- © 2005 American Society of Plant Biologists
Abstract
Mutant lines defective for each of the four starch debranching enzyme (DBE) genes (AtISA1, AtISA2, AtISA3, and AtPU1) detected in the nuclear genome of Arabidopsis (Arabidopsis thaliana) were produced and analyzed. Our results indicate that both AtISA1 and AtISA2 are required for the production of a functional isoamylase-type of DBE named Iso1, the major isoamylase activity found in leaves. The absence of Iso1 leads to an 80% decrease in the starch content in both lines and to the accumulation of water-soluble polysaccharides whose structure is similar to glycogen. In addition, the residual amylopectin structure in the corresponding mutant lines displays a strong modification when compared to the wild type, suggesting a direct, rather than an indirect, function of Iso1 during the synthesis of amylopectin. Mutant lines carrying a defect in AtISA3 display a strong starch-excess phenotype at the end of both the light and the dark phases accompanied by a small modification of the amylopectin structure. This result suggests that this isoamylase-type of DBE plays a major role during starch mobilization. The analysis of the Atpu1 single-mutant lines did not lead to a distinctive phenotype. However, Atisa2/Atpu1 double-mutant lines display a 92% decrease in starch content. This suggests that the function of pullulanase partly overlaps that of Iso1, although its implication remains negligible when Iso1 is present within the cell.
Amylopectin, the major polysaccharide fraction of starch granules, can be distinguished from glycogen by the presence of an asymmetric distribution of α-1,6 branches within the molecule (for review, see Buléon et al., 1998; Myers et al., 2000). It is thought that this concentration of branches in certain regions of the macromolecule is responsible for the clustering of chains, which in turn allows the formation of double helical structures that align and crystallize (Hizukuri, 1986; Manners, 1989; for review, see Ball and Morell, 2003). In 1996, Ball et al. suggested that this distribution and, consequently, the crystallization of polysaccharides in macrogranular structures result from the action of the isoamylase family of debranching enzymes (DBEs). This suggestion has been termed the amylopectin-trimming model, which hypothesized that misplaced, loosely branched chains are trimmed from a maturing structure to allow polysaccharide crystallization. This result was suggested by the description of mutants of maize (Zea mays), rice (Oryza sativa), and the monocellular alga Chlamydomonas reinhardtii that, in the absence of isoamylase, substituted amylopectin synthesis by that of glycogen (Pan and Nelson, 1984; James et al., 1995; Mouille et al., 1996; Nakamura et al., 1996; Kubo et al., 2005). In C. reinhardtii, the substitution of starch by glycogen synthesis was essentially complete and therefore thought to be mandatory to obtain macrogranular structures. Subsequently, mutants or antisense plants with reduced amounts of isoamylase activity were produced in barley (Hordeum vulgare), potato (Solanum tuberosum), and Arabidopsis (Arabidopsis thaliana; Zeeman et al., 1998; Burton et al., 2002; Bustos et al., 2004). Alternative interpretations to the trimming model were proposed, including a specific function in the clearing of water-soluble polysaccharides (WSPs) or in the priming of starch granule formation (Zeeman et al., 1998). A specific claim was made that the block on starch synthesis of defects in isoamylase activity was essentially incomplete. In Arabidopsis, mutants that lacked isoamylase activity were said to display a normal amylopectin structure, an observation thought to be incompatible with the trimming model (Zeeman et al., 1998).
In this study, we report the starch accumulation phenotypes of different insertion mutant lines of Arabidopsis defective for each of the four DBE isoforms detected in the Arabidopsis genome: AtISA1 (At2g39930), AtISA2 (At1g03310), AtISA3 (At4g09020), and AtPU1 (At5g04360). Moreover, we have generated and analyzed a double-mutant line that is defective for both pullulanase and the major form of an isoamylase-type DBE (AtISA2/AtPU1). Our results contradict the observation of Zeeman et al. (1998) and provide information about the specific function of each DBE in starch metabolism.
RESULTS
DBE Genes in Arabidopsis and Selection of the Corresponding Mutant Lines
Four genetically independent loci corresponding to DBEs were detected in the nuclear genome of Arabidopsis: At2g39930 (AtISA1), At1g03310 (AtISA2), At4g09020 (AtISA3), and At5g04360 (AtPU1; for a complete description of these loci, see http://www.starchmetnet.org/GeneList/GeneListFrameset.htm). The latter corresponds to a pullulanase-type DBE, whereas the former three correspond to an isoamylase-type DBE. Insertion mutant lines for each of these genes were selected from different insertion mutant collections (Nottingham Arabidopsis Stock Centre [NASC] and Génoplante). A rapid description of these lines is presented in Table I. Except for Atisa2-1, T-DNA is always inserted within coding sequences (Fig. 1). Homozygous mutant lines were selected from a pooled sample of T3 seeds for each gene of interest. Expression of mRNA was tested for each line by reverse transcription (RT)-PCR amplification of the corresponding transcript (18S RNA amplification was used as a positive control for each sample). The mRNA synthesis is completely abolished in Atisa2-1 (Fig. 1B) and Atpu1-1 (Fig. 1D). Conversely, a transcript is still synthesized and was amplified in Atisa1-1 (Fig. 1A), Atisa2-2 (Fig. 1B), and Atisa3-1 (Fig. 1C). Nevertheless, when the RT-PCR amplification was performed by a primer pair overlapping the insertion site of the T-DNA, no amplicon was observed in the mutant lines. It is likely that these mutations do not lead to the production of a protein and might therefore be considered as null mutations for the genes of interest.
Description of mutant lines used in this work
Molecular organization of DBE mutant alleles. Both gDNA and mRNA organizations are presented for each DBE gene of Arabidopsis. Red flags specify T-DNA insertion sites. Bold lines correspond to exons, while thin lines represent introns. Double-sided arrows indicate mRNA regions amplified by RT-PCR for the different primer pairs designed for each T-DNA insertion (see supplemental data for primer sequences). Images below mRNA structure correspond to agarose gels performed with RT-PCR amplification products. RT-PCR amplifications were produced upstream, downstream, and spanning the T-DNA insertion site. M, mutant line; W, wild-type line.
Double-mutant lines defective for both AtISA2 and AtPU1 were generated after crosses between lines CTI15 (Atisa2-1) and CQU5 (Atpu1-1). Two segregating lines, named 5FWB6 and 3FWF5, were selected from two independent crosses and further analyzed for starch accumulation phenotype. In the following paragraphs, only the results concerning line 5FWB6 will be presented, since the same results were obtained with line 3FWF5.
No significant modification in seed germination level, plant growth, flowering rate, or silique formation was noticed for the different mutant lines grown under the standard conditions used for this work (16-h light/8-h dark).
Enzymatic Characterization of the Mutant Lines
DBE activities were checked in all mutant lines through zymogram analysis (Fig. 2). This technique has proven to represent one of the best ways to correlate the alteration of a specific starch-metabolizing activity with one particular mutated gene. Soluble proteins of a leaf crude extract are loaded onto a starch-containing polyacrylamide gel and separated under native conditions. Their activity is revealed by iodine staining after incubation overnight at room temperature. Under these conditions, several starch-modifying activities were observed in Arabidopsis leaves ranging from starch-branching enzymes to β-amylase (Fig. 2, A and B). Nevertheless, some bands of unidentified activity were detected on these gels. These activities were not yet associated to a particular starch-metabolizing enzyme after mutant analysis.
Zymograms of starch-modifying enzymes. A and B, Approximately 100 μg of soluble proteins from a leaf crude extract were loaded onto a starch-containing polyacrylamide gel. After migration under native conditions, gels were incubated overnight at pH 6. Starch-modifying activities were visualized by iodine staining. C, Approximately 100 μg of soluble proteins from a leaf crude extract were loaded onto a red pullulan-containing gel. After migration under native conditions, the gel was incubated overnight at pH 8. White bands correspond to pullulanase activity where red pullulan was degraded.
Mutation at the AtPU1 locus in line CQU5 leads to the specific disappearance of a blue band (Fig. 2B). To confirm that this blue band actually corresponds to pullulanase, we have submitted the same leaf extract to a specific pullulanase zymogram performed with a red pullulan-containing gel (Fig. 2C). After migration, the gel was incubated overnight at room temperature. Only one clear band (corresponding to degradation of red pullulan) was detected in the wild-type extract, indicating that only one pullulanase activity is present in Arabidopsis leaves (as already suggested by genome sequence analysis). This band is lacking in Atpu1-1, but remains apparent in Atisa2-1, suggesting that a mutation at this locus does not affect pullulanase activity.
A pale blue band situated at the edge of the starch-containing gel is missing in Atisa1-1, Atisa2-1, and Atisa2-2 (Fig. 2, A and B). This band, which we have named Iso1, corresponds to the previously described isoamylase activity lacking in the dbe1 mutant, which was localized within the plastid (Zeeman et al., 1998). This result indicates that both AtISA1 and AtISA2 genes are required for the normal activity of this enzyme in Arabidopsis leaves. Nevertheless, it is yet impossible to determine whether both proteins belong to the same heteromultimeric complex or whether their activities are simply interdependent. Despite several attempts (performed under several conditions), it was impossible to restore any visible Iso1 activity on the zymogram by mixing crude extracts produced from single-mutant lines (i.e. lines Atisa1-1 and Atisa2-2). This suggests either the instability of monomers in the single-mutant lines or the presence of an intricate mechanism, not reproduced in our crude in vitro conditions, required for the formation of an active Iso1 enzyme. We have established by the zymogram technique that this enzyme is not only active on starch but also on various substrates such as maize amylopectin and β-limit dextrins (Fig. 3, A and B). However, this enzyme appears as a brown-staining band on glycogen- and phytoglycogen-containing zymograms (Fig. 3, C and D, respectively) that is specifically lacking in the lines mutated at genes AtISA1 and AtISA2. Such coloration for isoamylase activity on glycogen-containing gel was previously described for C. reinhardtii (Colleoni et al., 1999). Although it is not very obvious from the figure, a clearer region was observable in the center of the brownish band, indicating that the substrate is only slowly degraded by the enzyme within the gel (Fig. 3C). To ensure that this brown band is not related to α-4-glucanotransferases (d-enzyme), we have loaded onto the same glycogen-containing gel the cellular extracts prepared from two mutant lines defective for dpe1 and dpe2 genes (At5g64860 and At2g40840 genes, respectively, corresponding to both forms of d-enzyme found in the Arabidopsis genome). The brown band was still present in both d-enzyme mutant lines (Fig. 3C). This result suggests that glycogen and phytoglycogen do not represent the best substrates for Iso1, as these molecules seem to be only slowly degraded within the gel.
Zymograms of starch-modifying enzymes. Approximately 100 μg of soluble proteins from a leaf crude extract were loaded onto polyacrylamide gels containing different substrates. A to D, Maize amylopectin, maize β-limit dextrins, rabbit liver glycogen, and maize phytoglycogen, respectively. After migration and incubation overnight at room temperature at pH 6, starch-modifying activities were revealed by iodine staining. Atdpe1 and Atdpe2 lanes correspond to mutant lines defective for genes At5g64860 and At2g40840, respectively.
The situation is different in Atisa3-1. Although this gene is expressed in the leaves during the illuminated period (http://www.starchmetnet.org/Datapages/AtISA3/AtISA3Frameset.htm), no modification of the zymogram profile was reported in this mutant line (Fig. 2A). Zymograms performed with other types of polysaccharides (glycogen, amylopectin, maize phytoglycogen, β-limit dextrins) entrapped within the polyacrylamide matrix also failed to detect any modification of activity in Arabidopsis leaves (Fig. 3). The missing activity might thus require other components to be active in vitro or might be extremely sensitive to the surrounding conditions such as polysaccharide concentration and/or redox environment that are not reproduced in our in vitro conditions.
Both Iso1 and pullulanase activities were lacking in the two double-mutant lines Atisa2-1/Atpu1-1, as shown in Figure 2, B and C.
No other hydrolytic activity (as well as other starch-metabolizing enzymes; data not shown) was affected by the different mutations that were studied in this work.
Starch and WSP Accumulation Phenotypes
We have performed a complete study of the polysaccharide accumulation phenotype for every mutant line. Starch and WSP content was measured at the end of the illuminated period (Fig. 4). We have used both Wassilewskija (WS) and Columbia-0 (Col-0) ecotypes as wild-type references since the mutant lines were produced in both genetic backgrounds (WS for Génoplante lines and Col-0 for NASC lines). Starch levels were expressed in comparison to the corresponding wild type (starch content in wild-type lines was arbitrarily assigned a value of 100%).
Comparison of leaf starch and WSP content in wild-type and mutant lines. Both starch and WSP were assayed by amyloglucosidase assay. The results presented here are the mean of several independent extractions (±se shown as thin bars). Starch levels (A, C, and D) were standardized to the wild-type content, which was arbitrarily assigned a value of 100% (n = 4 for WS, Atisa2-1, Atpu1-1, and Atisa2-1/Atpu1-1; n = 3 for Col-0, Atisa1-1, and Atisa3-1; n = 2 for Atisa2-2 and 1CDB3). WSP amounts (B and E) are expressed in mg g−1 of fresh weight (n = 5 for Atisa2-1/Atpu1-1; n = 4 for Atisa2-1; n = 3 for WS; n = 2 for other lines). Samples were collected at the end of the light period (16-h-light/8-h-dark cycle), except for C, where samples were collected at the end of the dark period. FW, Fresh weight; WT, wild type.
Starch content was not significantly modified in mutant lines defective for pullulanase activity (Atpu1-1). Conversely, the starch level was strongly reduced in mutant lines affected at AtISA1 or AtISA2 loci that are both required for Iso1 activity in leaves. In these cases, starch amounts were only 15% to 21% of the wild-type content. Whereas WSP content was only slightly increased in line CQU5 (2-fold increase), a strong increase (50-fold increase) in WSP content was recorded in Iso1-defective mutants.
The starch accumulation level in the double mutant (Atisa2-1/Atpu1-1) was also strongly reduced when compared to the wild-type reference (about 8% of remaining starch). The starch content was even lower when compared to Atisa1 and Atisa2 single mutants (an approximately 2-fold decrease). WSP amounts were strongly increased when compared to the wild type and were even higher than those measured in the single-mutant lines defective for Iso1 activity. Therefore, our results indicate that the combination of mutations at both AtISA2 and AtPU1 genes leads to a more severe phenotype regarding starch and WSP content in leaves.
The situation was different in Atisa3-1. In this case, a strong increase of the starch amount was monitored in the leaves at the end of the day (a 2.5-fold increase on average at the end of the day), while no striking modification of WSP content was observed. The same increase in leaf starch content was measured at the end of the dark period, as shown in Figure 4C (a 3.5-fold increase in average at the end of the dark period). No modification in WSP level at the end of the dark period was reported in Atisa3-1 when compared to the wild type (data not shown).
Finally, no significant alteration of the amylase-to-amylopectin ratio was observed in the different mutant lines analyzed in this study (determined by size exclusion chromatography performed on a Sepharose CL-2B column; data not shown).
Structural Analysis of Mutant Amylopectin and WSP
The impact of the different mutations or combination of mutations on amylopectin structure was analyzed and compared to the wild type. Amylopectin was purified by size exclusion chromatography on a Sepharose CL-2B column and subsequently analyzed. The wavelength of the maximal absorbance of the iodine-polysaccharide complex (λmax) of amylopectin was recorded (Table II). The values measured clearly indicate that the amylopectin structure is modified in Atisa1 and Atisa2 mutant lines. Conversely, the amylopectin structure seems to remain unaffected by a mutation at AtISA3 or AtPU1 as no significant modification of the amylopectin λmax could be observed in these samples when compared to the wild-type (Table II) references. Surprisingly, the λmax value of the residual amylopectin of the double-mutant line (Atisa2-1/Atpu1-1) was slightly lower than that of the wild type and consequently much lower than that of the corresponding single-mutant lines.
λmax, Wavelength at the maximal absorbance of the amylopectin-iodine complex; Col-0, Columbia; WS, Wassilewskija
Purified amylopectin was debranched to completion by a mix of two bacterial DBEs (isoamylase and pullulanase). Resultant linear glucans were separated and quantified by fluorophore-assisted capillary electrophoresis (FACE) or high-performance anion exchange chromatography with pulsed amperometric detection (HPAEC-PAD; Fig. 5) and compared to the corresponding wild-type reference (Fig. 6). Chain-length (CL) distribution profiles for both Atisa1-1 and Atisa2-2 mutants were very similar but different from those of the wild-type reference (Fig. 5, B, C, and A, respectively). Both mutant amylopectins showed a strong increase in the number of small chains (5 < degree of polymerization [DP] < 9) together with a significant decrease in the number of longer chains (11–17 DP; (Fig. 6, A and B). The same results were obtained with Atisa2-1. This result could partly explain the increase in λmax of the residual amylopectin observed in both lines since glucans shorter than 12 Glc residues in length are unable to interact with iodine (Banks et al., 1971).
CL distribution profiles of amylopectin purified from different lines. After purification by size exclusion chromatography on Sepharose CL-2B columns, amylopectin was completely debranched by a mix of isoamylase and pullulanase. The resulting glucans were analyzed by FACE (A–E) or HPAEC-PAD (F and G). The diagrams represent the percentage (y axis) of each glucan (x axis) in the total population of glucans that can be detected for each sample. Thin bars correspond to sd calculated on three independent analyses for each line, except for lines Atisa3-1 and Atisa2-1/Atpu1-1 (n = 2) and for WT (n = 4). A, Wild-type amylopectin (WS and Col-0 profiles were essentially the same and were combined). B, Atisa1-1 (N542704). C, Atisa2-2 (N529442). D, Atisa3-1 (N604008). E, Atpu1-1 (CQU5). F, Line 1CDB3 (wild-type progeny from a cross between lines CTI15 and CQU5). G, Double-mutant line Atisa2-1/Atpu1-1 (5FWB6). Individual glucans were normalized to total peak area showed in the diagram.
Difference plots corresponding to CL distribution profiles presented in Figure 5 (mutant-wild-type profile). A, Atisa1-1-wild type. B, Atisa2-wild type. C, Atisa3-1-wild type. D, Atpu1-1-wild type. E, Atisa2-1/Atpu1-1-1CDB3. Mean profiles of each sample displayed in Figure 5 were used to draw the difference plot.
These results were strikingly different from those previously reported by Zeeman et al. (1998) for the dbe1 mutant line. No modification of amylopectin CL distribution was reported by these authors. It is not clear yet whether this difference is a consequence of culture conditions or starch extraction, purification methods or analysis procedures. Nevertheless, whatever the mutation (to be found at the AtISA1 or AtISA2 gene), the Iso1 complex seems to be required for normal amylopectin synthesis in Arabidopsis leaves.
On the contrary, no significant alteration of the CL distribution profile of amylopectin was observed in Atisa3-1 (Figs. 5D and 6C). On the other hand, the Atpu1-1 mutant displayed a small, yet significant, difference (Figs. 5E and 6D). This result suggests that AtIsa3 either is not extensively involved in amylopectin synthesis or is involved at such a low level that its absence does not lead to a significant modification of the amylopectin structure when the Iso1 complex is present and active. The result obtained with Atpu1-1 is suggestive of a small impact on amylopectin structure that is qualitatively similar, but not identical to, that of Iso1.
The amylopectin CL profile of the Atisa2-1/Atpu1-1 double mutant was established (Fig. 5G) and compared to that of the corresponding wild-type reference (Fig. 5F). The difference plot, shown in Figure 6E, displays the profile similar to those of the Atisa1 and Atisa2 single mutants, although to a significantly less severe extent. This result could explain why the λmax of the residual amylopectin in the double-mutant line is reduced when compared to wild-type and single-mutant lines.
Finally, we have investigated the structure of the WSP that accumulates in Iso1-defective (Atisa1 and Atisa2) and Iso1/pullulanase-defective (Atisa2-1/Atpu1-1) lines. After purification, WSPs were first submitted to size exclusion chromatography on a Sephadex TSK HW50 column (Fig. 7, A–C, respectively). Two peaks were observed for each mutant line. The elution volume of the first peak (peak I in Fig. 7, A–C) matches that of rabbit liver glycogen, while the elution volume of the second peak (peak II in Fig. 7, A–C) corresponds to that of Glc. About two-thirds of the loaded material were found in peak I and one-third in peak II (determined from amyloglucosidase test after subtracting the tiny amount of free hexose assayed in the samples) whatever the line under study. The CL distributions of these peaks were determined by high-performance anion-exchange chromatography pulsed-amperometric detection (HPAEC-PAD) after complete debranching with a mixture of both isoamylase and pullulanase. Peak I CL profiles of the three mutant lines were essentially the same (Fig. 7, E–G) and were very similar to those of glycogen (Fig. 7D). As for glycogen, chains of DP 7 were the most abundant glucans within this soluble polymer. Peak II was almost solely composed of maltose (only a tiny amount of free hexose was also measured in peak II; data not shown).
Structural characterization of WSPs. After purification, WSPs extracted from Atisa1-1 (N542704), Atisa2-1 (CTI15) single-mutant lines (A and B, respectively), and the Atisa2-1/Atpu1-1 (5FWB6) double-mutant line (C) were loaded onto a TSK HW50 column. Elution was carried out in 10% DMSO at a flow rate of 12 mL h−1. Two-milliliter fractions were collected and assayed for α-1,4-linked glucans by amyloglucosidase test. y axis, glucan content in milligrams (after subtracting free hexose content). x axis, Elution volume in milliliters. Peak I was submitted to complete debranching and the resultant linear glucans were separated and quantified by HPAEC-PAD analysis. E to G correspond to peak I glucans from Atisa1-1 (N542704), Atisa2-1 (CTI15), and Atisa2-1/Atpu1-1 (5FWB6), respectively. D corresponds to CL distribution profile of rabbit liver glycogen.
DISCUSSION
We have shown that insertions at either AtISA1 (At2g39930) or AtISA2 (At1g03310) specifically lead to the disappearance in Arabidopsis leaves of one form of isoamylase, named Iso1. The same activity, demonstrated to be located within the plastid, was specifically absent in the dbe1 mutant described by Zeeman et al. (1998). In both cases, the absence of Iso1 leads to a dramatic decrease of starch content together with the accumulation of large amounts of WSP generally not found in wild-type plants. However, we have shown that the residual starch is composed of a structurally modified amylopectin. These results strongly suggest a direct function of Iso1 during the synthesis process of amylopectin as proposed by Ball et al. (1996). Observations of the same nature have been performed in both maize and rice, where modification in the nature of the isoamylase activity is accompanied by a change in amylopectin CL distribution (Kubo et al., 1999; Dinges et al., 2001). By contrast, an indirect function limited to the clearing of soluble glucans whose synthesis would otherwise compete with that of amylopectin (Zeeman et al., 1998) is difficult to reconcile with our observations in Arabidopsis and with those obtained in maize and rice. Moreover, our results indicate that Iso1 is much less active on both glycogen and phytoglycogen (Fig. 3, C and D) when compared to starch, amylopectin, or β-limit dextrins. This suggests that, if the clearing process effectively occurs within the cell, this isoamylase activity might not represent the major enzyme involved in the degradation of WSP in vivo or at such a low level that WSP would have to be observable in the wild-type line during starch synthesis.
It is yet impossible to indicate precisely how both AtISA1 and AtISA2 function together to produce an active Iso1 enzyme. It was already suggested that different isoamylase isoforms might interact together to form a functional hetero- or homomultimeric complex (Beatty et al., 1999; Fujita et al., 1999; Dauvillée et al., 2001; Hussain et al., 2003). A comparable situation might occur within Arabidopsis leaves. In this context, AtIsa1 and AtIsa2 might represent two subunits of the same heteromultimeric complex. Absence of either subunit would, therefore, destabilize such a complex and lead to the same phenotype as those observed in the different mutant lines. Therefore, combination of both mutations at AtISA1 and AtISA2 genes in one single plant is likely to produce more or less the same phenotype as that observed for single-mutant lines analyzed in this work. Further investigation is required to check whether both proteins are catalytically active.
The strong increase in starch content at the end of the illuminated and dark periods displayed by the Atisa3 mutant, together with the absence of modification in the amylopectin structure, strongly suggest that this isoamylase isoform defines a major enzyme of starch catabolism. This represents a demonstration of the involvement of a particular isoamylase-like activity in starch degradation. This hypothesis is strongly supported by the transcript expression profile for AtISA3 provided by transcriptomic analysis performed on short (8 h) and normal (12 h) day length (http://www.starchmetnet.org/Datapages/AtISA3/AtISA3Frameset.htm; Smith et al., 2004, respectively). The mRNA level of the AtISA3 gene perfectly matches those of other genes known to be involved in transient starch mobilization at night (mRNA level is maximal at the transition between the light and dark period, whatever the length of the illuminated period). In addition, the potato recombinant protein corresponding to AtISA3 was shown to be predominantly active on β-limit dextrins (Hussain et al., 2003). AtIsa3 and the Escherichia coli GlgX DBE both display a significantly shorter loop (loop 4) in the enzyme structure that is thought to correlate with the substrate preference. Indeed, according to Abe et al., (1999), longer loops would be required to accommodate the longer outer chains of glycogen or amylopectin, while shorter loops would be adapted to hydrolysis of short outer chains such as those displayed by both phosphorylase or β-amylase-limit dextrins. It is striking to note that both the E. coli GlgX and the Arabidopsis AtIsa3 display similar phenotypes consisting of polysaccharide overproduction due to a defect in catabolism (Dauvillée et al., 2005). It has been known since the pioneering work of Scheidig et al. (2002) that β-amylolysis is the major pathway of starch catabolism in plant leaves. β-Amylase defines an exo-hydrolase that processively digests the outer chains of starch polymers. However, β-amylase cannot bypass the branches and will stop degrading two to three residues from the α-1,6 branch. We propose that AtIsa3 defines the major DBE required to assist β-amylases for starch degradation in plant leaves at night. This will lead to the production of both α-maltotriose and α-maltose, which need to be metabolized. While a disproportionating enzyme (Dpe1) is expected to efficiently metabolize maltotriose, maltose is expected to define a very poor substrate for this enzyme. Conflicting results have been published concerning the cellular location of amylomaltase (Dpe2) in potato and Arabidopsis (Chia et al., 2004; Lloyd et al., 2004; Lu and Sharkey, 2004). In theory, this enzyme would effectively metabolize the α-maltose produced by AtIsa3. The precise localization of the enzyme and its substrate preferences toward α-maltose and β-maltose needs to be readdressed.
No significant modification of the starch accumulation phenotype was observed in the Atpu1-1 mutant. Only a very small increase in WSP content was recorded. The modification of amylopectin structure reported in our study is small by comparison to those recorded for the Atisa1 and Atisa2 mutants. Nevertheless, the defective pattern is similar, suggesting that both activities may play partly redundant functions during starch biosynthesis. The relative increase in chains above DP20 is small and endowed with large errors. We think this general increase stems from the selective decrease of chains between DP9 to DP17, which is not compensated in the pullulanase mutant by an increase in the amount of very small chains. This result is similar to that previously reported in maize leaves (Dinges et al., 2003), although with less expressivity in the case of Arabidopsis. This difference might reflect distinct enzymatic situations between the Arabidopsis and maize leaf. Pullulanase is known to be very effective in the hydrolysis of dextrins produced through α-amylases during starch degradation. We believe that the absence of a clear impact of pullulanase defect on starch amounts suggests an alternative function of this activity during starch mobilization. Indeed, pullulanase, also named limit dextrinase, is expected to be important whenever α-dextrins are produced. The latter are defined as the product of degradation of starch by α-amylases. The absence of impact on starch accumulation levels of the absence of pullulanase may reflect the relatively low importance of α-amylases for starch degradation in leaves (Scheidig et al., 2002). It remains likely that pullulanase plays an important role in starch degradation in tissues or physiological situations different from those defined by the mature leaf.
However, when the mutated copy of the AtPU1 gene was combined to the mutation at the AtISA2 locus in one single plant, a more severe phenotype for starch and WSP accumulation was observed, while the structure of the residual amylopectin was less severely modified. This result strongly suggests that pullulanase is partially redundant with the function of Iso1 for the trimming of pre-amylopectin chains during starch synthesis. However, pullulanase may only account for a minor part of this process in Arabidopsis leaves. This result correlates to those already reported in maize and rice endosperm (Kubo et al., 1999; Dinges et al., 2003). The impact on amylopectin structure of the loss of pullulanase in the absence of Iso1 might reflect the different substrate specificity displayed by pullulanase when compared to isoamylase.
MATERIALS AND METHODS
Materials
A CL-2B Sepharose column and Percoll were obtained from Amersham-Pharmacia Biotech (Uppsala); a TSK HW50F Sephadex column was obtained from Tosoh Bioscience (Stuttgart, Germany); the starch assay kit (EnzyPlus) was purchased from Enzytec (Toulouse, France); and the protein assay kit (Lowry method) was obtained from Bio-Rad Laboratories (Hercules, CA).
Arabidopsis Lines, Growth Conditions, and Media
Wild-type (WS and Col-0) and mutant lines of Arabidopsis (Arabidopsis thaliana) were obtained from the T-DNA mutant collections at Unité de Recherche en Génomique Végétale (URGV), INRA, Versailles, France (Bechtold et al., 1993; Bouchez et al., 1993, Balzergue et al., 2001) and NASC (Alonso et al., 2003). Standard procedures were employed for plant germination and growth. The plants were grown on peat-based compost (seeds were previously incubated at 4°C for 3 d before sowing) under 16-h-light/8-h-dark cycles with temperatures ranging from 16°C during the night to 21°C during the illuminated period.
RT-PCR Amplification
RT-PCR amplification was performed with the one-step RT-PCR kit (Qiagen, Valencia, CA). Approximately 100 mg of fresh tissue were harvested (see culture conditions described above) for total RNA extraction with the Plant RNeasy kit (Qiagen) following the manufacturer's instructions. After extraction, total RNA preparations were treated by RNase-free DNase (1 unit for 30 min at 37°C; Promega, Madison, WI) to ensure the complete removal of any contaminating gDNA into samples. Eighty to 100 ng of purified total RNA were used to complete the RT-PCR amplification. Supplemental Table I lists the different primer pairs that were used during this work. To ensure that RNA extraction was correctly performed, we have systematically amplified, as a positive control, a 315-bp fragment corresponding to 18S RNA (Quantum RNA universal 18S; Ambion, Austin, TX).
Leaf Crude Extract Preparation
Three to four leaves were harvested at the middle of the light period and kept on ice during the whole extraction procedure. Leaves were homogenized in 100 μL of cold extraction buffer (MOPS 100 mm; EDTA 1 mm; dithiothreitol [DTT] 1 mm; glycerol 10%) using a polytron blender (Tissue Tearor; BioSpec Products, Bartlesville, OK). The homogenate is centrifuged twice for 10 min at 10,000g and 4°C. The resulting supernatant is immediately used to perform zymograms (after protein assay).
Zymograms of Hydrolytic Activities
One hundred micrograms of leaf extract proteins were loaded onto a native PAGE (7.5% acrylamide) and separated for 3 h at 4°C and 15 mA cm−1. The gels were incubated overnight at room temperature in the following buffer: trisodium citrate 50 mm, pH 6, Na2HPO4 50 mm, and DTT 5 mm, except for the pullulanase zymogram, which was incubated in Tris 25 mm, glycine 192 mm, pH 8.2, DTT 1 mm. Hydrolytic activities were revealed by iodine solution (I2 0.2% [w/v] and KI 2% [w/v]) except for pullulanase gel. Depending on the tested substrate, soluble potato (Solanum tuberosum) starch (0.3%), maize (Zea mays) amylopectin (0.3%), maize β-limit dextrins (0.3%), maize phytoglycogen (0.3%), rabbit liver glycogen (0.3%), or red pullulan (0.6%) was added to the gel (final concentration of the substrate in the gel is given within brackets).
Extraction of Starch from Arabidopsis Leaves
Arabidopsis leaves were harvested at the end of the light period unless otherwise specified. Approximately 10 g of fresh material were homogenized using a polytron blender in 30 mL of the following buffer: 100 mm MOPS, pH 7.2, 5 mm EDTA, 10% (v/v) ethanediol. The homogenate was filtered through two layers of Miracloth (Calbiochem, EMD Biosciences, La Jolla, CA) and centrifuged for 15 min at 4°C and 4,000g. The pellet was resuspended in 30 mL Percoll 90% (v/v) and centrifuged for 40 min at 4°C and 10,000g. The starch pellet was washed six times with sterile distilled water (10 min at 4°C and 10,000g). Starch was finally stored at 4°C in 20% ethanol.
Determination of Starch and WSP Levels and Spectral Properties of the Iodine-Starch Complex
A full account of amyloglucosidase assays and λmax (wavelength at the maximal absorbance of the iodine polysaccharide complex) measures can be found in Delrue et al. (1992).
Separation of Starch Polysaccharides by Size Exclusion Chromatography
A total of 1.5 to 2.0 mg starch were dissolved in 500 μL of 10 mm NaOH and subsequently applied to a Sepharose CL-2B column (0.5-cm i.d.×65-cm height), which was equilibrated and eluted with 10 mm NaOH. Fractions of 300 μL were collected at a rate of 1 fraction/1.5 min. Glucans were detected by their reaction with iodine and the levels of amylopectin and amylose were determined by amyloglucosidase assays after fraction assembling.
WSP Extraction and Purification
Approximately 10 g of fresh leaves were homogenized using a polytron blender in 30 mL of the following buffer: 100 mm MOPS, pH 7.2, 5 mm EDTA, 10% (v/v) ethanediol. The homogenate was centrifuged for 10 min at 4,000g at 4°C and the resulting supernatant was immediately boiled in a water bath for 30 min and centrifuged at 10,000g for 15 min at 4°C. The supernatant was subsequently lyophilized and the resulting pellet was dissolved in 10% dimethyl sulfoxide (DMSO) before loading onto a TSK HW50F column (90-cm height, 2.2-cm i.d.). Elution is performed by DMSO 10% at a flow rate of 12 mL h−1. Two-milliliter fractions were collected and analyzed for glucan content by amyloglucosidase assay. Peak fractions were collected, debranched, and analyzed by HPAEC-PAD. Debranching and HPAEC-PAD procedures were the same as those described for amylopectin (below). Approximately 1 mg of material was used for debranching reaction by isoamylase and pullulanase.
CL Distribution of Amylopectin after Complete Debranching
FACE of debranched amylopectin: After purification on a Sepharose CL-2B column, 500 μg of amylopectin were dialyzed against distilled water and subsequently lyophilized. The amylopectin pellet was resuspended in 1 mL of 55 mm sodium acetate, pH 3.5 buffer, and incubated overnight at 42°C with 20 units of isoamylase isolated from Pseudomonas amyloderamosa (Hayashibara Biochemical Laboratories, Okayama, Japan) and 1 unit of pullulanase from Klebsiella pneumoniae (Sigma, St. Louis). Salts were subsequently removed by passage through an extract-clean carbograph column (Alltech, Deerfield, IL).
Derivatization procedure: Glucans were derivatized with 8-amino-1,3,6-pyrenetrisulfonic acid (APTS) according to the manufacturer's recommendations (Beckman Coulter, Fullerton, CA). Briefly, 2 μL APTS in 15% acetic acid solution and 2 μL of 1 m of NaBH3CN in tetrahydrofolate were added and the coupling reaction was allowed overnight at 37°C in the dark.
Capillary electrophoresis analysis: Separation and quantification of APTS-coupled glucans were carried out on a P/ACE System 5000 (Beckman Coulter) equipped with a laser-induced fluorescence system using a 4-mW argon ion laser. The excitation wavelength was 488 nm and the fluorescence emitted at 520 nm was recorded on the Beckman P/ACE station software system (version 1.0). Uncoated fused-silica capillaries of 57-cm length×75-μm i.d. were used. Running buffers were from Beckman Coulter. Samples were loaded into the capillaries by electroinjection at 10 kV for 10 s and a voltage of 30 kV was applied for 20 min at a constant temperature of 25°C.
HPAEC-PAD (Dionex, Sunnyvale, CA) of debranched polysaccharides: A complete description of this technique can be found in Fontaine et al. (1993). Polysaccharides were debranched as described above before submission to HPAEC-PAD. HPAEC-PAD analysis was performed on Dionex DX600 equipped with a CarboPac PA100 column (4-mm i.d.×250-mm length).
Flow rate: 0.8 mL min−1. After a 1-min equilibration step with 100 mm NaOH, sample elution was carried out with a linear gradient of sodium acetate (1–40 min, 0–400 mm) in 100 mm NaOH. Peak integration was performed with a C-R8A Chromatopac integrator (Shimadzu, Duisburg, Germany).
Acknowledgments
We thank Frédéric Chirat and Yves Leroy for their successful help with FACE and HPAEC-PAC analysis.
Footnotes
-
Article, publication date, and citation information can be found at www.plantphysiol.org/cgi/doi/10.1104/pp.105.059295.
-
↵1 This work was supported by Génoplante (project no. Af2001030), by the Région Nord Pas de Calais, by the Ministère Délégué à la Recherche (Action Concertée Incitative Jeunes-Chercheurs no. JC5145), by the European Community (Feder, CPER2000–2006), and by the Centre National de la Recherche Scientifique.
-
↵2 These authors contributed equally to the paper.
-
↵3 Present address: National Institute of Agricultural Biology, Hintingdon Road, Cambridge CB3 0LE, UK.
-
↵[w] The online version of this article contains Web-only data.
- Received January 26, 2005.
- Revised March 1, 2005.
- Accepted March 3, 2005.
- Published April 22, 2005.
LITERATURE CITED
