on horizontal plates were separated from sucrose containing medium by a nylon mesh with 1 mm mesh size. For harvest, shoots, roots and, in one case, the terminal 2 mm of the root tip, were separated, and transferred to liquid nitrogen within 2 sec. Samples contained 10–20 plants, and typically at least five replicate samples were collected.
Gas exchange measurements
Gas exchange measurements were performed in a special custom-designed open system developed in collaboration with Walz (Effeltrich, Germany) as described in Muschak et al. (1999). The chamber was modified to enable measurements on intact plants. The surface of the pot was wrapped with thin rubber foil and epicotyl was surrounded with a gas-tight soft rubber to prevent any gas exchange with the soil. Measurements under growth conditions started 30 min after setting the plant in the chamber. After measurement, the leaf surface area of the plant was determined by photostating and afterwards measuring the area using MetaImaging Series-Metamorph (Visitron Systems GmbH, Puchheim, Germany). The diameter of plants ranged from 3 to 6 cm, when artefacts caused by leaf overlap are minimal (Eckert and Kaldenhoff, 2000). Fresh weight was determined immediately after taking the image.
Chemicals were purchased from Sigma (St Louis, MO, USA), except NADH (Roche Mannheim, Germany). Enzymes for analysis were purchased from Roche except invertase (Sigma) and UMP-kinase, which was overexpressed in Escherichia coli and purified as in Serina et al. (1995).
Extraction and assay of metabolites
Sucrose, glucose, fructose were determined in ethanol extracts as described in Geigenberger et al. (1996), starch was determined as in Hendriks et al. (2003) and glucose-6P as in Gibon et al. (2002). Assays were prepared in 96-well microplates using a Multiprobe II pipetting robot (Perkin-Elmer, Zaventem, Belgium). The absorbances were red at 340 or 570 nm in a Synergy or an ELX-800-UV microplate reader (Bio-Tek Friedrichshall, Germany).
Extraction and assay of enzyme activities
Aliquots of 20 mg FW of ground plant material were extracted as in Hendriks et al. (2003). The composition of the extraction buffer was 10% (v/v) glycerol, 0.25% (w/v) BSA, 0.1% (v/v) Triton-X100, 50 mm HEPES/KOH (pH 7.5), 10 mm MgCl2, 1 mm EDTA, 1 mm EGTA, 1 mm benzamidine, 1 mmε-aminocapronic acid, 1 mm PMSF, 10 μm leupeptin and 1 mm DTT. PMSF was added just prior to extraction. The extracts were further diluted in the extraction buffer so that the dilution of FW was 1000 (w/v) in the assay. Assays were prepared in 96-well microplates using a Multiprobe II pipetting robot equipped with a cooling block, an incubation block set to 25°C, a shaker and a gripper (Perkin-Elmer).
ADP-glucose pyrophosphorylase was assayed in the reverse direction by measuring the PPi-dependent production of ATP from ADP-glucose in a final volume of 50 μl. The concentrations used for the substrates and for the activator 3-phosphoglycerate were chosen according to Merlo et al. (1993). Extracts, as well as ATP standards prepared in the extraction buffer and ranging from 0 to 1 nmol, were incubated in a medium containing: 50 mm HEPES/KOH (pH 7.5), 5 mm MgCl2, 1 U ml−1 glycerokinase, 0 (blank) or 1 mm (maximal activity) ADP-glucose, 5 mm 3-phosphoglycerate, 1.5 mm sodium fluoride and 120 mm glycerol. The reaction was started by the addition of PPi to a final concentration of 2 mm. The reaction was stopped with 20 μl of 0.5 m HCl.
Sucrose phosphate synthetase was assayed in the forward direction by measuring the fructose-6P dependent production of UDP from UDP-glucose. The concentrations used for fructose-6P, glucose-6P and phosphate were chosen according to Huber et al. (1989). Extracts, as well as UDP standards prepared in the extraction buffer and ranging from 0 to 1 nmol, were incubated in a medium containing: 50 mm HEPES/KOH (pH 7.5), 10 mm MgCl2, 1 m EDTA, 1 U ml−1 UMP-kinase, 1 U ml−1 glycerokinase, 10 mm UDP-glucose, 50 μm ADP, 1 mm AMP, 120 mm glycerol, 0 (blank) or 12 mm (maximal activity) fructose-6P, 0 (blank) or 36 mm (maximal activity) glucose-6P. The reaction was stopped with 20 μl of 0.5 m HCl. To measure SPS activation, an additional assay was used that contained the same reagents at the same concentrations, except fructose-6P (2 mm), glucose-6P (6 mm) and phosphate (5 mm).
Both stopped assays led to formation of glycerol-3-P. After neutralization (20 μl of 0.5 m NaOH), glycerol-3-P was measured in both assays as the final product as described in Gibon et al. (2002), in the presence of 1.8 U ml−1 glycerol-3-P oxidase, 0.7 U ml−1 glycerol-3-P dehydrogenase, 1 m NADH, 1.5 mm MgCl2 and 100 mm Tricine/KOH (pH 8), in a final volume of 100 μl. The absorbance was red at 340 nm and at 30°C in a Synergy microplate reader (Bio-Tek) until the rates were stabilized. The rates of reactions were calculated as the decrease of the absorbance in mOD min−1 by using the KC4 software (Bio-Tek).
Analysis of AGPase activation
Extraction of AGPase for blotting and procedures for gels and quantification were performed as in Hendriks et al. (2003).
RNA isolation expression analysis with 22K Affymetrix arrays and data evaluation
Isolation of total ribonucleic acid (RNA), cDNA synthesis, cRNA labelling and the hybridization on the GeneChip Arabidopsis ATH1 genome array was carried out exactly as described by Thimm et al. (2004) and as recommended by the manufacturer (part no. 900385; Affymetrix UK Ltd, High Wycombe, UK). The microarray suite software package (MAS 5.0; Affymetrix) was used to evaluate probe set signals of the array. The generated data files (.cel) were the input for the software package RMAExpress, to normalize and estimate raw signal intensities (Bolstad et al., 2003), combining the microarray hybridizations from all experiments described. Ratios of signal intensities were calculated on log2 base, for either all probe sets on the array or only for those probe sets, which gave present calls by MAS 5.0 in all microarray hybridizations. PCA analysis was carried out using TIGR MeV software (Saeed et al., 2003).
RNA was isolated from ground rosette material using TRIZOL reagent (Invitrogen GmbH, Karlsruhe, Germany), according to the manufacturer's instructions. RNA concentration was measured and 150 μg of total RNA was digested with RNase-free DnaseI. Absence of genomic DNA contamination was subsequently confirmed by PCR, using primers designed on intron sequence of a control gene (Actin: At3g18780). Reverse transcription reactions were performed using the cDNA synthesis kit from Promega. Efficiency of cDNA synthesis was assessed by real-time PCR amplification of control genes encoding actin2, ubiquitin10, β-6-tubulin. Only cDNA preparations that yielded similar CT values (e.g. 20 ± 1) for the control genes were used for comparing transcript levels of the genes studied here.
Primers were designed and synthesized at MWG Biotech AG (Ebersberg, Germany) as described by Czechowski et al. (2004). PCR reactions were performed in 96-well plate with an Gene Amp 5700 Sequence Detection System (Applied Biosystems, Foster City, CA, USA), using SYBR® Green (Eurogentec, Cologne, Germany) to monitor dsDNA synthesis. Reactions contained 12.5 μl 2X Master Mix buffer, 0.75 μl SYBR® Green, 1.0 ng cDNA and 200 nm of each gene-specific primer in a final volume of 10 μl. The following standard thermal profile was used for all PCR reactions: 50°C for 2 min; 95°C for 10 min; 40 cycles of 95°C for 15 sec and 60°C for 1 min. Data were analysed using the SDS 2.0 software (Applied Biosystems). To generate a baseline-subtracted plot of the logarithmic increase in fluorescence signal (ΔRn) versus cycle number, baseline data were collected between cycles 3 and 15. CT values for all analysed genes were normalized to the CT value of ubiquitin10 (At4g05320), which was the most constant of the three house-keeping genes included in each PCR run (actin2, At3g18780; β-6-tubulin At5g12250) PCR efficiency (E) was estimated as described by Czechowski et al. (2004).
This research was supported by the Max Planck Society and the BMBF funded project GABI Verbund Arabidopsis III Gauntlets, ‘Carbon and Nutrient Signalling: Test Systems, and Metabolite and Transcript Profiles’ (0312277A). Thanks are due to Dr Octavian Barzu (Institut Pasteur, Paris, France) for the generous gift of the clone encoding the UMP-kinase from Escherichia coli. Thanks are also due to Sonya Köhler for technical assistance and to Dr Florian Wagner at the RZPD company (Berlin, Germany) for carrying out the microarray hybridizations.
Table S1. Comparison of the expression of 105 genes encoding proteins involved in photosynthesis in wild-type Col0 and pgm, at the end of the day and the end of the night. The plants were grown in a 12 h light/12 h dark cycle, and harvested at the end of the night (these are the same samples as those given in Table 1) and the end of the day in two separate experiments. The log2 transformed signals were averaged for the biological replicates, and ratio between the signal in pgm and the signal in Col0 calculated for the end of the night (EN) and the end of the day (ED) (see Figure 1)
Table S2. Gene accession references (AGI codes) and oligonucleotide sequences used in real-time RT-PCR measurements
Table S3. Inclusion of low concentrations of sucrose in the medium inhibits the accumulation of starch in short days but not in long days. Levels of starch at the beginning and end of the day in Arabidopsis Col0 plants grown in nutrient agar added with various concentrations of sucrose on horizontal plates. The results are the mean ± SD (n = 5 separate samples, each of four individual plants)
Table S4. Original data from microarray hybridisations