SEARCH

SEARCH BY CITATION

Keywords:

  • starch metabolism;
  • maltose metabolism;
  • cytosolic heteroglycans;
  • cytosolic transglucosidase;
  • cytosolic phosphorylase;
  • Arabidopsis thaliana L

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The recently characterized cytosolic transglucosidase DPE2 (EC 2.4.1.25) is essential for the cytosolic metabolism of maltose, an intermediate on the pathway by which starch is converted to sucrose at night. In in vitro assays, the enzyme utilizes glycogen as a glucosyl acceptor but the in vivo acceptor molecules remained unknown. In this communication we present evidence that DPE2 acts on the recently identified cytosolic water-soluble heteroglycans (SHG) as does the cytosolic phosphorylase (EC 2.4.1.1) isoform. By using in vitro two-step 14C labeling assays we demonstrate that the two transferases can utilize the same acceptor sites of the SHG. Cytosolic heteroglycans from a DPE2-deficient Arabidopsis mutant were characterized. Compared with the wild type the glucose content of the heteroglycans was increased. Most of the additional glucosyl residues were found in the outer chains of SHG that are released by an endo-α-arabinanase (EC 3.2.1.99). Additional starch-related mutants were characterized for further analysis of the increased glucosyl content. Based on these data, the cytosolic metabolism of starch-derived carbohydrates is discussed.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

During photosynthesis plant cells utilize up to 50% of the fixed carbon to form transitory starch. Photosynthesis-competent cells of higher plants (and many algae as well) synthesize starch particles exclusively in the stromal space of the chloroplasts (for review see Ball and Morell, 2003). When conditions are unfavorable to photosynthesis plastidial starch is rapidly degraded and the degradation products are exported into the cytosol. The biochemistry of this process has not yet been fully elucidated. Surprisingly, in Arabidopsis thaliana L. the activity of α-amylase (EC 3.2.1.1) is not essential for this process (Yu et al., 2005) but at least one plastidial isoform of β-amylase (EC 3.2.1.2) is indispensable (Scheidig et al., 2002; Smith et al., 2005). In any case, the plastidial path of starch degradation finally leads to the formation of neutral sugars which enter the cytosol via transporters of the inner chloroplast envelope membrane (Niittyläet al., 2004; Weber, 2004; Weise et al., 2004). As leaves of a mutant lacking the functional maltose transporter (designated as mex1) possess both an extremely high maltose content and a starch-excess phenotype the main path of the starch degradation appears to include the intraplastidial formation and the subsequent export of maltose (Niittyläet al., 2004). Once transported into the cytosol, the starch-derived maltose is thought to be subjected to a complex metabolism in which the action of a transglucosidase, initially designated as disproportionating (iso)enzyme 2 (DPE2; EC 2.4.1.25), is indispensable. The phenotype of leaves of Arabidopsis mutants deficient in DPE2 resembles that of mex1 mutants. The maltose content exceeds that of the wild type by approximately two orders of magnitude. The DPE2-mediated transfer reaction is thus likely to be an essential step within the cytosolic maltose metabolism that cannot be replaced or bypassed by other reactions. Starch degradation during the dark period is impeded (Chia et al., 2004; Lu and Sharkey, 2004). Thus, the carbon fluxes within the chloroplasts appear to be controlled by feedback from the cytosolic compartment. The mechanism of this control remains to be elucidated. As a further complication, both the α- and β-anomers of maltose occur in leaves, and their ratio varies during the light–dark regime (Weise et al., 2005).

Under in vitro conditions DPE2 transfers glucosyl residues to glycogen, using maltose as the glucosyl donor and releasing glucose (Chia et al., 2004). There are no reports of glycogen-like glucans in the cytosolic compartment of plant cells, so DPE2 presumably utilizes this highly branched homoglucan as a non-physiological substitute for some endogenous carbohydrates that reside in the cytosol but have not yet been identified.

Similar to DPE2, the cytosol-specific isoform of the phosphorylase (designated as PHS2 in Arabidopsis; EC 2.4.1.1) exhibits a high affinity towards glycogen, which in fact exceeds that of the mammalian muscle phosphorylase (Steup and Schächtele, 1981). Based upon these kinetic properties, a pool of highly branched polyglycans has been postulated to exist in the cytosol of plant cells (Yang and Steup, 1990).

Water-soluble heteroglycans have recently been isolated from various plant tissues that contain, as their most prominent constituents, arabinose, galactose and glucose. Their pattern of glycosidic linkages is highly complex, comprising more than 20 different linkages. Both low- and high-molecular-weight heteroglycans have been shown to reside in the cytosol of mesophyll cells. As revealed by in vitro assays performed with glycans isolated from leaves of A. thaliana L. or Solanum tuberosum L., they act as glucosyl acceptors for the phosphorylase-catalyzed transfer reaction (Fettke et al., 2005a,b). The interaction between the cytosolic phosphorylase isoform and the cytosolic heteroglycans is highly selective. The enzyme does not noticeably react with water-soluble apoplastic glycans and the rabbit muscle phosphorylase, which is otherwise kinetically closely related to the cytosolic plant phosphorylase, does not utilize any of the plant-derived heteroglycans (Fettke et al., 2005a,b). Similarly, the plastidial phosphorylase isozymes do not react with the cytosolic heteroglycans (Fettke et al., 2005b).

In this paper we provide evidence that the cytosolic heteroglycans also act as substrates for DPE2. Several approaches were chosen. First, in vitro assays were performed using glycan preparations from leaves of wild-type Arabidopsis plants and either DPE2-containing protein fractions or recombinant DPE2. Second, we analyzed the action of DPE2 and the cytosolic phosphorylase isoform and studied the glucosyl acceptor/donor sites used by both enzymes. Third, we investigated the substrate specificity of DPE2 under in vitro conditions. Fourth, we analyzed soluble heteroglycans prepared from the leaves of both DPE2-deficient and wild-type Arabidopsis plants. Finally, we characterized soluble heteroglycans isolated from different starch-related Arabidopsis mutants, such as starch-deficient plants and mutants lacking a functional plastidial transglucosidase, DPE1. For all these analyses, heteroglycans were isolated from leaves harvested either at the end of the light or of the dark period.

The data clearly show that both DPE2 and the cytosolic phosphorylase act on the cytosolic heteroglycans and can use the same acceptor sites. Furthermore, evidence is presented that the cytosolic heteroglycans possess a high degree of structural variability depending on the metabolic status of the leaf cells. This variability quantitatively affects the interactions between the two cytosolic enzymes and the heteroglycans. By comparing the cytosolic heteroglycans from the various mutants we provide evidence that these polysaccharides are involved in the cytosolic metabolism of both plastid-derived monosaccharides and disaccharides.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Plant-derived water-soluble glycans (designated as SHG0) are highly diverse, differing in physicochemical properties, subcellular locations and biochemical functions. In addition, SHG0 contains mono- and disaccharides (such as glucose, fructose and sucrose). Due to this heterogeneity, any quantification or detailed characterization of SHG0 is not meaningful. Recently, a method has been established that allows the extracted glycans to be resolved into distinct pools (Fettke et al., 2004, 2005a). This procedure includes three essential steps. First, all compounds having a size below 1 kDa are removed to yield SHGT that, as its main constituents, contains arabinose, galactose and glucose but is still functionally heterogeneous. Secondly, SHGT is separated into glycans having an apparent size of either below or above 10 kDa (designated as SHGS and SHGL, respectively). Finally, SHGL is separated into subfractions I and II, which can be achieved either by field flow fractionation (FFF; Wyatt, 1993) or by reaction with the β-glucosyl Yariv reagent (Nothnagel, 1997; Yariv et al., 1962). The heteroglycans from both subfractions I and II are homogeneous with respect to their subcellular distribution and biochemical function. Subfraction I resides in the cytosol (in a strict sense) of mesophyll cells, and in vitro it selectively acts as an acceptor for a glucosyl transfer catalyzed by the cytosolic phosphorylase isozyme (Fettke et al., 2004, 2005a). In contrast, subfraction II does not interact with this phosphorylase isoenzyme and it is located in the apoplastic space of the leaf tissue. Unlike subfractions I and II, SHGS includes carbohydrates that still are topologically and functionally heterogeneous. It comprises at least 20 distinct oligosaccharides as resolved by high-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD). Some of the oligosaccharides are recovered inside isolated microsomal vesicles and, presumably, undergo the process of synthesis/export via the endomembrane system to yield constituents of the extracellular matrix. However, other oligosaccharides in the SHGS fraction reside in the cytosol sensu stricto and, as shown for both A. thaliana L. and S. tuberosum L., some of these act as acceptors for the glucosyl transfer catalyzed by the cytosolic phosphorylase isozyme (Fettke et al., 2005a,b).

DPE2 interacts with SHGT under both catalytic and non-catalytic conditions

In order to test for a possible interaction between DPE2 and plant-derived heteroglycans a solid-phase assay was established that utilizes immobilized SHGT and a complex mixture of buffer-soluble proteins extracted from the leaves of Arabidopsis wild-type plants. The SHGT was immobilized by coupling to epoxy-activated Sepharose. As a control, SHGT-free but otherwise identically treated Sepharose particles were used. The binding assay was performed in a buffer containing 0.1 m NaCl in order to minimize non-selective interactions. Non-interacting proteins were removed by washing with the same buffer. Subsequently, bound proteins were eluted using a high-salt buffer (1 m NaCl) followed by a final elution step performed under denaturing conditions (6 m urea). The various eluate fractions were subjected to SDS-PAGE, transferred to nitrocellulose and the blots were analyzed using antibodies directed against DPE2 (Figure 1a) or the cytosolic phosphorylase isozyme (Figure 1b). The two Western blots clearly indicate that both enzymes interacted with the immobilized glycans and were quantitatively eluted under high-salt conditions. As they did not bind to the SHGT-free Sepharose gel any non-selective protein–carbohydrate or protein–matrix interaction can be ruled out.

image

Figure 1.  DPE2 interacts with plant-derived heteroglycans. (a, b) Western blot analysis of a solid-phase assay using SHGT covalently bound to Sepharose beads (P). As a control, a glycan-free Sepharose was used (C). Following application of buffer-soluble proteins extracted from leaves of wild-type plants of A. thaliana L. the Sepharose gels were eluted with buffer containing 0.1 m NaCl, 1 m NaCl, and with 6 m urea. The eluate fractions were collected separately and processed for SDS-PAGE and Western blotting. Blots were analyzed using antibodies directed against DPE2 (a) or the cytosolic phosphorylase isozyme (b). The open and closed arrows mark the position of DPE2 and of AtPHS2, respectively. (c) Affinity electrophoresis and affinophoresis performed with proteins extracted from Arabidopsis leaves followed by staining for DPE2 activity. The separation gel contained 0.1% (w/v) glycogen (lane a) or no immobilized carbohydrate (lanes b to e). Twenty-five micrograms of protein was loaded on lanes a and b. Following electrophoresis, lanes a and b were incubated overnight (37°C) in a mixture containing 0.1% (w/v) soluble starch, 25 mm maltose and 100 mm citrate-NaOH pH 6.5. Prior to loading on lanes c to e, proteins were pre-incubated for 15 min at 4°C in a final volume of 30 μl. The following pre-incubation mixtures were used: 20 μg protein in 0.4% (w/v) glycogen (lane c); 20 μg protein and 0.4% (w/v) SHGT (lane d); 20 μg protein without any exogenous carbohydrate (lane e). The entire volume of the respective pre-incubation mixture was applied to each lane. Following electrophoresis, gels were stained overnight (37°C) in a mixture containing 25 mm maltose and 100 mm citrate-NaOH pH 6.5.

Download figure to PowerPoint

The data obtained with the solid-phase assay were confirmed by electrophoresis studies (Figure 1c). In these experiments buffer-soluble proteins from Arabidopsis leaves were subjected to a non-denaturing discontinuous PAGE that separates under moderately alkaline conditions (Steup, 1990). These conditions are often beneficial if both the native status of a carbohydrate-active enzyme and selective protein–carbohydrate interactions are to be maintained. Following native PAGE, the slab gels were stained for DPE2 activity. A single band of activity was observed both in a glycogen-containing and a glycogen-free separation gel, but the position of the enzyme activity differs strikingly (Figure 1c, lanes a and b). In the polysaccharide-free gel DPE2 moved towards the anode, whereas it was essentially immobile in the presence of glycogen. This indicates that during electrophoresis the enzyme strongly binds to the immobilized polyglucan as has been previously described by Chia et al. (2004).

In another series of experiments the buffer-soluble proteins were pre-incubated with either glycogen or SHGT and were then subjected to non-denaturing PAGE using a carbohydrate-free separation gel. As a control, pre-incubation was performed in the absence of any glycan. Proteins were allowed to interact with carbohydrates during pre-incubation and the resulting protein–glycan complexes were expected to be retained in the subsequent electrophoresis. In this type of native PAGE, which is often referred to as ‘affinophoresis’ (Shimura, 1990; Shimura and Kasai, 1998), the ligand is mobile and the electrophoretic mobility of the protein–ligand complexes can be predicted only if both the structure and the charge of the complex are known. Following electrophoresis, the separation gels were stained for DPE2 activity using a reaction mixture that contained maltose but lacked any high-molecular-weight glucosyl acceptor. Pre-incubation with either glycogen (Figure 1c, lane c) or SHGT (Figure 1c, lane d) resulted in a staining of the DPE2 activity, but in the control sample no staining was detectable (Figure 1c, lane e). It should be noted that in these experiments the actual ligands of DPE2 are unknown as hydrolytic activities may partially degrade glycogen and/or SHGT during pre-incubation thereby forming the actual ligands of DPE2. Nevertheless, these data clearly indicate that DPE2 interacts with carbohydrates derived from both glycogen and SHGT. The carbohydrate–DPE2 complexes formed are retained during electrophoresis. Subsequently, DPE2 utilizes the non-covalently bound glycans as acceptors for repetitive maltose-dependent glucosyl transfer reactions that form iodine-stainable α-helical structures.

Similar affinophoresis results have been previously obtained for the cytosolic phosphorylase isoenzyme from Pisum sativum L. interacting with SHGT (Fettke et al., 2004).

Both the transglucosidase and the extra-plastidial phosphorylase isozyme use the cytosolic heteroglycans as a substrate and share common acceptor/donor sites

The data shown in Figure 1 indicate that under in vitro conditions both DPE2 and PHS2 from Arabidopsis interact with heteroglycans. Therefore we asked whether or not both enzymes act on the same glycan molecules and, if so, on identical acceptor sites. To answer this question, two 14C-labeling experiments were performed, each of which consists of several steps. In the first labeling experiment (Figure 2a), SHGL was subjected to a glucosyl transfer catalyzed by a recombinant cytosolic phosphorylase isozyme (Fettke et al., 2004). The 14C-labeled glucose 1-phosphate served as the glucosyl donor. This reaction was performed under conditions (such as a relatively short incubation period and relatively high glycan concentrations) under which the formation of extended helical structures was prevented. Subsequently, the residual glucose phosphate was removed. In the next step, the labeled SHGL was incubated for 2 h at 37°C with a DPE2-containing protein fraction prepared from Arabidopsis wild-type leaves and glucose. As a control, the incubation was performed in the absence of glucose under otherwise identical conditions. The second incubation was terminated by separation of SHGL from carbohydrates having a size of less than 10 kDa and the latter were analyzed by HPAEC. The eluate fractions were collected separately and their 14C contents were quantified. In addition, the radioactivity retained in SHGL was monitored. In the presence of glucose approximately 40% of the 14C content of SHGL was lost during the second incubation. The loss of 14C was reduced to approximately 10% when glucose was omitted.

image

Figure 2. 14C-labeling of heteroglycans (SHGL) by the cytosolic phosphorylase isozyme or by DPE2. (a) DPE2 utilizes the same glycan molecules and the same glucosyl residues previously transferred to SHGL by the cytosolic phosphorylase. The SHGL (25 μg) was labeled with 14C-glucosyl residues by the action of recombinant phosphorylase (0.2 nkat) for 1 h at 37°C. Following termination of the reaction, the labeled SHGL was divided into two equal parts each of which was incubated for 2 h at 37°C with a DPE2-enriched protein fraction (30 μg each) from leaves of wild-type A. thaliana L. either in the absence or in the presence of glucose (final concentration 20 mm). Subsequently, the two reaction mixtures were passed through a membrane filter (MWCO 10 kDa) and the two filtrates were analyzed by HPAEC-PAD (incubation in the presence of glucose, dark line; incubation in the absence of glucose, light line). Samples eluting between 10 and 14 min were collected as three distinct fractions and their 14C contents were monitored. In the insert the 14C content of the three eluate fractions are given (dark columns, complete reaction mixture; light columns, glucose-free incubation mixture). (b) The cytosolic phosphorylase utilizes glucosyl residues previously transferred to SHGL by the action of DPE2. The SHGL (25 μg) was incubated with 14C maltose (final concentration 20 mm) and the recombinant DPE2 (2 μg) for 30 min at 37°C. Following termination and removal of the residual maltose and the glucose formed, the labeled SHGL was divided into two equal parts. Each part was incubated with recombinant phosphorylase (0.2 nkat) for 2 h at 37°C either in the absence or in the presents of orthophosphate (final concentration 200 mm). Subsequently, the two samples were passed through a membrane filter [as in (a)] and the two filtrates were analyzed by HPEAC-PAD. Samples eluting between 22 and 30 min were collected and their 14C contents were monitored. (c) DPE2 reacts with cytosolic heteroglycans (subfraction I) but not with the apoplastic glycans (subfraction II). Recombinant DPE2 (5 μg each) was incubated with subfractions I or II (7μg each) and 20 mm14C-labeled maltose at 37°C. As a control, the reaction mixture did not contain any polysaccharide. In another series, the plant-derived heteroglycans were replaced by glycogen (0.2 μg each) using otherwise identical conditions. After 20, 30 and 40 min the reaction mixtures were subjected to filtration using a 10 kDa membrane filter and the retentates were repeatedly washed with water. Subsequently, radioactivity in the >10 kDa fraction was monitored. The bars are the mean ± SD of three independently performed experiments.

Download figure to PowerPoint

The HPAEC-PAD chromatograms of the glycans of less than 10 kDa and the 14C contents of the eluate fractions are shown in Figure 2(a). A labeled compound that exactly co-chromatographed with authentic maltose was observed when the SHGL was incubated in the presence of both DPE2 and glucose (Figure 2a, insert). This compound was undetectable when glucose was omitted in the incubation. Thus, a major proportion of labeled glucosyl residues that had been incorporated into the high-molecular-weight glycans by the phosphorylase isozyme were transferred to glucose to yield maltose. As the maltose formation strictly depends on glucose it is due to the action of DPE2 rather than to any hydrolytic enzyme activity.

In the second 14C-labeling experiment, the order of the two glucosyl transfer reactions was altered (Figure 2b). SHGL was first labeled using 14C-maltose and recombinant DPE2. Following the removal of the residual maltose, the labeled SHGL was incubated with the recombinant cytosolic phosphorylase isozyme and orthophosphate. As a control, orthophosphate was omitted. Subsequently, the formation of 14C-glucose 1-phosphate was monitored. Under these conditions labeled glucose 1-phosphate was found but its formation was strictly dependent on the addition of orthophosphate to the second incubation mixture (Figure 2b). Taken together, these data clearly show that both DPE2 and the cytosolic phosphorylase are capable of acting on the same glycan molecules and of using, at least in part, the same glucosyl acceptor/donor sites.

SHGL contains both the cytosolic subfraction I and apoplastic subfraction II. As previously shown, the cytosolic phosphorylase isozyme acts on the entire subfraction I but is unable to use subfraction II as a glucosyl acceptor (Fettke et al., 2005a); therefore subfraction II is irrelevant in the labeling experiment shown in Figure 2(a). In order to determine the glycan specificity of DPE2 we incubated the recombinant protein with 14C-labeled maltose and equal amounts of either subfraction I or subfraction II (or glycogen as well). At intervals, the 14C labeling of the various glycans was monitored. The DPE2-dependent incorporation into subfraction I increased with time, but labeling of subfraction II was not discernible from that observed in the (glycan-free) control (Figure 2c). Compared with subfraction I, glycogen was more effective, even though, based upon the total monomer content, a lower concentration of glycogen than of SHG subfraction I was used. Presumably, this reflects the larger number of acceptor sites present in the highly branched homoglucan. A higher priming efficiency of glycogen has also been observed for the phosphorylase-dependent glucosyl transfer (Fettke et al., 2005a).

Selectivity of the DPE2-catalyzed glucosyl transfer

Glycosidic linkage analyses performed with subfraction I prepared from several plant species strongly suggest that the cytosolic heteroglycans possess a highly complex structure. Similarity to starch or other α-glucans appears to be very low as the typical starch-like linkages were generally below the limit of detection (Fettke et al., 2004, 2005a,b). Therefore, the question arises of whether or not the DPE2-catalyzed reaction is strictly dependent on glucose-like acceptors. To answer this question DPE2 was characterized using two approaches.

First, the recombinant enzyme was incubated with various hexoses/hexuloses or pentoses as potential glucosyl acceptors. Glycogen was added as a highly effective (but non-physiological) glucosyl donor. Following incubation, the low-molecular-weight compounds were separated from the other constituents of the incubation mixtures and were analyzed by HPAEC-PAD. The occurrence of additional peaks eluting in the disaccharide region of the chromatogram indicates that under in vitro conditions the respective monosaccharide can act as a glucosyl acceptor (Figure 3a). In addition, the monomer composition of some of the disaccharides observed was determined by acid hydrolysis and rechromatography of the hydrolysates (data not shown). By using this approach, we confirmed that DPE2 utilizes glucose as an acceptor yielding maltose, but both mannose and xylose were similarly efficient acceptors. Both fucose and galactose also acted as glucosyl acceptors, although less efficiently than glucose, mannose and xylose. No disaccharide formation was detectable when arabinose or fructose were applied (data not shown). Thus, DPE2 does not exert a high selectivity with respect to the monosaccharides that act as glucosyl acceptors and, consequently, is capable of forming a variety of disaccharides in addition to maltose. Currently, the metabolic implications of these results are difficult to assess. A fixed monosaccharide concentration (30 mm each) was applied in all the experiments shown in Figure 3(a). More detailed kinetic studies are very difficult to perform, for several reasons. First, the mechanism of the DPE2-catalyzed reaction has not yet been elucidated nor has the thermodynamic equilibrium of the glucosyl transfer reaction(s) been determined. Furthermore, in all these experiments glycogen was used as a glucosyl donor that acts as a non-physiological substitute for cytosolic heteroglycans and, therefore, kinetic constants derived from these data have a limited relevance for the in vivo action of DPE2. Due to the relatively wide size distribution, kinetics with subfraction I are expected to be highly complex.

image

Figure 3.  Selectivity of the DPE2-catalyzed glucosyl transfer. (a) Glucosyl transfer from glycogen to various monosaccharides. Recombinant DPE2 (2.5 μg each) was incubated for 90 min at 37°C with glycogen and various monosaccharides (30 mm each). Following incubation, the reaction mixtures were passed through a 10 kDa filter and the resulting filtrates were analyzed by HPAEC-PAD. All monosaccharides used were tested by HPAEC for purity (data not shown). (b) DPE2-catalyzed glucosyl transfer using maltotriose (DP 3) or maltotetraose (DP 4). In a final volume of 30 μl, recombinant DPE2 was incubated for 90 min at 37°C with DP4 or DP3 (25 μg each) without any mono- or disaccharide or with glucose (Glc) and maltose, respectively. Following incubation, the reaction mixtures were analyzed by HPAEC-PAD. (c) DPE2-catalyzed glucosyl transfer from maltose to a mixture of maltohexaose and maltoheptaose. In a final volume of 40 μl recombinant DPE2 was incubated for 60 min at 37°C with DP 6/DP 7 (25 μg). Subsequently, the low-molecular-weight compounds were analyzed by HPAEC-PAD. As two controls, either the maltose or the DPE2 was omitted in the otherwise identical incubation mixtures. Following incubation the controls were treated identically and analyzed by HPAEC-PAD.

Download figure to PowerPoint

In a second approach, we analyzed the action of DPE2 on low-molecular-weight glucans. Glycogen was omitted in all reaction mixtures and DPE2 was incubated with defined maltodextrins, such as maltotriose, maltotetraose or a mixture of maltohexaose and maltoheptaose. Following incubation, the oligosaccharides were analyzed by HPAEC-PAD. In the presence of maltose the formation of larger maltodextrins was observed (Figure 3b,c). Thus, under the conditions used, all these maltodextrins act as a glucosyl acceptor that is elongated by a repetitive glucosyl transfer, with maltose being the donor. The patterns of the elongated maltodextrins do not indicate any preference of DPE2 for longer maltodextrins as the glucosyl acceptor. The data shown in Figure 3(b) and (c) are not consistent with the recent characterization of the DPE2 activity, indicating that DPE2 does not act on maltodextrins (Chia et al., 2004). The latter data were, however, obtained with a crude extract containing DPE2 rather than with a purified recombinant enzyme and, presumably, this type of reaction remained undetectable due to the low DPE2 activity applied and/or interfering activities of other enzymes present in the crude extract.

When DPE2 was incubated with maltotriose but maltose was replaced by glucose there was no detectable formation of maltose (Figure 3b). Thus, DPE2 did not catalyze the disproportionating reaction G3 + G1 ⇄ 2G2. Following the incubation of DPE2 with maltotetraose and glucose, low amounts of both maltose and maltotriose were observed (Figure 3b). Thus, under the conditions used, DPE2 appears to be able to convert maltotetraose to maltotriose but with a very low efficiency. In the absence of high-molecular-weight glycans maltotetraose seems to represent the minimum size of the glucosyl donor.

Physicochemical characterization of heteroglycans in a DPE2-deficient Arabidopsis mutant

A T-DNA insertion mutant of A. thaliana L. (designated as dpe2-3; see Chia et al., 2004) was used for a phenotypical characterization of the leaf heteroglycans. In this mutant, DPE2 is undetectable at the protein and enzyme activity level. Remarkably, in this mutant expression of the cytosolic phosphorylase isozyme (AtPHS2) is increased three- to fourfold (Chia et al., 2004). The heteroglycan isolation procedure that has recently been elaborated results in a highly reproducible monomer composition. On a fresh weight basis the SHG yield is within a wide range of different amounts of starting leaf material, unchanged (Fettke et al., 2005a). It is, therefore, highly likely that the target carbohydrates are quantitatively extracted and retained during the entire isolation procedure. This is an important prerequisite if the glycan analyses are to form part of the phenotypic characterization of a mutant. The previous study performed with Arabidopsis wild-type plants revealed that the cytosolic heteroglycans vary within the light–dark cycle and, presumably, depending on the metabolic status of the leaves (Fettke et al., 2005a). Therefore, the status of the starting material has to be defined.

In the present study, leaves from the dpe2-3 mutant and from wild-type plants were harvested at the end of the light or the dark period. The total amounts of SHGT isolated from leaves are given in Table 1. Under both conditions, the DPE2-deficient mutant possesses a slightly increased SHGT content on a fresh weight basis.

Table 1.   Absolute SHGT and relative SHGL content of leaves from a DPE2-deficient mutant (dpe2) and from wild-type (wt) plants of A. thaliana L.
Leaf materialaSHGT (μg g−1 fresh weight)b SHGL (%)c
  1. aLeaves were harvested either after 14 h illumination (light) or after 8 h darkness (dark).

  2. bData for the absolute SHGT content are means ± SD (n = 4). The absolute SHGT content is based on the fresh weight of the leaf material.

  3. cThe relative SHGL content is given as % of SHGT.

dpe2 light170.3 ± 5.480.1 ± 2.14
dpe2 dark152.6 ± 11.276.2 ± 1.43
wt light159.3 ± 7.277.5 ± 1.84
wt dark140.5 ± 6.778.4 ± 3.12

For a more detailed analysis, both the monomer composition and the size distribution of the heteroglycans were determined. Leaves from wild-type and DPE2-deficient mutant plants were harvested during the light or dark period and the water-soluble heteroglycans were isolated. The monomer patterns of the four SHGT preparations, as obtained by HPAEC-PAD analysis of the hydrolysates, are shown in Figure 4(a). As estimated from the amperometric signals, the SHGT isolated from the DPE2-deficient mutant has an approximately sixfold higher glucose content than the SHGT preparation from wild-type leaves.

image

Figure 4.  Monosaccharide patterns of heteroglycans from leaves of a DPE2-deficient mutant and a wild-type plant from A. thaliana L. SHGT (a) was isolated from leaves of a DPE2-deficient mutant (dpe2) or wild-type (wt) plants from A. thaliana L. Leaves were harvested at the end of either the light or the dark period. Following acid hydrolysis, 1 μg glucose equivalents of each sample was applied to HPAEC-PAD. Aliquots of the unhydrolyzed SHGT preparations were separated into SHGS (b) and SHGL (c). Following hydrolysis, 3 μg glucose equivalents each (SHGS) or 1 μg each (SHGL) was analyzed by HPAEC-PAD. All chromatograms were normalized to arabinose. Fuc, fucose; Rha, rhamnose; Ara, arabinose; Gal, galactose; Glc, glucose; Xyl, xylose; Man, mannose.

Download figure to PowerPoint

For a further characterization, SHGT was separated into heteroglycans having a size of less than 10 kDa (SHGS) and above 10 kDa (SHGL). The monomer patterns of the two glycan fractions were determined separately (Figure 4b,c). Both SHGS and SHGL prepared from the mutant possessed higher glucose levels than the respective glycans derived from the wild type. Most of the additional glucose in the DPE2-deficient mutant is to be found in SHGL (see also Table 1). However, SHGL is an inhomogeneous polysaccharide preparation as it contains high-molecular-weight glycans originating both from the cytosol (designated as subfraction I) and the apoplast (subfraction II). Each of the four SHGL preparations was resolved into these two subfractions using FFF (Figure 5). Two observations can be made. First, the ratio between subfraction I and subfraction II is not significantly altered in the DPE2-deficient mutant. Secondly, in the DPE2-deficient mutant the total pool of subfraction I is essentially unchanged throughout the light–dark regime. In contrast, in SHGL isolated from wild-type leaves the total amount of subfraction I is increased during darkness and decreased in the light. This difference is consistently observed in Arabidopsis wild-type leaves (Fettke et al., 2005a) and, presumably, reflects the varying metabolic roles of the cytosolic heteroglycans during the light and dark period.

image

Figure 5.  Size distribution and quantification of the SHGT-derived subfractions I and II as revealed by field flow fractionation followed by delta refractive index measurement (FFF-DRI). The SHGL was isolated from leaves of a DPE2-deficient mutant and of wild-type plants. Leaves were harvested during the middle of the light or dark period (as indicated). From each SHGT preparation 75 μg were separated by FFF and subfractions were quantified by DRI. Wild type, full lines; DPE2-deficient mutant, broken lines.

Download figure to PowerPoint

Following separation by FFF, subfractions I and II from the four SHGL preparations were separately subjected to acid hydrolysis and their monosaccharide patterns were determined (Figure 6a). The patterns of all the subfraction II preparations are dominated by galactose and arabinose. They are essentially indiscernible from leaves of wild type and the mutant plants, irrespective of the time of harvest. This indicates that the apoplastic heteroglycans do not respond to altered cytosolic levels of the DPE2 protein or activity and are not noticeably affected by the light–dark regime. As opposed to subfraction II, the four monosaccharide patterns for the (cytosolic) subfraction I differed: heteroglycan preparations obtained from DPE2-deficient plants contained more glucose than those from the control leaves. In addition, the content of xylose and mannose was increased. Furthermore, the glucose content was higher when the leaves were harvested during the light period.

image

Figure 6.  Monosaccharide patterns of the SHGL-derived subfractions I and II and the oligosaccharide patterns of SHGS from leaves of the DPE2-deficient mutant (dpe2) and of wild-type (wt) plants from A. thaliana L. harvested after 14 h illumination or 8 h darkness. (a) The SHGL-derived subfractions (sub) I and II were separated by FFF. Following hydrolysis, approximately 1 μg of each sample was applied to the HPAEC-PAD. All chromatograms were normalized to arabinose. (b) SHGS (25 μg each) were analyzed by HPAEC-PAD using the oligosaccharide mode. Abbreviations are as in Figure 4.

Download figure to PowerPoint

Unlike subfractions I and II, the native oligosaccharides (SHGS) are accessible to an anion-exchange chromatographic analysis (Figure 6b). For all four SHGS preparations complex oligosaccharide patterns were obtained consisting of more than 20 compounds. The patterns for wild-type leaves harvested during the light and dark period differed, but were more similar than those from the DPE2-deficient mutant. As shown previously, some of the SHGS-derived oligosaccharides reside in the cytosol (in a strict sense) but others are located inside the cellular endomembrane system (Fettke et al., 2005a). Nevertheless, the SHGS preparation from the DPE2-deficient mutant clearly deviates from that of the wild type.

Enzymatic characterization of heteroglycans in a DPE2-deficient mutant

For a further characterization of the soluble heteroglycans from the DPE2-deficient mutant we analyzed in vitro protein–carbohydrate interactions in more detail. First, we estimated the priming efficiency for the cytosolic phosphorylase isozyme. SHGL was prepared from illuminated or darkened leaves of mutant and wild-type plants. Samples (10 μg each) of SHGL were incubated with recombinant cytosolic phosphorylase and 20 mm [U-14C]glucose 1-phosphate. At intervals, reaction was terminated and the incorporation of 14C-labeled glucosyl residues into SHGL was determined (Table 2). The glucosyl transfer into the heteroglycans differed depending on the starting material. In wild type-derived heteroglycans incorporation was higher when SHGL had been prepared from illuminated leaves compared with the dark control. In the latter, the amount of the phosphorylase-reactive subfraction I is slightly increased (Figure 5), and therefore the data presented in Table 2 slightly underestimate the differences in priming efficiency. Furthermore, the kinetics of the incorporation vary between the two glycans. The rate of incorporation into SHGL derived from darkened leaves is essentially constant, whereas labeling of SHGL derived from illuminated leaves follows a more complex kinetic. Both effects were also observed with SHGL prepared from the DPE2-deficient mutant.

Table 2.   Glucosyl transfer from glucose 1-phosphate to SHGL as catalyzed by the cytosolic phosphorylase isozyme. Recombinant phosphorylase (4 nkat) was incubated at 37°C in a mixture containing 20 mm [14C]-glucose 1-phosphate and SHGL (10 μg each) isolated from leaves of the DPE2-deficient mutant (dpe2) or wild-type (wt) plants that had been harvested after 14 h illumination or 8 h darkness. At intervals (as indicated) the reaction was terminated and the 14C content was determined. One representative experiment (of three) is shown
Incubation time (min)Glucosyl moieties incorporated into SHGL (nmol)
wt lightwt darkdpe2 lightdpe2 dark
 533.210.436.015.2
1539.926.942.337.7
2571.753.181.976.9

Because the phosphorylase-related priming capacity of SHGL is restricted to subfraction I, these data clearly indicate that the large cytosolic heteroglycans possess a functional variability. Thus, the efficiency of use and/or the number of glucosyl acceptor sites that are used by the cytosolic phosphorylase appear to vary during the light–dark regime.

The functional variability of subfraction I was further analyzed by in vitro assays using recombinant DPE2 and 14C-labeled maltose. In these experiments a constant reaction time (40 min) and two fixed maltose concentrations (5 or 20 mm), but varying amounts of SHGL, were applied. The SHGL was isolated from either illuminated or darkened wild-type leaves and was incubated with the recombinant DPE2 (Table 3). Over the entire range of SHGL levels, heteroglycans from the DPE2-deficient mutant possessed a much higher acceptor efficiency than those from wild-type plants. Glycans from illuminated leaves of the mutants were more effective than those of darkened leaves.

Table 3.   Glucosyl transfer from maltose to SHGL as catalyzed by DPE2. Recombinant DPE2 (2.5 μg) was incubated at 37°C in a mixture containing 5 or 20 mm [14C]-maltose and SHGL isolated from leaves of the DPE2-deficient mutant (dpe2) and wild-type (wt) plants harvested at the end of the light or dark period, respectively. After 40 min the reaction was terminated and the 14C-content (nmol glucosyl moieties) was determined. The average of two experiments is shown
Glucosyl donorLeaf materialSHGL
10 μg20 μg30 μg
5 mm maltosewt light9.9210.2316.86
dpe2 light160.32189.83201.02
wt dark5.375.988.65
dpe2 dark117.25160.07178.84
20 mm maltosewt light9.8511.9116.46
dpe2 light171.93196.68187.95
wt dark6.096.518.67
dpe2 dark133.93168.38172.92

It is likely that DPE2 catalyzes a repetitive glucosyl transfer to multiple acceptor sites of the heteroglycan molecules, which presumably results in both structural and kinetic diversification of the heteroglycans. As a result of these complications no detailed kinetic analysis was performed.

Presumably, the differences in the glucose content in subfraction I (see Figure 6a) are related to the varying acceptor efficiencies as documented in Table 3. This assumption is supported by an in vitro approach in which subfraction I prepared from illuminated leaves of the DPE2-deficient mutant was incubated with various enzymes (such as phosphorylase or α-amylase) that act on α-1,4-interglucose linkages. Following incubation, changes in the monosaccharide patterns of subfraction I were monitored by ion chromatography (Figure 7). Most of the glucosyl residues that are present in the subfraction I derived from the DPE2-deficient mutant were removed by both α-amylase and phosphorylase. From these data, two conclusions can be drawn. First, most (if not all) of the additional glucosyl residues that are present in the subfraction I from the DPE2-deficient mutant carry α-1,4-glucosidic linkages. Second, these glucosyl residues are accessible to both the hydrolytic and the phosphorolytic cleavage. As the action of phosphorylases (EC 2.4.1.1) is restricted to the non-reducing ends of α-glucan-like structures most (if not all) of the additional glucosyl residues are attributed to oligosaccharyl chains forming non-reducing ends.

image

Figure 7.  Enzymatic removal of glucosyl residues from SHGL prepared from leaves of a DPE2-deficient Arabidopsis mutant. In a final volume of 30 μl, SHGL (5 μg each) was incubated for 90 min at 37°C in a mixture containing either α-amylase or the recombinant cytosolic phosphorylase isozyme in the presence of orthophosphate (cyto. Pho + Pi). As a control, SHGL was incubated in the absence of any enzyme (control). Following incubation, low-molecular-weight compounds were removed and the remaining SHGL was subjected to acid hydrolysis and HPAEC-PAD analysis. As a further control (enzyme control), incubation mixtures were prepared that contained the α-amylase or the cytosolic phosphorylase isozyme (as above) but SHGL was omitted. These mixtures were processed identically and analyzed by HPAEC-PAD to ensure that no contaminating carbohydrates were present in the enzyme preparations used. Because of identity only one chromatogram is shown.

Download figure to PowerPoint

A recombinant fungal endo-α-1,5-arabinanase (EC 3.2.1.99) was applied as another degrading enzyme. This enzyme selectively cleaves α-1,5-linkages between arabinosyl residues. As previously shown, it acts on subfraction I of plant-derived SHGL preparations and liberates oligosaccharyl chains that carry glucosyl acceptor sites related to the cytosolic phosphorylase (Fettke et al., 2005b). The four SHGL preparations were incubated with arabinanase and the oligosaccharides released were analyzed by HPAEC-PAD (Figure 8a). The resolved oligosaccharide patterns differed strikingly. SHGL preparations from the DPE2-deficient mutant gave a much more complex pattern of oligosaccharides than was obtained from preparations from wild-type plants. The most prominent compound eluted slightly before maltose. Two other major compounds were recovered after approximately 18 and 30 min. Currently, none of these compounds has been identified.

image

Figure 8.  Analysis of heteroglycans from the DPE2-deficient mutant and wild-type plants by degradation using arabinanase. (a) SHGL was prepared from leaves of the DPE2-deficient Arabidopsis mutant (dpe2) and from wild-type (wt) plants. Leaves were harvested after 14 h illumination or after 8 h darkness. The four SHGL preparations (100 μg each) were incubated with arabinanase for 7 h at 40°C. The oligosaccharides released were separated by membrane filtration and analyzed by HPAEC-PAD. For comparison, a standard maltodextrin mixture is included and the degrees of polymerization (DP) are given. Three prominent oligosaccharides that are more abundant in the mutant-derived samples are marked with closed arrows. In the chromatograms of the mutant-derived samples areas that deviate strongly from those of the wild-type controls are underlined. (b, c) Following HPAEC-PAD, the oligosaccharides released by arabinanase were further analyzed. Glycans eluting between 0 and 10 min (b) or between 10 and 17 min (c) were collected and hydrolyzed separately. The monosaccharide patterns of both samples were analyzed by HPAEC-PAD. (d) Following arabinanase treatment and removal of the oligosaccharides the remaining SHGL heteroglycans were subjected to acid hydrolysis and analyzed by HPAEC-PAD. Abbreviations are as in Figure 4.

Download figure to PowerPoint

Two pools of the oligosaccharides resolved by HPAEC were collected and hydrolyzed. The monosaccharide patterns of the earlier (0 to 10 min) and later (10 to 17 min) eluting pool are shown in Figure 8(b) and (c), respectively. In all mutant-derived samples the glucose content was much higher than in the wild-type controls. Furthermore, the later eluting mutant-derived oligosaccharide pool contained, in addition to glucose, more non-glucose compounds, such as arabinose, xylose or mannose. This strongly suggests that the higher glucosyl content in the oligosaccharides released by the action of arabinanase results in a shift towards higher degrees of polymerization (DP) and, therefore, towards the later eluting region of the chromatogram.

Following arabinanase treatment, the residual high-molecular-weight glycans that represent the vast majority of the entire SHGL preparations were hydrolyzed. The monosaccharide patterns are shown in Figure 8(d). It is remarkable that the glucosyl content of preparations from the DPE2-deficient mutant is higher than that from wild-type plants. Thus, it appears that some of the glucosyl acceptor sites used in vivo are not susceptible to the arabinanase treatment. A similar conclusion has been recently reached for heteroglycans from S. tuberosum L. (Fettke et al., 2005b).

In summary, these data clearly point to a phenotypical difference in the cytosolic heteroglycans (subfraction I). The DPE2 deficiency results in the formation of heteroglycans whose outer oligosaccharide chains are larger and enriched in glucose residues. In contrast, the apoplastic heteroglycans (subfraction II) are unaffected by the DPE2 deficiency.

Analysis of additional Arabidopsis mutants with altered starch metabolism

The increased glucosyl content in the cytosolic heteroglycans (subfraction I) isolated from the DPE2-deficient mutant is unexpected. For a better understanding of this phenotype, we analyzed the SHGL-derived subfractions I and II from starch-deficient Arabidopsis mutants that lack either the plastidial ADPglucose pyrophosphorylase (adg1; EC 2.2.7.27; Lin et al., 1988) or the plastidial phosphoglucomutase (pgm1; EC 5.4.2.2; Caspar et al., 1985). In the mutants the maltose levels are close to or below the limit of detection (Niittyläet al., 2004). As shown in Figure 9(a) and (b), these mutants also possess an increased glucosyl content, which is restricted to the cytosolic heteroglycans (subfraction I). This effect is pronounced in the light period, whereas only small differences between the mutants and the appropriate wild type are observed during darkness. Thus, the increased glucosyl content of subfraction I reflects an additional intracellular carbon flux rather than the export of starch-derived maltose into the cytosol and its subsequent metabolism.

image

Figure 9.  Analysis of Arabidopsis mutants with altered starch metabolism and wild-type plants. (a, b) SHGL was isolated from Arabidopsis mutants deficient in the plastidial phosphoglucomutase (P-PGM) or in the plastidial AGPase (AGPase). As a control, SHGL was isolated from wild-type plants (wt) grown under identical conditions to the two mutants. The SHGL-derived subfractions (sub) I and II were separated by FFF. Following acid hydrolysis, approximately 2 μg of each sample was applied to the HPAEC-PAD. All chromatograms were normalized to galactose. The SHGL was isolated from leaves harvested after 14 h illumination (a) or after 8 h darkness (b). (c) Enzymatic removal of glucosyl residues from SHGL isolated from illuminated leaves of the two starch-deficient mutants. In a final volume of 30 μl, SHGL (5 μg each) was incubated for 90 min at 37°C in a mixture containing α-amylase. As a control, SHGL was incubated in the absence of enzyme. Following incubation, low-molecular-weight compounds were removed and the remaining SHGL was subjected to acid hydrolysis and the hydrolysate was analyzed by HPAEC-PAD. (d) Affinity electrophoresis of buffer-soluble proteins from the two starch-deficient mutants and from wild-type leaves followed by phosphorylase activity staining. The separation gels contained 0.25% (w/v) glycogen. Twenty micrograms of protein was loaded to each lane. Following electrophoresis, the gels were incubated overnight (37°C) in a mixture containing 20 mm glucose 1-phosphate and 100 mm citrate-NaOH pH 6.5 and stained with iodine. I and II, cytosolic phosphorylase isoform; III, plastidial phosphorylase isoform. (e) Heteroglycan analysis of the DPE1-deficient mutants. The SHGL was isolated from DPE1-deficient mutants and from wild type (Wassilewskija, Ws). Leaves were harvested after 14 h illumination or after 8 h darkness. Following total hydrolysis, approximately 5 μg of each hydrolysate was applied to the HPAEC-PAD. Abbreviations are as in Figure 4.

Download figure to PowerPoint

The additional glucosyl residues of subfraction I can be released by α-amylase treatment, indicating that they are linked via α-1,4 glucosidic bonds (Figure 9c). In the DPE2 mutant of A. thaliana L. the cytosolic phosphorylase isozyme is increased three- to fourfold compared with the wild type (Chia et al., 2004). As revealed by affinity electrophoresis, the two starch-deficient mutants also possess an elevated level of the cytosolic phosphorylase isozyme (Figure 9d). In contrast, the level of the plastidial isoform is unchanged. As shown previously, the cytosolic phosphorylase isozyme from A. thaliana L. occurs in two states that differ in their apparent affinities towards the immobilized glycogen (Fettke et al., 2005a). Both states were increased in the starch-deficient mutants. Thus, the increased AtPHS2 levels are not specifically linked to an increased maltose content.

As opposed to the two starch-deficient mutants, the DPE1 mutant from Arabidopsis possesses starch levels exceeding those of wild-type plants. This mutant lacks the plastidial disproportionating enzyme DPE1 (Chritchley et al., 2001). This enzyme appears to be involved in plastidial maltodextrin metabolism and is required for the formation of starch-derived glucose that is exported into the cytosol via a glucose transporter of the chloroplast envelope (Smith et al., 2005). Interestingly, the cytosolic heteroglycans (subfraction I) from this mutant differ from all the other investigated heteroglycans as they possess a decreased glucosyl content (Figure 9e). These data indicate that the plastidial glucose export is correlated with the relative glucosyl content of subfraction I.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Cytosolic heteroglycans whose major constituents are arabinose, galactose and glucose have been isolated from leaves of P. sativum L., A. thaliana L. and S. tuberosum L. (Fettke et al., 2004, 2005a,b). They have also been found in heterotrophic organs, such as potato tubers (Fettke et al., 2005a) and appear to be common in higher plants. More than 20 different glycosidic linkages were observed for the cytosolic heteroglycans and, therefore, a large number of cytosolic carbohydrate-active enzymes is expected to be required for the turn-over of the cytosolic heteroglycans.

In this paper we provide evidence that the recently identified cytosolic transglucosidase, designated as DPE2 (Chia et al., 2004), can use these heteroglycans as a substrate for a bidirectional glucosyl transfer as does the cytosolic phosphorylase. Thus, currently two glucosyl residue-transferring enzymes are known that have a potential relevance for the in vivo biochemistry of the cytosolic heteroglycans. However, none of the enzymes that transfer non-glucosyl residues towards or from the heteroglycans have yet been identified.

As glucosyl residues transferred to the heteroglycans by the phosphorylase isozyme are converted to maltose in a subsequent DPE2-catalyzed reaction (and vice versa) both enzymes are able to use the same glucosyl acceptor/donor sites in vitro (Figure 2a,b). However, this two-step transfer is not quantitative and, therefore, some of the sites appear to be selective for either of the two enzymes.

DPE2 and cytosolic phosphorylase use different low-molecular-weight acceptors for the glucosyl transfer from the high-molecular-weight glycan. For the latter enzyme the acceptor is orthophosphate (or, with a far lower efficiency, arsenate), whereas DPE2 utilizes a monosaccharide and forms a disaccharide. The selectivity of DPE2 for the monosaccharide acceptor (and, consequently, for the formation of disaccharides) is surprisingly low. Glucose, mannose and xylose are similarly effective acceptors, and galactose and fucose are also used, albeit with a lower efficiency (Figure 3a). In contrast, DPE2 apparently does not use fructose and arabinose.

Presumably, two features of a monosaccharide are important for an efficient acceptor function: first, a pyranose ring having a configuration similar to that of glucose and, second, the orientation of the hydroxyl group at C4. In contrast, both the orientation of the hydroxyl group at C2 and the presence or absence of C6 appear to be of little relevance.

Based on the data shown in Figure 3(a) it cannot be excluded that DPE2 also transfers glucosyl residues from maltose (or other disaccharides) to acceptor sites of the high-molecular-weight glycan that lack a terminal glucosyl residue. Interestingly, both xylose and mannose appear to be abundant in the vicinity of some of the acceptor sites (Figure 8b,c). DPE2 also acts on maltodextrins and catalyzes maltose-dependent glucosyl transfer to maltotriose or maltotetraose and, therefore, gives rise to the formation of a series of maltodextrins (Figure 3b,c). Possibly, DPE2 acts on some oligosaccharides that are constituents of SHGS. Similar data have been obtained for the cytosolic phosphorylase isozyme from A. thaliana L. and from S. tuberosum L. (Fettke et al., 2005a,b).

Whilst all these in vitro data clearly demonstrate that DPE2 can act on soluble heteroglycans they do not provide any unambiguous information on the in vivo function of the enzyme. Therefore, heteroglycans from a DPE2-deficient mutant were characterized and compared with those from a wild-type control. The characterization included the analyses of physical and chemical parameters, such as the size distribution or the monomer patterns, and also biochemical features, such as the priming efficiencies for glucosyl transferases. In addition, we analyzed oligosaccharyl chains liberated by the action of an endo-α-1,5-arabinanase. When taken together, the data clearly demonstrate that the cytosolic heteroglycans from the DPE2-deficient mutant differ significantly from those of the wild type. In contrast, no difference was observed for the apoplastic heteroglycans, subfraction II. These results concur with the fact that in Arabidopsis leaves the biosynthetic pathways of subfractions I and II are largely separated (Fettke et al., 2005a).

With the cytosolic heteroglycans three differences were observed between mutant and wild-type plants that presumably are to some extent interconnected. First, in the mutant the pool size of subfraction I was unchanged during the light–dark cycle. In contrast, the size of the subfraction I pool of the wild type increased during darkness (Figure 5). Second, the cytosolic heteroglycans possessed a higher glucosyl content than the wild-type controls. Most of the additional glucosyl moieties are located in outer chains accessible to a hydrolytic or phosphorolytic attack (Figure 7), and arabinanase treatment results in the liberation of oligosaccharides enriched in glucosyl moieties (Figure 8b,c). Third, the priming efficiency of the cytosolic heteroglycans for both the cytosolic phosphorylase and DPE2 was higher in mutant than wild-type plants (Tables 2 and 3). It is reasonable to assume that these differences are, at least in part, caused by changes in the close vicinity of the actual acceptor sites, e.g. the size and/or structure of non-reducing oligoglucosyl termini. This implies that the cytosolic heteroglycans possess a high degree of submolecular variability that strongly affects the action of the two carbohydrate-active enzymes.

It is highly unlikely that the increased glucose content that was observed in the cytosolic heteroglycans from the DPE2-deficient mutant is due to an unnoticed contamination of starch-derived glucans. No correlation between the starch content of the starting material and the monomer patterns of SHG has been observed. In fact, the SHG preparations obtained from two starch-deficient Arabidopsis mutants contained either an unchanged or an even higher glucose level (Figure 9a,b). Likewise, the DPE1-deficient mutants possess an elevated starch content (Chritchley et al., 2001) but this does not correlate with the glucose level of the SHG preparations. Furthermore, a much higher glucose content was observed in oligosaccharides released from the heteroglycans by the action of an endo-α-arabinanase (Figure 8). In addition to glucose, these oligosaccharide chains contained xylose, mannose, galactose and arabinose and, therefore, are hetero-oligosaccharides. As shown previously, the arabinanase used does not attack homoglucans under the conditions used (Fettke et al., 2005b). It is important to note that, compared with the wild-type samples, the glucose-rich oligosaccharides are shifted towards the higher-molecular-weight region. A similar phenomenon has recently observed when heteroglycans were elongated in vitro by the action of the recombinant cytosolic phosphorylase isozyme (Fettke et al., 2005b). However, the shift in size presented in this paper is entirely due to in vivo processes. Thus, in summary the quantitative changes in the SHG monomer patterns as observed for the DPE2-deficient mutant reflect structural alterations in the cytosolic heteroglycans.

The higher glucose content of the mutant-derived heteroglycans is difficult to explain if DPE2 is the only enzyme that transfers glucosyl residues to the cytosolic heteroglycans. In this case the knock-out of DPE2 would be expected to result in high maltose levels but a lower glucosyl content of the heteroglycans. High maltose levels have been observed in the DPE2-deficient mutant (Chia et al., 2004) but, as shown in this paper, the heteroglycans possess an increased rather than a decreased glucose content.

In this context it is important to consider the phenotype of other starch-related mutants. We have studied heteroglycans from two starch-deficient mutants that are incapable of forming starch-derived maltose. Whilst the apoplastic glycans (subfraction II) remained unchanged the cytosolic heteroglycans (subfraction I) from both mutants also possess an elevated glucosyl content, and this effect is most pronounced during the light period (Figure 9a,b). Almost all of these glucosyl residues were removed by an in vitro treatment with α-amylase (Figure 9c). This indicates that they are linked via α-1,4 interglucose bonds and have a peripheral position within the heteropolysaccharides. Similar to the DPE2 mutant, both starch-deficient mutants exhibit an increased level of the cytosolic phosphorylase isozyme whilst the plastidial isoform as well as DPE2 remain unchanged (Figure 9d; data not shown). In contrast to all other mutants studied in this paper, the DPE1-deficient mutant that lacks a functional plastidial transglucosidase (Chritchley et al., 2001) possesses heteroglycans whose glucosyl content is significantly lower than that of the wild-type control (Figure 9e). In this mutant the level of cytosolic phosphorylase isoform is not increased (Chritchley et al., 2001). Based on these data the following conclusions are reached.

First, two cytosolic transferases, the cytosolic phosphorylase isoform and DPE2, are involved in cytosolic heteroglycan-related carbon metabolism.

Second, we propose that the cytosolic heteroglycans are linked to the plastidial carbon metabolism via at least two routes. In one route glucosyl residues from maltose are transferred onto the cytosolic heteroglycan via DPE2, then removed and converted to glucose 1-phosphate via the cytosolic phosphorylase. In the leaf, glucose 1-phosphate is then available either for sucrose synthesis or for general cellular metabolism. In a second route the cytosolic heteroglycans are linked via glucose to the plastidial carbon metabolism. In the latter the cytosolic phosphorylase isoform is also involved. The ratio between both fluxes is, to some extent, variable: a restriction of the maltose-dependent path results in an enhancement of the glucose-mediated route and an elevated level of cytosolic phosphorylase isoform.

Third, the interaction between the cytosolic heteroglycans and the plastidial carbon metabolism is not restricted to the period of net starch degradation but rather occurs also during the light period and in the absence of any functional starch biosynthesis. Thus, cytosolic heteroglycan metabolism is driven by carbohydrates that are derived from either starch or intermediates of the Calvin cycle.

These conclusions are further consistent with the recently described phenotype of the dpe2/dpe1 double mutant (Smith et al., 2005). This mutant possesses an extreme reduction in growth even compared with other starch-related mutants. It appears that in these mutants both the maltose- and the glucose-dependent link between the plastidial and the cytosolic carbon metabolism are not functional. In this double mutant none of the two routes can partially compensate for each other and consequently a far more severe phenotype is obtained.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Plant material, isolation of soluble heteroglycans and carbohydrate quantification

Wild-type plants and mutants of A. thaliana L. were grown under controlled conditions as recently described (Fettke et al., 2005a). Water-soluble heteroglycans (SHG0) were isolated from leaves of wild-type plants or mutants of A. thaliana L. and further processed as previously described (Fettke et al., 2005a). Carbohydrates were quantified after acid hydrolysis as described (Fettke et al., 2004).

High performance anion-exchange chromatography and field flow fractionation

Anion exchange chromatography was done with an HPAEC-PAD DX-600 system (Dionex, Idstein, Germany). For separation of monosaccharides the column (CarboPac®-PA1) was flushed with 200 mm NaOH for 15 min and was then equilibrated with water for 20 min. Analytes were eluted with water (flow rate of 1 ml min−1) for 60 min. The sensitivity of the pulsed amperometric detection was enhanced by the post-column addition of 300 mm NaOH (flow rate 0.5 ml min−1). Alternatively a CarboPac®-P20 (diameter 3 mm) column was used, flushed with 200 mm NaOH for 10 min, equilibrated with 3 mm NaOH for 15 min and eluted with 3 mm NaOH for 25 min (flow rate of 0.5 ml min−1) without post-column addition. Depending on the program used the order of the eluted rhamnose and abrabinose differed.

Oligosaccharides were eluted using a linear gradient of sodium acetate (5 to 500 mm in 30 min), followed by an elution with 500 mm sodium acetate dissolved in 100 mm NaOH (10 min). Alternatively, the oligosaccharides were eluted in a sodium acetate gradient (5 to 500 mm in 110 min).

Field flow fractionation was performed as previously described (Fettke et al., 2005a).

Protein preparation

Buffer-soluble proteins were extracted from leaves of wild-type plants of A. thaliana L. using a mixture of 100 mm HEPES-NaOH pH 7.5, 1 mm EDTA, 5 mm dithioerythritol (DTE), and 200 μm phenylmethanesulfonyl fluoride as grinding buffer (1 ml g−1 fresh weight). Following centrifugation (12 min at 20 000 g; 4°C) the supernatant was passed through a nylon net (and designated as buffer-soluble proteins). An aliquot of the buffer-soluble proteins were subjected to precipitation with ammonium sulphate (30 to 70% saturation). Precipitated proteins were collected, dissolved in 1 ml grinding buffer and passed through a PD 10 column equilibrated with grinding buffer. The proteins eluting in the void volume are designated as the DPE2-containing protein fraction.

Heterologous expression and purification of DPE2

Full-length DPE2 cDNA was amplified from total RNA extracted from rosette leaves of A. thaliana ecotype Col-0. The forward and reverse primers were designed with attB sites for recombination reactions into the Gateway® cloning system (Invitrogen GmbH, Karlsruhe, Germany). Forward primer: 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTCCatgatgaatctaggatctctttcg-3′. Reverse primer: 5′-GGGGACCACTTTGTACAAGAAAGTCGGGTAttatgggtttggcttagtcg-3′.

The PCR product was introduced into the donor vector pDONR201 via the BP reaction, then recombined into the destination vector pDEST17 via the LR reaction. The resulting construct expresses an N-terminal polyhistidine fusion of DPE2 and this was transformed into the Rosetta strain of Escherichia coli by heat-shock.

Enzyme purification was from 5 l of bacterial culture grown in Luria Bertrani broth at 21°C for 16 h after induction with isopropyl-β-D-thiogalactopyranoside (IPTG). Cells were pelleted by centrifugation and stored at −80°C until use. After thawing, cells were agitated for 15 min at room temperature in 200 ml Bugbuster and 50 μl Benzonase (Merck Biosciences Ltd, Novagen Brand, Darmtadt, Germany). All subsequent steps were at 0–4°C. After clarification by centrifugation, the supernatant was applied at 5 ml min−1 to a cobalt-charged HiTrap chelating column attached to an Akta fast protein liquid chromatograph (GF Healthcare Europe GmbH, Munich, Germany) and equilibrated with 0.1 m K phosphate (pH 7.5), 0.1 m NaCl and 20 mm imidazole. After washing with this medium, the column was eluted with a 50-ml gradient from 20 to 500 mm imidazole in the same medium, and 5-ml fractions were collected and assayed for DPE2 activity and protein. The three fractions of highest specific activity were combined, concentrated to 100 μl by centrifugation through a microconcentration device, and loaded onto a Superdex 200 column equilibrated with 50 mm Tris (pH 7.5), 100 mm KCl, 2 mm dithiothreitol, at a flow rate of 0.5 ml min−1. Fractions of 0.5 ml were collected. The DPE2 protein emerged as a single peak that was judged to be at least 95% pure by SDS-PAGE analysis. Peak fractions were combined, concentrated fivefold by centrifugation through a microconcentration device, brought to a glycerol concentration of 20%(v/v), and stored at −80°C. Little or no loss of activity was observed over several weeks of storage. Specific activity is about 0.5 μmol min−1 mg−1 protein.

Heterologous expression and purification of the cytosolic phosphorylase isozyme

The recombinant (untagged) cytosolic phosphorylase isoform (Pho 2 from Vicia faba L.) was expressed in E. coli and purified by affinity chromatography as described elsewhere (Fettke et al., 2004).

Protein quantification

Soluble proteins were quantified according to Bradford (1976) using the microversion of the Bio-Rad protein assay (Bio-Rad, Munich, Germany) and bovine serum albumin as standard.

In vitro binding assays

For coupling to Sepharose beads SHGT (22 mg glucose equivalents) was dissolved in 2.5 ml water and brought to pH 13 by adding NaOH. For a (SHGT-free) control 2.5 ml water were adjusted to pH 13. Two aliquots (2.5 g each) of epoxy-activated Sepharose 6B (Amersham Biosciences, Uppsala, Sweden) were prepared according to the instructions of the manufacturer and were then resuspended in water (2.5 ml each). Following the addition of the alkaline SHGT solution or the alkaline water the suspension was incubated for 22 h at 40°C. Coupling of SHGT was terminated by washing with water (40 ml per mixture) and then by blocking with 1 m ethanolamine (pH 8) for 5 h at 40°C. The Sepharose beads were then subjected to three cycles of washing with an acidic (0.5 m NaCl, dissolved in 0.1 m acetate pH 4.0) and alkaline (0.5 m NaCl, dissolved in 0.1 m Tris-HCl pH 8.0) solution. Finally, the Sepharose suspensions were washed several times with water. For binding assays, the Sepharose gels were transferred into small columns and the buffer-soluble proteins (2 mg each) that had been extracted from wild-type leaves in grinding buffer (as above) were passed through the gel. Subsequently, the two gels were subjected to a three-step elution procedure. First, gels were washed with 0.1 m NaCl, dissolved in grinding buffer (12 ml each). From this elution step, most of the volume (9 ml) was discharged but the last 3 ml each were collected and analyzed for proteins. Second, the gels were eluted with 1 m NaCl dissolved in grinding buffer (3 ml each) and the two eluates were collected. Third, the gels were eluted with 6 m urea (3 ml each) and the eluates were collected. Each of the six eluate fractions were concentrated and washed by filtration through a 10 kDa filter. The samples were then mixed with a SDS-containing denaturing buffer and were heated for 10 min at 95°C. Following SDS-PAGE using a 7.5% (w/v) total monomer concentration (T) separation gel, proteins were transferred to nitrocellulose and analyzed using a polyclonal antibody directed against DPE2 (Chia et al., 2004) or against the cytosolic phosphorylase isozyme (Fettke et al., 2004). Electrophoresis and Western blotting were performed as previously described (Fettke et al., 2004).

Native PAGE and activity staining

Native PAGE, affinity electrophoresis and affinophoresis were performed as described elsewhere (Fettke et al., 2004). For staining of the DPE2 activity the separation gel was incubated in a mixture containing 100 mm citrate-NaOH pH 6.5 and 25 mm maltose. As a control, maltose was omitted in the incubation mixture. Following incubation overnight at 37°C, the slab gels were stained with iodine.

In vitro14C-labeling experiments

For glycan labeling by a phosphorylase-catalyzed glucosyl transfer SHGL (10 μg each except where stated) was incubated at 37°C with 20 mm glucose 1-phosphate, [U-14C]glucose 1-phosphate (specific radioactivity 5.4 kBq μmol−1), 100 mm citrate-NaOH pH 6.5 and recombinant (untagged) Pho 2 (4 nkat). The final volume was 60 μl. At intervals, the reaction mixtures were heated for 5 min at 95°C and were then centrifuged for 5 min at 12 000 g. The supernatants were separated from low-molecular-weight compounds by filtration through a 10 kDa filter and the radioactivity content of the retentate was monitored using a liquid scintillation counter (for details see Fettke et al., 2005a). For glycan labeling by a DPE2-catalyzed glucosyl transfer 14C-labeled maltose [5.26 kBq μmol−1 (5mm); 1.32 kBq μmol−1 (20 mm)] was used. A mixture of 100 mm citrate-NaOH pH 6.5 and 2 mm DTE served as buffer. The incubation temperature throughout was 37°C.

Glucosyl acceptor specificity of DPE2

The transfer of glucosyl residues from glycogen or oligoglucans to monosaccharides was analyzed by HPAEC following incubation of recombinant DPE2 (2.5 μg each) for 90 min at 37°C in a mixture containing 100 mm citrate-NaOH pH 6.5, 30 mm monosaccharide (as stated) and glycogen (20 μg) or maltodextrins (30 μg). Following incubation the reaction mixtures were passed through a membrane filter [molecular weight cut-off (MWCO) 10 kDa] and the filtrate was analyzed by HPAEC-PAD.

Enzymatic treatment of SHGL

In a final volume of 50 μl, SHGL (100 μg each) was incubated with a fungal endo-α-1,5-arabinanase (EC 3.2.1.99; 23 μg each) and 25 mm sodium acetate buffer pH 5.5 for 7 h at 40°C. Subsequently, the reaction mixtures were passed through a membrane filter (MWCO 10 kDa) and each retentate was washed with 50 μl water. The two filtrates from each incubation mixture were combined. Aliquots were analyzed by HPAEC-PAD (oligosaccharide mode). Eluate fractions were collected as stated and hydrolyzed for HPAEC analysis. Similarly, the washed retentates were hydrolyzed and the monosaccharides released were analyzed by HPAEC-PAD (monosaccharide mode).

Removal of glucosyl residues from SHGL prepared from the DPE2-deficient mutant were performed with SHGL (5 μg each) incubated for 90 min at 37°C in a mixture containing either α-amylase from Bacillus amyloliquefaciens (4.8 units) and 25 mm sodium acetate buffer pH 5.5 or the recombinant cytosolic phosphorylase isozyme (2 nkat) in 250 mm orthophosphate buffer pH 7.0. The reaction mixtures were passed through a membrane filter (MWCO 10 kDa) and were washed repeatedly with water. The resulting retentates were hydrolyzed and analyzed by HPAEC-PAD.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Financial support by the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 429, Molecular Physiology, Energetics, and Regulation of Plant Primary Metabolism’ TP B2) to MS and a fellowship of the State of Brandenburg to JF are gratefully acknowledged. The authors thank Professor Dr T. Linker and Ms Anja Schulenburg (both University of Potsdam) for valuable discussions on carbohydrate structures and the Interdisciplinary Center for Advanced Protein Technologies (APT; University of Potsdam) for support. The arabinanase preparation used in this study was a generous gift from Dr Kirk Matthew Schnorr, Novozymes, Denmark. The authors are indebted to Ms Silke Gopp for excellent technical assistance. The authors thank Ms Andrea Mohrenweiser and Dr Simon Poeste for kindly providing the recombinant cytosolic phosphorylase from Vicia faba L.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Ball, S.G. and Morell, M.K. (2003) From bacterial glycogen to starch: understanding the biogenesis of the plant starch granule. Annu. Rev. Plant Biol. 54, 207233.
  • Bradford, M.M. (1976) A rapid and sensitive method for the quantification of microgram quantities of proteins utilizing the principle of protein-dye binding. Anal. Biochem. 72, 238245.
  • Caspar, T., Huber, S.C. and Somerville, C. (1985) Alterations in growth, photosynthesis, and respiration in a starchless mutant of Arabidopsis thaliana (L.) deficient in chloroplast phosphoglucomutase activity. Plant Physiol. 79, 1117.
  • Chia, T., Thorneycroft, D., Chapple, A., Messerli, G., Chen, J., Zeeman, S.C., Smith, S.M. and Smith, A.M. (2004) A cytosolic glucosyltransferase is required for conversion of starch to sucrose in Arabidopsis leaves. Plant J. 37, 853863.
  • Chritchley, J.H., Zeeman, S.C., Takaha, T., Smith, A.M. and Smith, S.M. (2001) A critical role for disproportionating enzyme in starch breakdown is revealed by a knock-out mutant in Arabidopsis. Plant J. 26, 89100.
  • Fettke, J., Eckermann, N., Poeste, S., Pauly, M. and Steup, M. (2004) The glycan substrate of the cytosolic (Pho 2) phosphorylase isozyme from Pisum sativum L.: identification, linkage analysis and subcellular localization. Plant J. 39, 933946.
  • Fettke, J., Eckermann, N., Tiessen, A., Geigenberger, P. and Steup, M. (2005a) Identification, subcellular localization and biochemical characterization of water-soluble heteroglycans (SHG) in leaves of Arabidopsis thaliana L.: distinct SHG reside in the cytosol and in the apoplast. Plant J. 43, 568586.
  • Fettke, J., Poeste, S., Eckermann, N., Tiessen, A., Pauly, M., Geigenberger, P. and Steup, M. (2005b) Analysis of cytosolic heteroglycans from leaves of transgenic potato (Solanum tuberosum L.) plants that under- or overexpress the Pho 2 phosphorylase isozyme. Plant Cell Physiol. 46, 19842004.
  • Lin, T.-P., Caspar, T., Somerville, C. and Preiss, J. (1988) Isolation and characterization of a starchless mutant of Arabidopsis thaliana (L.) lacking ADP-glucose pyrophosphorylase activity. Plant Physiol. 86, 11311135.
  • Lu, Y. and Sharkey, T.D. (2004) The role of amylomaltase in maltose metabolism in the cytosol of photosynthetic cells. Planta, 207, 271274.
  • Niittylä, T., Messerli, G., Trevisan, M., Chen, J., Smith, A.M. and Zeeman, S.C. (2004) A previously unknown maltose transporter essential for starch degradation in leaves. Science, 303, 8789.
  • Nothnagel, E.A. (1997) Proteoglycans and related components in plant cells. Int. Rev. Cytol. 174, 195291.
  • Scheidig, A., Fröhlich, A., Schulze, S., Lloyd, J.R. and Kossmann, J. (2002) Down-regulation of a chloroplast-targeted β-amylase leads to a starch-excess phenotype in leaves. Plant J. 30, 581591.
  • Shimura, K. (1990) Progress in affinophoresis. J. Chromatogr. 510, 251270.
  • Shimura, K. and Kasai, K. (1998) Capillary affinophoresis as a versatile tool for the study of biomolecular interactions: a minireview. J. Mol. Recognit. 11, 134140.
  • Smith, A.M., Zeeman, S.C. and Smith, S.M. (2005) Starch degradation. Annu. Rev. Plant Biol. 56, 7398.
  • Steup, M. (1990) Starch degrading enzymes. In Methods in Plant Biochemistry, Vol. 3 (Lea, P.J., ed.). London: Academic Press, pp. 103128. ISBN 0-12-46 1013-7.
  • Steup, M. and Schächtele, C. (1981) Mode of glucan degradation by purified phosphorylase forms from spinach leaves. Planta, 153, 351361.
  • Weber, A.P.M. (2004) Solute transporters as connecting elements between cytosol and plastid stroma. Curr. Opin. Plant Biol. 7, 247253.
  • Weise, S.E., Weber, A.P.M. and Sharkey, T.D. (2004) Maltose is the major form of carbon exported from the chloroplast at night. Planta, 218, 474482.
  • Weise, S.E., Kim, K.S., Stewart, R.P. and Sharkey, T.D. (2005) β-maltose is the metabolically active anomer of maltose during transitory starch degradation. Plant Physiol. 137, 756761.
  • Wyatt, P.J. (1993) Light scattering and the absolute characterization of macromolecules. Anal. Chim. Acta, 272, 140.
  • Yang, Y. and Steup, M. (1990) Polysaccharide fraction from higher plants which strongly interacts with the cytosolic phosphorylase isozyme. I. Isolation and characterization. Plant Physiol. 94, 960969.
  • Yariv, J., Rapport, M.M. and Graf, L. (1962) The interaction of glycosides and saccharides with antibody to the corresponding phenylazo glycoside. Biochem. J. 85, 383388.
  • Yu, T.-S., Zeeman, S.C., Thorneycroft, D. et al. (2005) α-amylase is not required for breakdown of transitory starch in Arabidopsis leaves. J. Biol. Chem. 280, 97739779.