ADP-Glucose Pyrophosphorylase Is Activated by Posttranslational Redox-Modification in Response to Light and to Sugars in Leaves of Arabidopsis and Other Plant Species^,

Abstract

ADP-glucose pyrophosphorylase (AGPase) catalyzes the first committed reaction in the pathway of starch synthesis. It was recently shown that potato (Solanum tuberosum) tuber AGPase is subject to redox-dependent posttranslational regulation, involving formation of an intermolecular Cys bridge between the two catalytic subunits (AGPB) of the heterotetrameric holoenzyme (A. Tiessen, J.H.M. Hendriks, M. Stitt, A. Branscheid, Y. Gibon, E.M. Farré, P. Geigenberger [2002] Plant Cell 14: 2191–2213). We show here that AGPase is also subject to posttranslational regulation in leaves of pea (Pisum sativum), potato, and Arabidopsis. Conversion is accompanied by an increase in activity, which involves changes in the kinetic properties. Light and sugars act as inputs to trigger posttranslational regulation of AGPase in leaves. AGPB is rapidly converted from a dimer to a monomer when isolated chloroplasts are illuminated and from a monomer to a dimer when preilluminated leaves are darkened. AGPB is converted from a dimer to monomer when sucrose is supplied to leaves via the petiole in the dark. Conversion to monomeric form increases during the day as leaf sugars increase. This is enhanced in the starchless phosphoglucomutase mutant, which has higher sugar levels than wild-type Columbia-0. The extent of AGPB monomerization correlates with leaf sugar levels, and at a given sugar content, is higher in the light than the dark. This novel posttranslational regulation mechanism will allow starch synthesis to be regulated in response to light and sugar levels in the leaf. It complements the well-characterized regulation network that coordinates fluxes of metabolites with the recycling of phosphate during photosynthetic carbon fixation and sucrose synthesis.

During photosynthesis, triose-phosphates (triose-P) are exported to the cytosol where they are converted to end products, including Suc. This releases inorganic orthophosphate (Pi), which is recycled to the chloroplast in counterexchange with triose-P (Edwards and Walker, 1983). Some of the photosynthate is retained in the chloroplast to synthesize starch. Leaf starch represents a transient store, which is remobilized during the night to support leaf metabolism, and continued synthesis and export of Suc (Geiger and Servaites, 1994). Its importance is demonstrated by the phenotype of starch-deficient mutants, which grow poorly or die in short-day conditions (Caspar et al., 1986; Schulze et al., 1991; Geiger et al., 1995; Sun et al., 2002). A consensus has developed that leaf starch synthesis is regulated by changes in the levels of phosphorylated metabolites and Pi that are generated when the rate of photosynthesis increases or when rising levels of sugars lead to feedback regulation of Suc synthesis. This paper presents evidence that light and sugars also regulate starch synthesis more directly via redox-dependent posttranslational activation of ADP-Glc pyrophosphorylase (AGPase).

AGPase catalyzes the first committed step in the pathway of starch synthesis (Preiss, 1988; Martin and Smith, 1995). The higher plant enzyme is a heterotetramer that contains two “regulatory” (AGPS, 51 kD) and two slightly smaller “catalytic” (AGPB, 50 kD; Morell et al., 1987; Okita et al., 1990) subunits. AG-Pase is exquisitely sensitive to allosteric regulation, with glycerate-3-phosphate (3PGA) acting as an activator and Pi as an inhibitor (Sowokinos, 1981; Sowokinos and Preiss, 1982; Preiss, 1988). Studies with isolated chloroplasts led to the concept that starch synthesis is stimulated when low Pi restricts carbon export from the plastid (Heldt et al., 1977). In these conditions, ATP falls, leading to an inhibition of 3PGA reduction. A rising 3PGA to Pi ratio provides a sensitive signal that carbon fixation is exceeding the rate of export, and activates AGPase. An analogous situation arises in leaves when the rate of end product synthesis falls below the rate of photosynthesis. For example, feedback regulation of Suc synthesis will lead to the accumulation of phosphorylated intermediates, depletion of Pi, and activation of AGPase by the rising 3PGA to Pi ratio, resulting in a compensatory stimulation of starch synthesis (see Herold, 1980; Stitt et al., 1987).

For more than a decade, this biochemical model has provided the framework to explain how the photosynthate allocation between Suc and starch is regulated. Support has been provided by biochemical analyses of changes of metabolites, enzyme activities, and fluxes during the diurnal cycle (Gerhardt et al., 1987; Stitt et al., 1988), by genetic evidence that AGPase colimits starch synthesis in leaves (Neuhaus et al., 1990), and by genetic evidence that reduced expression of various enzymes in the pathway of Suc synthesis (Neuhaus et al., 1989; Neuhaus and Stitt, 1990; Zrenner et al., 1996; Geigenberger and Stitt, 2000; Scott et al., 2000; Häusler et al., 2000; Draborg et al., 2001) or overexpression of Suc phosphate synthase (SPS; Baxter et al., 2001; Laporte et al., 2001) leads to the predicted stimulation and inhibition, respectively, of starch synthesis. Details of the biochemical mechanisms that contribute to the feedback inhibition of Suc synthesis have been elucidated. In spinach (Spinacia oleracea), rising Suc leads to posttranslational inhibition of SPS (Stitt et al., 1988; Neuhaus and Stitt, 1990). In other species, Suc is hydrolyzed to reducing sugars, which are rephosphorylated (Huber, 1989; Goldschmidt and Huber, 1992). Rising cytosolic hexose phosphates (Gerhardt et al., 1987) lead via activation of Fru-6-phosphate 2-kinase and inhibition of Fru-2,6-bisphosphatase (Stitt et al., 1987; Villadsen and Nielsen, 2001; Markham and Kruger, 2002) to an increase of Fru-2,6-bisphosphate, which inhibits cytosolic Fru-1,6-bisphosphatase (cFBPase; Stitt et al., 1987; Neuhaus et al., 1989, 1990). The predicted decrease of ATP and increase of 3PGA has been confirmed when Suc synthesis is inhibited by decreased expression of transporters and enzymes in the pathway (see above), agents that sequester phosphate (Stitt et al., 1987) and low temperature (Stitt and Grosse, 1988).

Curiously, the evidence is less convincing for treatments that modify partitioning by altering sugar levels in the leaf. Starch synthesis was stimulated in the absence of an increase of 3PGA when sugars were supplied to detached spinach leaves (Krapp et al., 1991), when spinach leaves were cold-girdled to decrease export (Krapp and Stitt, 1995), and when phloem transport was inhibited by phloem-specific expression of Escherichia coli pyrophosphatase in tobacco (Nicotiana tabacum; Geigenberger et al., 1996). In at least some conditions, decreased expression of SPS in Arabidopsis leads to decreased rather than increased starch synthesis (Strand et al., 2000). Further, transgenic potato (Solanum tuberosum) plants with increased levels of 3PGA due to antisense inhibition of cytosolic phosphoglycerate mutase did not show any increase of starch in their leaves (Westram et al., 2002). These results indicate that there are gaps in our understanding of the regulation of photosynthate partitioning.

When potato AGPB and AGPS are heterologously overexpressed in E. coli, an intermolecular bridge forms between the cys82 residues of the two AGPB subunits. To obtain active enzyme, it was necessary to incubate the complex with dithiothreitol (DTT) or thioredoxin to break this link (Fu et al., 1998; Ballicora et al., 2000). It was recently shown that an analogous process occurs in planta in potato tubers (Tiessen et al., 2002). In both cases, reduction of the intermolecular bridge leads to a dramatic increase of activity, due to a decrease of the K_m(ATP) and increased sensitivity to activation by 3PGA. Activation of AGPase in planta correlated closely with the tuber Suc content across a range of physiological and genetic manipulations, indicating that redox modulation is part of a novel regulatory loop that channels incoming Suc toward synthesis of storage starch (Tiessen et al., 2002). Crucially, it allows the rate of starch synthesis to be increased in response to external inputs and independently of any increase in the levels of glycolytic intermediates. The following paper asks whether AGPase is regulated by an analogous mechanism in leaves and investigates its contribution to the regulation of photosynthetic metabolism.

RESULTS

AGPB Expressed in Leaves Contains a Conserved N-Terminal Cys

Almost all dicotyl plant AGPB sequences contain a conserved SQTCLDPDAS motif at the N terminus, which includes the Cys shown by Fu et al. (1998) to be involved in formation of the intermolecular Cys bridge in the potato enzyme (see Supplementary Material). The only exception is Sumatra orange (Citrus unshui). Monocots contain two types of AGPB transcript: One encodes proteins that contain this motif, and the other encodes proteins that lack it. The latter may represent cytosolic isoforms that occur in the endosperm of growing cereal seeds (Sikka et al., 2001; Burton et al., 2002).

The full genome sequence for Arabidopsis contains one reading frame (At5g48300) with a high homology to AGPB in other higher plants. A second open reading frame, which is annotated as a putative AGPB (At1g05610), shows considerable deviations from all other AGPB sequences (see Supplementary Material). Diversification occurs throughout the sequence and includes the loss of many highly conserved amino acids. At1g05610 is the only gene that falls outside of the group for plant AGPB sequences when a tree is calculated with ClustalX (Thompson et al., 1997), using only those stretches of the alignment selected for tree calculation by Gblocks 0.91b (Castresana, 2000). It even does not group with plant AGPS and bacterial AGP sequences (Supplementary Material). We therefore suspect that this gene does not code for a functional AGP protein. Further, RNA primers designed to distinguish between transcripts for these two genes detected At5g48300 but not At1g05610 in Arabidopsis leaf extracts (data not shown). We conclude that an AGPB protein that contains the conserved Cys is expressed in the leaves of most if not all plants.

AGPB Exists as a Dimer in the Dark and Becomes Monomerized after Illumination of Potato, Arabidopsis, and Pea (Pisum sativum) Leaves

Pea, potato, and Arabidopsis leaves were harvested during the second half of the light period and toward the end of the dark period to investigate whether leaf AGPB undergoes reversible dimerization in vivo. Extracts were rapidly prepared in degassed SDS solutions and subjected to non-reducing SDS-PAGE, and AGPB protein was detected using a rabbit-antibody raised against AGPB from potato (see Tiessen et al., 2002). When extracts from growing potato tubers are analyzed in this way, they contain a mixture of monomeric (molecular mass of approximately 50 kD) and dimeric (molecular mass of approximately 100 kD) AGPB (Tiessen et al., 2002). In initial experiments with leaves, all of the AGPB protein ran with an apparent molecular mass of about 100 kD, irrespective of whether extracts were prepared from illuminated or darkened plants (data not shown). When extracts from leaves were mixed 1:1 with extracts from growing potato tubers, the immunosignal was also obtained at only 100 kD (data not shown). These results indicated that leaf extracts contain unknown compounds that rapidly oxidize AGPB to a dimer. To prevent this, leaves were extracted in a trichloroacetic acid (TCA)-diethyl ether mixture to rapidly denature AGPase and physically separate any AGPB subunits that were present as monomers. This new procedure revealed that pea leaf AGPB is completely dimerized in the dark, and partly converted to a monomer in the light (Fig. 1A). When the extracts were separated in a reducing gel (including DTT), immunosignal was found only at 50 kD, showing that the intermolecular link involves a Cys bridge. Similar results were obtained for potato and Arabidopsis (Fig. 1B) leaves. The proportion converted to a monomer was lower in Arabidopsis, possibly reflecting the lower growth light intensity.

Figure 1.

Dimerization of AGPB varies between the day and night in potato, pea, and Arabidopsis leaves. A, Western of pea leaf (cv Marcia) tissue harvested during the second half of the day (day) and at the end of the night (night). Samples were prepared with TCA ether and run directly (non-reducing) or after adding 4 mm DTT to part of the sample (reducing). B, Non-reducing westerns of leaf material of Arabidopsis and potato harvested during the day and the night.

AGPB Dimerization Is Accompanied by a Decrease of AGPase Activity

In potato tubers, dimerization increases the K_m(ATP) and decreases sensitivity to activation by 3PGA (Tiessen et al., 2002). This change is reversed by incubation with DTT. We investigated whether dimerization of Arabidopsis leaf AGPase is also accompanied by changes in AGPase activity. Appearance of the monomer was accompanied by an increase in AGPase activity when assayed in absence of 3PGA but not in the presence of saturating amounts of 3PGA (Fig. 2A). The ratio of activity in the two conditions (–3PGA/+3PGA) changed from 0.15 in the dark to 0.67 in the light. Illumination led to a marked increase of the affinity for ATP in the absence of 3PGA, which could be overcome at high 3PGA concentrations (Fig. 2, B–C). The sensitivity to activation by 3PGA is also changed. Whereas AGPase from illuminated leaves attained significant activities in the absence of 3PGA and was stimulated 4- to 10-fold by 3PGA depending on the ATP level, activity of AGPase from leaves at the end of the night was very low in the absence of 3PGA and was stimulated 10-fold by 3PGA in the presence of high ATP and up to 25-fold in the presence of low ATP (Fig. 2D). The small increase of overall AGPase activity in the dark is frequently seen in Arabidopsis leaves and reflects changes in the amount of AGPase protein (data not shown).

Figure 2.

Increased monomerization of AGPB in the light leads to a change in the kinetic properties of AGPase. A, AGPase activity in presence of 0 or 1 mm 3PGA in presence of 1.5 mm ATP and 1.5 mm G1P in Arabidopsis leaves harvested during the end of the night (▪) or the second half of the day (□). B, ATP substrate saturation curves of AGPase from Arabidopsis leaves harvested at the end of the day (▿, ○) and end of the night (▾, •) assayed in presence of no (○, •) or 3 mm 3PGA (▿, ▾) and 1.5 mm G1P. C, Lineweaver-Burk presentation of the data in B. D, The activation factor by 3 mm 3PGA compared with 0 mm 3PGA for the day (□) and the night sample (▪). The data points in presence of 3 mm 3PGA and 5 mm ATP are omitted from C and D, because they showed substrate inhibition. Leaves were taken from 8-week-old plants.

Light-Dark Transitions Are Accompanied by Rapid Changes in AGPB Dimerization

The appearance of AGPB monomer during the day in leaves could be due to illumination or it could be an indirect effect due, for example, to leaves containing more sugars in the light. To investigate whether there are rapid light-dependent changes in AGPB monomerization, we darkened pre-illuminated plants. This treatment was chosen because it leads to an abrupt change in photosynthesis, whereas illumination leads to only slow changes due to the need to induce photosynthesis and increase stomatal conductance.

Arabidopsis plants were illuminated for 6.5 h, samples were taken in the light, the remaining plants were darkened, and samples taken 6, 15, and 60 min later. In the light, a small proportion of AGPB was present as a monomer (Fig. 3A; see also Fig. 1B). After darkening, the monomer decreased within 6 min and almost totally vanished within 15 min. Sugar levels were measured in the same leaf material (Fig. 3B). There were no significant changes of Suc, Glc, or Fru levels in the first 6 min and only small changes in the first 60 min after darkening. Similar results were obtained for pea plants (data not shown).

Figure 3.

Darkening rapidly reverses the light-dependent monomerization of AGPB. A, Non-reducing western blot of leaf samples of 6.5-week-old Arabidopsis harvested 6.5 h into the day, and after darkening the plants for 6, 15, and 60 min. B, Sugar content in these leaves (▪, Glc;, Fru; and, Suc).

Light Leads to Monomerization of AGPB in Isolated Pea Chloroplasts

To provide independent evidence that light promotes monomerization of AGPB protein, we investigated the responses in isolated chloroplasts. Chloroplasts do not contain or synthesize Suc or other sugars. These experiments were carried out with chloroplasts from young pea plants. Pea chloroplasts have the advantage that it is possible to manipulate the adenylate content. Addition of inorganic pyrophosphate (PPi) leads to the loss of adenylates from the chloroplast, which can be reversed by adding ATP or ADP (Lunn and Douce, 1993).

AGPB occurred almost exclusively as a dimer when chloroplasts were incubated in the dark with PPi, ATP, and 3PGA (Fig. 4A). A large proportion was converted to monomer after 6 min of illumination. This paralleled the increase of plastid FBPase activity (Fig. 4B). Addition of 30 mm Suc to isolated chloroplasts did not lead to monomerization of AGPB in the dark over a 15-min period (data not shown).

Figure 4.

Changes in dimerization of AGPB also occur in isolated pea chloroplasts. A and B, Non-reducing western-blot (A) and FBPase activity (B) of chloroplasts incubated in the dark, and 3 and 6 min after turning the lamp on in presence of 0.67 mm NaPPi, 1 mm 3PGA, and 1 mm ATP. C through E, Non-reducing western-blot (C), FBPase (D), and NADP-malate dehydrogenase activity (E) in chloroplasts in the dark (▪) and after 6 min in the light (□) in the presence of 0.67 mm PPi, 1 mm 3PGA, and 1 mm ATP (all), or when either 3PGA or ATP was left out of the incubation medium.

In vitro experiments with heterologously expressed potato tuber AGPase have shown that monomerization can be mediated by thioredoxins (Ballicora et al., 2000). In many cases, substrate levels modulate the activation of thioredoxin-regulated enzymes (Scheibe, 1991: Stitt, 1996; Schürmann and Jacquot, 2000). Activation of several Calvin cycle enzymes is promoted by high substrate concentrations, and activation of NADP-malate dehydrogenase by a high NADPH to NADP ratio. We investigated the effect of illuminating pea chloroplasts in full medium and in the absence of 3PGA or ATP on AGPB monomerization. Appearance of the monomer was suppressed when 3PGA was omitted and was stimulated when ATP was omitted (Fig. 4C). In contrast, activation of plastidic FBPase was high in the absence of 3PGA but decreased when ATP was omitted (Fig. 4D). Activation of NADP-malate dehydrogenase was relatively low in full medium, increased when 3PGA was omitted, and rose further when ATP was omitted (Fig. 4E). These results indicate that light-dependent monomerization of AGPB does not require high levels of ATP or a high NADPH to NADP ratio, but is promoted when metabolites, in particular 3PGA, are high.

Supplying Suc to Leaves in the Dark Leads to Conversion of AGPB from a Dimer to a Monomer and to Increased Rates of Starch Synthesis

A second set of experiments was carried out to investigate whether sugars promote AGPB monomerization. Leaves were harvested from Arabidopsis plants at the end of the normal day and supplied via their petiole with zero, 50, 100, or 200 mm Suc for 13 h in the dark (Fig. 5). AGPB was present almost exclusively as a dimer in leaf material at the end of the night and in leaves incubated in the dark without sugars. Suc led to the appearance of monomer. The proportion converted by Suc in the dark was similar to that seen in the light under normal growth conditions (Fig. 5A).

Figure 5.

AGPB is converted into the monomeric form by supplying sugars to leaves in the dark. Leaves from 8-week-old Arabidopsis plants were fed via their petioles with buffer and varying concentrations of Suc in the dark during their natural night (0, 50, 100, and 200 mm). For comparison, leaves harvested from intact plants at the start (end of day [ED]) and at the end of the experiment (end of night [EN]) are also shown. A, Non-reducing western blot of AGPB; B, sugar content (▪, Glc;, Fru; and, Suc); C, 3PGA content; and D, starch content of the leaves. E, In parallel incubations, high specific [U-¹⁴C]Glc was supplied together with the various concentrations of unlabeled Suc, to investigate the rate of starch synthesis.

Feeding sugars led to a progressive increase in the levels of sugar (Fig. 5B), but 3PGA remained unaltered (Fig. 5C). There was also an increase of starch (Fig. 5D). This might be due to a stimulation of starch synthesis or to slower breakdown of starch during the 13-h dark treatment. To measure the rate of starch synthesis, the unlabeled Suc was spiked with high specific activity [¹⁴C]Glc. The rate of starch synthesis was calculated by dividing the label incorporated into starch by the specific activity of the hexose phosphate pool (for a detailed discussion of this approach, see Geigenberger et al., 1997). Suc feeding led to a concentration-dependent stimulation of starch synthesis in the dark (Fig. 5E).

Time-of-Day Dependent Changes in AGPB Dimerization in the Starch-Deficient Phosphoglucomutase Mutant (pgm) Show That Light and Leaf Sugar Levels Interact to Regulate AGPB Activation

To provide further evidence that sugars increase monomerization of AGPB, we carried out a set of experiments comparing diurnal changes in wild-type Columbia-0 (Col0) and the pgm mutant (Caspar et al., 1986). The pgm mutant is deficient in starch synthesis, due to a loss-of-function mutation in a unique gene encoding plastid phosphoglucomutase (Kofler et al., 2000). In leaves of wild-type Col0, sugars rise to a plateau soon after illumination (Fig. 6, A–B, Fru not shown). Starch accumulates in a linear manner through the light period and is degraded during the night (Fig. 6C). In the pgm mutant (see also Caspar et al., 1986), large amounts of sugars accumulate in the leaf during the light period. They are depleted during the night, falling to levels at the end of the night that are lower than in wild-type Col0 (Fig. 6, A–B).

Figure 6.

Monomerization of AGPB is increased in the starch-deficient pgm mutant. A to C, Glc (A), Suc (B), and starch content (C) in Arabidopsis leaves of 6-week-old Col0 (•) and 11-week-old pgm (○) during a night/day cycle (indicated by black and white bar above the figures). D, The reduction state of the AGPase antigen in the same samples. E and F, Non-reducing western of leaves of Col0 (E) and pgm (F) at the change of light. G, Non-reducing western of pgm samples harvested at the indicated times after the start of the illumination.

In the same samples, the proportion of AGPB present as monomer was determined on western blots and quantified after scanning the films (Fig. 6D). Typical examples of immunoblots are shown in Figure 6, E through G. In wild-type Col0, AGPB is present almost exclusively as dimer at the end of the night, did not show a marked shift after 15 min illumination, was gradually converted to a monomer as the day progressed, and rapidly reverted to dimer after darkening (see also Fig. 3). The response was markedly changed in pgm. AGPB became partly monomerized within 15 min after illumination. Monomerization increased further during the next 2 to 3 h and by the second part of the light period AGPB was almost totally converted to monomer. 3PGA levels were comparable with those in Col0 (data not shown). After darkening, a substantial proportion of AGPB remained as monomer for the first 2 h of the night.

The data in Figure 6, A, B, and D, are replotted in Figure 7 to show the relation between monomerization of AGPB and total sugars. When extracts from darkened leaves are compared, there is a correlation between leaf sugar levels and the appearance of AGPB monomer. Illumination leads to increased monomerization at a given sugar content in the dark.

Figure 7.

Relation between AGPB dimerization and the leaf sugar content in the light and dark. The data from Figure 6 were replotted showing the relation between the AGPB dimerization state and the total sugar content (sum of Fru, Glc, and Suc) for Col0 (•, ○) and pgm (▾, ▿) during the night (•, ▾) and the day (○, ▿).

AGPase activity was measured with limiting ATP (0.2 mm) in the presence and absence of 1 mm 3PGA in extracts from wild-type Col0 harvested at the end of the night and pgm harvested at midday. These represent the most extreme changes obtained in our experiments. The shift from dimer to monomer was accompanied by a 7-fold stimulation of activity (from 25 to 176 nmol min^–1 g^–1 fresh weight) in the absence of 3PGA, whereas activity was not affected in the presence of 3PGA (583 and 559 nmol min^–1 g^–1 fresh weight, respectively).

DISCUSSION

A consensus has developed that starch synthesis is regulated in response to changes of metabolism in the cytosol (see introduction). When the rate of triose-P use for the synthesis of Suc and other end products is lower than the rate of photosynthesis, falling Pi is proposed to lead to a restriction of ATP synthesis and 3PGA reduction. The resulting increase of the 3PGA to Pi ratio activates AGPase, leading to an increased rate of starch synthesis and increased recycling of Pi within the chloroplast.

Tiessen et al. (2002) recently discovered that AGPase is subject to posttranslational regulation in potato tubers. The mechanism involves formation of an inter-molecular Cys bridge between the AGPB subunits in the AGPase heteroteramer. Rising Suc leads to monomerization and activation of AGPase, channeling carbon toward starch and away from respiration when the Suc supply increases. The experiments in the present paper show that AGPase is also subject to posttranslational redox regulation in leaves of Arabidopsis, potato, and pea. As in potato tubers, this involves reversible interconversion between a less active form in which AGPB is present as a dimer and an active form in which AGPB is present as monomers. To routinely monitor the dimerization state, we used an extreme extraction method in which trichloroacetic acid is used to rapidly dissociate the protein and physically separate AGPB monomers to prevent post extracto formation of a intermolecular Cys bridge. This was essential, because the AGPB monomer is rapidly converted to a dimer in leaf extracts. This lability probably explains why this important posttranslational mechanism was overlooked in earlier studies.

The shift from a dimer to a monomer is accompanied by an increase in leaf AGPase activity. Preiss and coworkers have shown for heterogeneously overexpressed potato tuber AGPase that DTT or thioredoxin lead to monomerization of AGPB and a concomitant increase in AGPase activity (Fu et al., 1998; Ballicora et al., 2000). DTT also led to monomerization of AGPB and an increase of AGPasae activity in potato tuber extracts (Tiessen et al., 2002). We have not yet been able to provide direct evidence that the shift from dimer to monomer causes the change in leaf AGPase activity, because DTT seriously interferes with AGPase assay in leaf extracts (data not shown). However, it is a reasonable assumption that the changes in monomerization and AGPase activity are causally related.

The increase of AGPase activity involved a change in the kinetic properties, including an increased affinity for ATP and altered sensitivity to regulation by 3PGA. The increase in activity is less marked than Tiessen et al. (2002) reported for potato tuber AGPase. This may be due to technical difficulties in retaining AGPB in the in planta status in leaf extracts (see above). Alternatively, it may reflect a real difference in sensitivity between AGPase in different plants or organs due, for example, to association with a different AGPS isoform. It will be necessary to carry out detailed studies with purified AGPase to resolve this point.

At least two inputs modulate the posttranslational redox-activation of AGPase in leaves. The first input is a light-dependent signal. This is analogous to the way that several Calvin cycle enzymes and other proteins involved in photosynthesis are regulated (Scheibe, 1991; Schürmann and Jacquot, 2000). Evidence for a light-dependent input is provided by two independent observations: AGPB monomerization decreases rapidly after darkening wild-type leaves even though sugar levels do not change and increases rapidly after illumination of isolated chloroplasts. The second input is a sugar-related signal, which is analogous to the situation in potato tubers. Evidence is provided by two independent observations: AGPase monomerization is increased by supplying exogenous sugars to wild-type leaf material in the dark and is also increased in starch-deficient pgm mutants in the light and at the start of the night when this mutant contains higher levels of sugars than wild-type plants. Comparison of the relation between light, internal sugar levels, and AGPase monomerization in wild-type Col0 and the starch-deficient pgm mutant indicates that light and sugars act in an additive manner to increase AGPase activation.

The reductive activation of heterologously expressed AGPase can be mediated in vitro by thioredoxin (Fu et al., 1998; Ballicora et al., 2000). Arabidopsis contains a family of thioredoxins, of which several are targeted to the plastid. Whereas Calvin cycle enzymes are activated by thioredoxin-f, other targets including NADP-malate dehydrogenase, cfATPase, and Rubisco activase are regulated preferentially by thioredoxin-m. Further studies are needed to identify which thioredoxin interacts with AGPB. The light-dependent redox activation of AGPase can be envisaged to be a direct result of increased reduction of thioredoxin. It is however not yet clear how increased levels of Suc modify this process (for discussion, see Tiessen et al., 2002). Intriguingly, antisense inhibition of a SNF1-homolog strongly attenuates the reductive activation of AGPase after adding Suc in potato tubers (Tiessen et al., 2003). This implies that the transduction pathway that regulates the reductive activation of AGPase in plastids and the regulatory network that controls the expression and phosphorylation of cytosolic enzymes have some common components.

The light-dependent activation of photosynthetic enzymes by thioredoxin is modulated by metabolites, which modify the mid-redox potential of the Cys in the target protein (Scheibe, 1991; Schürmann and Jacquot, 2000). This provides an elegant mechanism to fine-tune the activity of enzymes at different sites around the Calvin cycle and poise ATP and NADPH levels production (Stitt, 1996). Monomerization of AGPase in isolated chloroplasts is promoted by 3PGA but not by ATP. This indicates that reductive activation of AGPase may be promoted by high 3PGA in leaves. This could provide a mechanism to prevent depletion of phosphorylated intermediates due to excessive posttranslational activation of AGPase. It should, however, be noted that an increase of 3PGA level is not involved in the posttranslational regulation of AGPase in response to sugars in leaves (see Figs. 4 and 5; also comments in the preceding paragraph) or potato tubers (Tiessen et al., 2002).

Posttranslational regulation of AGPase allows starch synthesis to be modulated in response to light or the accumulation of sugars, without any requirement for changes in the levels of phosphorylated intermediates or Pi. Maintenance of an appropriate balance between phosphorylated intermediates and Pi is of crucial importance during photosynthesis. Triose-P are exported from the chloroplast and converted into Suc in the cytosol, and the Pi that is released is recycled to the chloroplast to support further photosynthesis. Excessive triose-P export will inhibit photosynthesis because it depletes the levels of Calvin cycle intermediates and inhibits regeneration of the CO₂ acceptor ribulose-1,5-bisphosphate, and inadequate triose-P export will inhibit photosynthesis because Pi is sequestered in phosphorylated intermediates leading to depletion of free Pi and an inhibition of ATP synthesis (Edwards and Walker, 1983). Because the exchange of triose-P and Pi via the triose phosphate:phosphate translocator (Häusler et al., 2000) is a passive process, the rate of triose-P export depends on the rate of consumption in the cytosol (Stitt et al., 1987; Stitt, 1996). Sophisticated mechanisms act on the cFBPase and SPS to coordinate the rate of Suc synthesis with the rate of photosynthesis (Stitt et al., 1987; Stitt, 1996). Rising levels of triose-P and 3PGA and falling Pi inhibit Fru-6-phosphate,2-kinase and stimulate Fru-2,6-bisphosphatase. The resulting decrease of Fru-2,6-bisphosphate (Stitt et al., 1987; Villadsen and Nielsen, 2001; Markham and Kruger, 2002) in combination with rising Fru-1,6-bisphosphate, stimulates cFBPase activity (Stitt et al., 1987; Stitt, 1997), leading to increased synthesis of hexose phosphates. A rising Glc-6-phosphate to Pi ratio activates SPS allosterically (Stitt et al., 1987) and modulates protein kinase and phosphatase activities resulting in posttranslational activation of SPS (Toroser et al., 2000; Winter and Huber, 2000). Feedback regulation of Suc synthesis (see introduction) in effect reverses this chain of events. This carries a concomitant risk that it may inhibit photosynthesis. Posttranslational regulation of AGPase by light and leaf sugar levels will stabilize the regulation network, because it allows partitioning between Suc and starch to be altered without this necessarily requiring changes in the levels of phosphorylated intermediates and Pi. As pointed out in the introduction, there are several reports in the literature in which starch synthesis changed independently of overall AGPase activity and the levels of phosphorylated intermediates. They all involved manipulations that alter sugar levels in the leaf and, according to the results in the present paper, will therefore probably have led to posttranslational redox regulation of AGPase.

In conclusion, three mechanisms interact to regulate AGPase activity in leaves. (a) Allosteric regulation allows instantaneous changes of AGPase activity when the 3PGA to Pi ratio changes. Although it may in some conditions be part of a regulatory sequence that links sugar accumulation to an increase of starch synthesis, its main significance is more likely to be to rapidly increase the recycling of Pi in the stroma when there is a transient imbalance between photosynthesis and triose P export. (b) Posttranslational redox regulation provides a mechanism that allows direct light activation of starch synthesis in leaves and also allows starch synthesis to be increased when sugars accumulate in the leaf. Crucially, this mechanism allows starch synthesis to be increased without an increase of the 3PGA to Pi ratio as a necessary intervening step. This will increase the flexibility of the regulatory network, because it allows photosynthetic carbon allocation to be regulated independently of the poising of intermediary photosynthetic metabolism. (c) Expression of AGPB and AGPS is increased by sugars (Salanoubat and Belliard, 1989; Müller-Röber et al., 1990; Sokolov et al., 1998) and decreased by nitrate (Scheible et al., 1997) and phosphate (Nielsen et al., 1998). Transcriptional control may operate mainly to allow starch accumulation to respond to sustained changes in the carbon or nutrient status of the plant.

MATERIALS AND METHODS

All experiments were reproduced at least once with independent biological material. Data points are at least the average of duplicate measurements of the same biological sample. When error bars are shown, they represent the sd of the average of the measurements on at least two biological samples of the same experiment.

Plant Growth

Pea (Pisum sativum cv Marcia) was grown either in a greenhouse with a 16-h day of 180 μE, 21°C/19°C (day/night), and 50% humidity or in a high-light phytotron with a 14-h day, 20°C/16°C, and 60%/75% humidity. The pea cv Kelvedon Wonder was grown in a short-day phytotron (8-h day of 180 μE, 20°C/16°C, and 60%/75% humidity day/night). Arabidopsis var Col0, wild type, and a plastidic pgm (Caspar et al., 1986) were grown in the same short-day phytotron. At least 2 weeks before their use, the plants were transferred into a small growth cabinet with a 10-h day of 160 μE and 20°C throughout the day/night cycle. Potato (Solanum tuberosum cv Desiree) plants were grown in a greenhouse at 400 μE, 20°C/16°C day/night, and 50% humidity throughout.

Harvesting Procedure, Sample Storage

Leaves were harvested while leaving the plants in place. Only source leaves that were not shaded by other leaves were selected. The leaves were put directly into liquid nitrogen, and stored at –80°C until use.

Incubation of Leaves with Sugars in the Dark

At the end of the light period, plants were taken from the growth cabinet. Non-shaded source leaves were cut, and their petioles were recut under buffer solution. The recut petioles were inserted into the feeding solution, containing 2 mm MES, pH 6.5, and varying concentrations of Suc. The leaves were returned to the growth cabinet and incubated there during the night. At the end of the night, leaves were frozen immediately in liquid nitrogen, after excising that part of the petiole, which had been immersed in the feeding solution.

Chloroplast Preparation

Ten- to 16-d-old pea seedling were subjected to an extended night by 5 h to deplete the internal starch pools and subsequently were transferred to light for about 30 min to induce photosynthesis. Chloroplasts were then prepared essentially as described by Lunn et al. (1990), but using 10 mm MES, pH 6.5, as buffer in the blending medium. The chlorophyll content of the final preparation was determined in MeOH extracts (Porra et al., 1989).

Incubations of Chloroplasts, Measurement of Photosynthesis

Oxygen evolution was measured in an oxygen electrode at 25°C on a chloroplasts suspension of 50 μg chlorophyll mL^–1 in resuspension buffer containing 4 mm HCO₃ and additives as indicated in the figure legends. The cuvette was darkened for 5 min before the dark sample was taken. After restabilization of the evolution trace, the sample was illuminated using the beam of a slide projector.

Extraction of AGPase for Blotting, Procedures for Gels

Frozen leaf material was homogenized using a liquid nitrogen cooled ball-mill, and 50 mg of leaf material was extracted in cold 16% (w/v) TCA in diethyl ether, mixed, and stored at –20°C for at least 2 h. The pellet was collected by centrifugation at 13,000 rpm for 5 min at 4°C. The pellet was washed three times with ice-cold acetone, dried briefly under vacuum, and resuspended in 1× Laemmli sample buffer containing no reductant (Laemmli, 1970). After heating the sample for 3 min at 95°C, the insoluble material was settled by a 1-min spin, and the supernatant was used for gel electrophoresis on 10% (w/v) acrylamide gels in presence of SDS. Proteins coming from 0.5 or 1 mg of fresh weight were loaded per small or broad lane, respectively. The gels were blotted onto polyvinylidene difluoride according to standard procedures. AGPase antigen was detected using a primary rabbit antibody raised against the His-tagged AGPB of potato (Tiessen et al., 2002) and a peroxidase-conjugated secondary goat anti-rabbit antibody (Bio-Rad Laboratories, Hercules, CA). The peroxidase was detected on film using the ECL kit of Amersham Biosciences (Uppsala). To quantify the amount of AGPase present as monomer, the films were scanned with standardized settings, saved as tif files, and analyzed with Tina 2.10i software (Raytest Isotopenmessgeräte GmbH, Straubenhardt, Germany). Gel samples from chloroplasts were prepared by mixing 1 volume of chloroplasts, taken directly from the oxygen electrode, with 1 volume of 2× Laemmli sample buffer without reductant. The samples were heated at 95°C for 3 min and stored at room temperature until use. Proteins coming from 6.25 μg of chlorophyll were loaded per lane. Electrophoresis, blotting, and immunolabeling procedures were as described above. For detection of the peroxidase, the ECL or ECL advance kit was used (Amersham Biosciences).

Extraction and Assay of AGPase

Activity measurements were performed essentially as described (Tiessen et al., 2002): Fifty milligrams of material was extracted with 0.5 mL of extraction buffer (50 mm K-HEPES, pH 7.5, 5 mm MgCl₂, 1 mm EGTA, 1 mm EDTA, 1 mm benzamidine, and 1 mm ε-aminocaproic acid). The sample was centrifuged for 30 s at 4°C. The supernatant was used directly at 1/10 or 1/20 of the volume in the activity assay containing 50 mm K-HEPES, pH 7.5, 5 mm MgCl₂, 1.5 mm G1P, and varying amounts of ATP and 3PGA. After 10 min at 30°C, the reaction was stopped by boiling for 5 min. After a 5-min centrifugation, the supernatant was stored at 4°C or –80°C until the ADP-Glc content was determined by HPLC as described (Tiessen et al., 2002).

Extraction and Assay of FBPase and NADP Malate Dehydrogenase

FBPase and NADP-malate dehydrogenase activities in chloroplasts were measured by mixing 20 μL of the chloroplasts solution from the oxygen electrode with 180 μL of reaction mixture containing 50 mm K-Tricine, pH 8.0, 5 mm MgCl₂, and 0.1% (v/v) Triton X-100. For FBPase, the mixture additionally contained 0.1 mm NADP⁺, 40 μm Fru-1,6-bisphosphate, and 1.75 mm EDTA. The reaction was stopped by the addition of 20 μL of 1 m NaOH either directly or after a 3- or 10-min incubation at room temperature. The reaction mix for NADP-malate dehydrogenase assays instead contained 0.1 mm NADPH, and 0 or 2 mm oxaloacetate additionally. It was stopped after 10 min by addition of 20 μL 1 m HCl/0.1 m Tricine, pH 9. In both cases, the difference in NADP(H) content of the two samples was taken as a measure for enzyme activity. The samples were stored at 4°C until further processing. Heating of the samples for 5 min at 95°C ensured the complete disrupture of all unused nucleotide-adenine substrate (NADP or NADPH). Five or 10 μL of the reaction was brought to pH 9 by the addition of 25 mm HCl/50 mm Tricine, pH 9, for FBPase or 0.1 m NaOH for MDH. The NADP(H) content was determined directly after the pH adjustment by an enzymatic cycling assay (Gibon et al., 2002). Both assays were shown to be linear with time for over 10 min.

Extraction and Assay of Suc, Reducing Sugars, Starch, Hexose Phosphates, and 3PGA

Suc, Glc, Fru, and starch were determined in ethanol extracts as described by Geigenberger et al. (1996), and hexose phosphates as by Gibon et al. (2002). For starch determination, the pellets of the ethanol extraction were solubilized by heating them to 95°C in 0.1 m NaOH for 30 min. After acidification to pH 4.9 with an HCl/sodium-acetate, pH 4.9, mixture, part of the suspension was digested overnight with amyloglucosidase and α-amylase. The Glc content of the supernatant was then used to assess the starch content of the sample. 3PGA was determined in perchloric acid extracts using an enzymatic cycling assay (Gibon et al., 2002).

Labeling Experiments and Label Separation

Labeling experiments were carried out with whole Arabidopsis leaves cut directly from the plant, with ends of petioles re-cut under water. Leaves were incubated in the dark for 13 h at 20°C (humidity of 60%) in medium containing 2 mm MES-KOH (pH 6.5) and 0.66 mm or 0.33 mm [U-¹⁴C]Glc (specific activity, 111 mBq mm^–1; Amersham-Buchler, Braunschweig, Germany) together with various concentrations of Suc (see legends to figures for details). Incubations were done in petri dishes (5-mL volume). Wet ends of petioles of incubated leaves were cut and discarded, and leaves were frozen immediately in liquid nitrogen. After ethanol extraction, the soluble fraction was further separated into neutral, anionic, and cationic components by ion-exchange chromatography as by Geigenberger et al. (1997), and the insoluble material left after ethanol extraction was resuspended and digested overnight as described above and counted for starch. Label in the hexose phosphate pool was analyzed as by Geigenberger et al. (1997), and total carbon in the hexose phosphate pool was determined in ethanol extracts as described above using non-radioactive replicates incubated in parallel.