Antisense inhibition of plastidial phosphoglucomutase provides ...

27 downloads 5180 Views 328KB Size Report
Page 1. Antisense inhibition of plastidial phosphoglucomutase provides compelling evidence that potato tuber amyloplasts import carbon from the cytosol in ...
The Plant Journal (2000) 23(1), 43±53

Antisense inhibition of plastidial phosphoglucomutase provides compelling evidence that potato tuber amyloplasts import carbon from the cytosol in the form of glucose-6-phosphate Eva Tauberger¶, Alisdair R. Fernie¶,*, Michael Emmermann², Andreas Renz³, Jens Kossmann, Lothar Willmitzer and Richard N. Trethewey§ Max-Planck-Institut fuÈr Molekulare P¯anzenphysiologie, Am MuÈhlenberg 1, 14476 Golm, Germany Received 7 March 2000; revised 19 April 2000; accepted 20 April 2000. *For correspondence (fax +49 331567 8250; e-mail [email protected]). ² Present address: PlantTec Biotechnologie GmbH F and E, Hermannswerder 14, 14473 Potsdam, Germany. ³ Present address: BASF Aktiengesellschaft, Carl-Bosch-Straûe 38, 67056 Ludwigshafen, Germany. § Present address: Metanomics GmbH and Co KGaA, Tegeler Weg 33, 10589 Berlin, Germany. ¶ These authors contributed equally to the work described in this paper.

Summary The aim of this work was to establish whether plastidial phosphoglucomutase is involved in the starch biosynthetic pathway of potato tubers and thereby to determine the form in which carbon is imported into the potato amyloplast. For this purpose, we cloned the plastidial isoform of potato PGM (StpPGM), and using an antisense approach generated transgenic potato plants that exhibited decreased expression of the StpPGM gene and contained signi®cantly reduced total phosphoglucomutase activity. We con®rmed that this loss in activity was due speci®cally to a reduction in plastidial PGM activity. Potato lines with decreased activities of plastidial PGM exhibited no major changes in either whole-plant or tuber morphology. However, tubers from these lines exhibited a dramatic (up to 40%) decrease in the accumulation of starch, and signi®cant increases in the levels of sucrose and hexose phosphates. As tubers from these lines exhibited no changes in the maximal catalytic activities of other key enzymes of carbohydrate metabolism, we conclude that plastidial PGM forms part of the starch biosynthetic pathway of the potato tuber, and that glucose-6-phosphate is the major precursor taken up by amyloplasts in order to support starch synthesis. Keywords: phosphoglucomutase, glucose-6-phosphate, potato, tuber, starch, amyloplast.

Introduction The pathway by which sucrose is converted to starch in the potato tuber and the subcellular location of the enzymes involved in this pathway are well documented (Kruger, 1997; Pozueta-Romero et al., 1999; ap Rees and Morrell, 1990; Trethewey and Willmitzer, 1998). Imported sucrose is cleaved in the cytosol by sucrose synthase, resulting in the formation of UDP-glucose and fructose, the UDP-glucose then being converted to glucose-1-phosphate by UDPglucose pyrophosphorylase (UGPase). The second product of the sucrose synthase reaction, fructose, is ef®ciently phosphorylated to fructose-6-phosphate by fructokinase. Fructose-6-phosphate is freely converted to glucose-6phosphate and subsequently to glucose-1-phosphate by ã 2000 Blackwell Science Ltd

the action of phosphoglucose isomerase and phosphoglucomutase, respectively. Categorical evidence that the carbon destined for starch synthesis enters the potato amyloplast at the level of hexose monophosphates, rather than triose phosphates, was provided by determination of the degree of randomization of radiolabel in glucose units isolated from starch following incubation of potato tuber discs with glucose labelled at the C1 or C6 positions (Hatzfeld and Stitt, 1990; Viola et al., 1991). These data are in agreement with the observation that potato tubers lack plastidial fructose-1,6bisphosphatase (FBPase) activity (Entwistle and ap Rees, 1990), and the failure to ®nd expression of plastidic FBPase 43

44

Eva Tauberger et al.

in tubers (Kossmann et al., 1992). There is, however, still some debate as to whether it is glucose-1-phosphate (Kosegarten and Mengel, 1994; Naeem et al., 1997) or glucose-6-phosphate (Kammerer et al., 1998; Schott et al., 1995; Wischmann et al., 1999) that is transported into the tuber amyloplast. Within the plastid, the biosynthesis of starch proceeds via the ADP-glucose pyrophosphorylase pathway (Martin and Smith, 1995; Trethewey and Smith, 1999). Phosphoglucomutases (PGM) are present at high activities in many plant tissues, e.g. tobacco (Hanson and McHale, 1988) and Arabidopsis leaves (Caspar et al., 1985), potato tubers (Sweetlove et al., 1999), maize roots (Manjunath et al., 1998) and pea embryos (Harrison et al., 1998). They are thought to operate close to equilibrium, catalysing the interconversion of glucose-1-phosphate and glucose-6-phosphate. It therefore follows that, if glucose-6phosphate is the imported carbon substrate for potato tuber amyloplasts, then both cytosolic and plastidic isoforms of phosphoglucomutase must be involved in the sucrose to starch interconversion. Early protein puri®cation studies of potato tuber revealed the presence of only two isoforms (Takamiya and Fukui, 1978), the elution pro®le of which correspond to the cytosolic and plastidic isoforms determined following puri®cation of Nicotiana silvestris phosphoglucomutases (Hanson and McHale, 1988). This is in contrast to Arabidopsis thaliana (Caspar et al., 1985) and Zea mays (Manjunath et al., 1998), which are both thought to contain three isoforms. Plastidial phosphoglucomutase mutants of Arabidopsis thaliana (Caspar et al., 1985) and Nicotiana sylvestris (Hanson and McHale, 1988) were found to contain dramatic reductions in the level of leaf transitory starch. Furthermore, recent studies have shown that the rug3 locus of pea encodes a plastidial phosphoglucomutase, and that mutation at this locus results in a severe depletion of starch levels in pea embryos (Harrison et al., 1998). In addition, these data are supported by uptake studies performed with isolated pea embryo amyloplasts (Hill and Smith, 1991) which provide further evidence that glucose-6-phosphate is the imported substrate for starch synthesis within the pea embryo. Furthermore, uptake experiments performed on isolated cauli¯ower bud amyloplasts (Batz et al., 1993; Batz et al., 1994) indicated that glucose-6-phosphate is also the transported substrate within this tissue. The recent cloning of the glucose-6phosphate transporter from maize endosperm, pea roots, potato tubers and cauli¯ower buds (Kammerer et al., 1998) supports these ®ndings. Caution must be taken when interpreting uptake experiments due to the technical dif®culty of these studies and the fact that the isolated amyloplast may not provide an accurate representation of the in vivo system. Thus the only direct, in planta evidence of carbon uptake into the

plastid of non-photosynthetic tissues remains that provided by the studies of the rug3 mutants of pea (Harrison et al., 1998). In this paper, we present the cloning of the plastidial isoform of phosphoglucomutase from potato and describe the generation and biochemical characterization of transgenic potato lines in which the activity of this isoform of phosphoglucomutase has been modulated using the antisense approach. We conclude that plastidial phosphoglucomutase is involved in the starch synthetic pathway in potato tubers and that carbon enters the amyloplast mainly at the level of glucose-6-phosphate. Results Isolation and molecular characterization of StpPGM A full-length cDNA encoding StpPGM was isolated from a Solanum tuberosum tuber-speci®c cDNA library (Kossmann et al., 1991) by an oligonucleotide screening approach as described in Experimental procedures. Using this strategy, we isolated a 2.4 kb full-length clone. The sequence of this gene has been deposited with GenBank under the accession number AJ240053; sequence analysis revealed an open reading frame of 632 amino acids. Comparison with the functionally characterized cytosolic phosphoglucomutase of Zea mays (Manjunath et al., 1998) and the phosphoglucomutases of Escherichia coli (Lu and Kleckner, 1994) and Saccharomyces cerevisiae (Boles et al., 1994) revealed 77, 48 and 67% similarity, respectively. However, the StpPGM bears an N-terminal domain of 76 amino acids which does not align with either the Z. mays or E. coli phosphoglucomutase sequences. This domain bears an exceptionally high degree of hydroxylated amino acids (mainly serine and threonine residues) and a central sequence bearing 29 positively charged residues. These features are characteristic of plastidial transit peptides (Gavel and von Heijne, 1990; Keegstra and Olsen, 1989), indicating a plastidial location for the enzyme encoded by the StpPGM cDNA. Using the computer program ChloroP (Emanuelsson et al., 1999), we predict that the transit peptide is 57 residues in length and has a molecular mass of approximately 8 kDa, the mature protein having a mass of approximately 62 kDa. Analysis of mRNA Northern blots using the StpPGM cDNA as a probe suggests constitutive expression of this gene; the transcript was present at similar levels in young and mature leaves, stems, roots, stolons and developing and mature tubers (data not shown). In vitro chloroplast import of StpPGM In order to con®rm that StpPGM encodes a plastidtargeted protein, we performed in vitro import experiments with isolated chloroplasts. Full-length StpPGM ã Blackwell Science Ltd, The Plant Journal, (2000), 23, 43±53

Phosphoglucomutase and starch synthesis protein was synthesized by in vitro transcription and translation, and was incubated with isolated pea chloroplasts in the presence of either 50 mM or 2 mM ATP. Following this incubation, the chloroplasts were subjected to a further incubation in the presence, or absence, of thermolysin. Polypeptides taken up into the stroma were then analysed by SDS±PAGE. Figure 1 shows that the fulllength StpPGM protein was imported into the chloroplasts and processed into the mature form under both reaction conditions, suggesting that protein import is independent of the ATP concentration. In both cases, thermolysin treatment resulted in digestion of the (70 kDa) in vitro translation product but not of a smaller polypeptide (62 kDa), which therefore can be concluded to be inaccessible to the protease within the chloroplasts. The molecular mass of the mature protein is in excellent agreement with the estimated molecular mass calculated from the amino acid sequence data.

45

(Figure 2b) and to assess the phosphoglucomutase activity in these lines (Figure 2c). The phosphoglucomutase activity was determined in an assay optimized for potato tubers as described by Trethewey et al. (1998). From these initial studies, four lines were selected for further study (P-5, P-9, P-40 and P-60) which showed a signi®cant decrease in phosphoglucomutase activity, being dramatically reduced by up to 60% in mature tubers (Figure 2). In order to test whether the reduction in phosphoglucomutase activity is speci®c to the plastidial isoform, we investigated the subcellular distribution of the enzyme

Preparation and selection of transgenic antisense plants The cDNA encoding plastidial PGM was cloned in the antisense orientation into the transformation vector pBinAR-Kan between the CaMV 35S promoter and the ocs terminator (Figure 2a; Liu et al., 1990). Wild-type potato leaves were transformed with this construct using an Agrobacterium-mediated protocol (Rocha-Sosa et al., 1989), and 80 transgenics were transferred to the greenhouse for screening. Potato plants were grown in 2 litre pots and leaf samples were taken for Northern analysis (data not shown). From the initial 80 lines, nine were selected that showed a reduction in the StpPGM transcript level with respect to the wild-type. These lines were ampli®ed and six replicates from each line were grown in 2 litre pots in the greenhouse to con®rm the decrease in transcription in leaves (data not shown) and tubers

Figure 1. Import of StpPGM into pea chloroplasts. Autoradiographs of 10% SDS±polyacrylamide gels are shown. Lanes 1 and 6 show the in vitro translation product, lanes 2±5 show the same translation product including pea chloroplasts (20 mg chlorophyll per 9 ml of translation reaction). Lanes 3 and 5: thermolysin to a ®nal concentration of 0.18 mg ml±1 has been added. Lanes 2 and 3 were supplemented with ATP at a ®nal concentration of 50 mM. Lanes 4 and 5 were supplemented with ATP at a ®nal concentration of 2 mM.

ã Blackwell Science Ltd, The Plant Journal, (2000), 23, 43±53

Figure 2. Antisense inhibition of plastidial PGM in potato tubers. (a) Construction of a chimeric gene for expresssion of plastidial phosphoglucomutase antisense RNA: (i) a 540 bp fragment encoding the CaMV 35S promoter, (ii) a 2400 bp Asp718 fragment encoding the StpPGM cDNA in the antisense orientation, (iii) polyadenylation signal of the octopine synthase gene, vector pBinAR (Liu et al., 1990). (b) Northern blot analysis of transgenic plants with altered expression of plastidial phosphoglucomutase. RNA was extracted from developing tubers of greenhouse-grown plants. The ®lter was hybridized with the full-length StpPGM cDNA. (c) Phosphoglucomutase activities of tubers of transgenic plants with altered expression of plastidial phosphoglucomutase. Tuber samples were taken from 14-week-old plants. Data are presented as the mean 6 SE of determinations on six individual plants per line.

46

Eva Tauberger et al.

(Figure 3). When isolated chloroplasts were applied to a DE-52 column, phosphoglucomutase eluted as a single major peak. When whole-leaf extracts were applied, approximately 45% of the phosphoglucomutase activity eluted in the second peak, corresponding to the plastidic isoform. These data are in contrast to the subcellular location of phosphoglucomutase in tobacco, in which only 23±29% of the activity is located in the plastid (Hanson and McHale, 1988). When we applied developing wild-type tuber extracts to the column, we found that the cytosolic isoform accounts for 55.3 6 1.7% (mean 6 SE, n = 4) and the plastidial isoform the remaining 44.7 6 1.7% (mean 6 SE, n = 4) of the total phosphoglucomutase activity (a representative example of the elution pro®le is given in Figure 4a). We then applied extracts from each of the transgenic lines. For this purpose, the amount of extract applied was set so that the total amount of phosphoglucomutase activity applied

Figure 3. Separation of phosphoglucomutase activities from potato leaf. Extracts were fractionated on a DE-52 column by elution with a KCl gradient. Each fraction was assayed for phosphoglucomutase activity. Enzyme activity is expressed as nmol min±1 fraction±1 (left axis), and the applied salt gradient is indicated (right axis). (a) Whole-leaf extract from wild-type potato plants; (b) isolated chloroplasts from wild-type leaves.

Figure 4. Separation of phosphoglucomutase activities from potato tuber tissue. Potato tuber extracts were fractionated on a DE-52 column by elution with KCl gradient. Each fraction was assayed for phosphoglucomutase activity. Enzyme activity is expressed as nmol min±1 fraction±1 (left axis), and the applied salt gradient is indicated (right axis). (a) Wild-type; (b) line P-40; (c) line P-9; (d) line P-5; (e) line P-60.

ã Blackwell Science Ltd, The Plant Journal, (2000), 23, 43±53

Phosphoglucomutase and starch synthesis was proportional to the total phosphoglucomutase activity of the developing tuber samples (data not shown). In the transgenic lines, the elution pro®le of the ®rst peak (cytosolic isoform) seemed to be relatively unchanged, whereas there was a striking reduction in the second peak (plastidial isoform). The activity in fractions from the second peak decreased by 29, 69, 75 and 83% in P-40, P60, P-9 and P-5, respectively, whereas the activity in the ®rst peak remained unaltered, despite minor changes in the elution pro®le (representative examples of the pro®les are shown in Figure 4b±e). These data show that inhibition of phosphoglucomutase activity occurs exclusively in the plastidic compartment. Phenotypic analysis Having con®rmed the decrease in plastidic phosphoglucomutase activity, the selected antisense lines were grown in 5 litre pots in the greenhouse alongside wildtype controls, and tuber yield and plant morphology were determined. As it is our experience that yield trials are subject to considerable variation, we grew at least 13 plants per line to ensure that a reliable data set was obtained. There was no difference in the total tuber yield in the transgenic lines (Table 1) or in the morphology of either the plants or the tubers. In addition, no trends could be observed in the number of tubers per plant or in the mean tuber fresh weight with respect to the residual phosphoglucomutase activ-

47

ity. We determined the density of the tubers in order to gain an accurate and direct indication of the starch content. Interestingly, we found that repression of plastial phosphoglucomutase led to a tendential decrease in tuber density, signi®cant in lines P-5, P-9 and P-60. Line P-40 shows no change in tuber density; however, it exhibits the smallest change in phosphoglucomutase activity. Determination of carbohydrate content of the transgenic lines In order to con®rm the results from the density trials, we determined the starch content of the tubers harvested from senescing plants by direct enzymatic measurement (Table 2). A signi®cant and dramatic reduction in starch content was observed in lines P-5, P-9 and P-60; the starch content in line P-60 was reduced by 39% with respect to wild-type. When taken in combination with the large-scale density measurements, these data provide strong evidence that the starch content is directly related to the activity of phosphoglucomutase. We also determined the sucrose and glucose contents of the transgenic tubers. Most interesting were the changes in sucrose, which signi®cantly increased in lines P-5, P-9 and P-60, and given the reduced starch content the starch/sucrose ratio in these lines was also dramatically altered. The glucose content was strikingly and signi®cantly increased in

Table 1. Yield, density, mean tuber size and number of the antisense StpPGM transgenic lines Parameter Tuber number Tuber yield Mean tuber fresh weight (g) Density

Wild-type 26.1 6 1.8 697 6 23 26.7 1.078

P-40 21.9 6 1.7 704 6 17 32.1 1.079

P-60 26.5 6 1.7 694 6 26 26.2 1.058

P-5 27.5 6 2.1 666 6 26 24.2 1.056

P-9 19.5 6 1.4 673 6 25 34.5 1.063

Potato plants were grown in the greenhouse in 5 litre pots. Transgenic tuber yield (total tuber fresh weight), tuber number and density determinations for mature tubers from fully senescent plants were performed with 13±15 plants per line harvested in the Summer. Values are means 6 SE. Table 2. Carbohydrate content (mmol g±1 fresh weight) of the antisense StpPGM transgenic tubers

Starch Sucrose Glucose Starch/Sucrose

Wild-type

P-40

P-60

P-5

P-9

657 6 65 18.5 6 2.6 1.7 6 0.2 45.4 6 8.4

691 6 20 20.1 6 2.1 0.4 6 0.1* 35.0 6 6.3

406 6 44* 36.1 6 3.5* 9.4 6 1.6* 8.2 6 2.4*

515 6 42* 34.2 6 1.7* 7.3 6 1.0* 7.9 6 2.7*

491 6 73* 27.3 6 0.3* 9.3 6 1.6* 15.4 6 1.7*

Potato plants were grown in the greenhouse in 3.5 litre pots. Developing tubers were harvested in the spring after 10 weeks of growth, and the sucrose, glucose and starch content were determined. Data are presented in mmol g±1 FW (mmol glucose g±1 fresh weight, in the case of starch), and represent the mean 6 SE of determinations on six individual plants per line. *Values that were determined by the t test to be signi®cantly different from the wild-type. ã Blackwell Science Ltd, The Plant Journal, (2000), 23, 43±53

48

Eva Tauberger et al.

lines P-5, P-9 and P-60, perhaps as a result of the elevated sucrose content. When assessed in combination with the normal phenotype and normal growth rate, the increased sucrose concentration implies that the observed changes are tuber-speci®c effects.

key enzymes of glycolysis (phosphoglucose isomerase (PGI), phosphofructokinase (PFK), pyruvate kinase (PK), enolase and aldolase) or ADP-glucose pyrophosphorylase (Table 3). Determination of hexose phosphates

Measurement of the activities of key enzymes of carbohydrate metabolism In order to verify that changes in the starch content were not due to changes in the activities of enzymes other that phosphoglucomutase, we measured the activities of a range of enzymes in the wild-type and transgenic lines. The activities were measured using assays speci®cally optimized for potato tubers and using an extraction procedure that has previously been shown to be reliable (Trethewey et al., 1998). Determination of the maximum catalytic activities revealed no signi®cant changes in either cytosolic enzymes of the sucrose±starch pathway (sucrose synthase (SuSy) or UDP-glucose pyrophosphorylase (UGPase)),

The glucose-1-phosphate (G1P), glucose-6-phosphate (G6P) and fructose-6-phosphate (F6P) contents of developing tubers are shown in Table 4. These values have low standard errors, and wild-type values are comparable to those previously reported for potato tubers (Jelitto et al., 1992; Trethewey et al., 1998). In the lines exhibiting the strongest reduction in steady-state StpPGM mRNA levels and total PGM activity, there is a moderate, but signi®cant, increase in the level of all three hexose phosphates. Despite the increased hexose phosphate pool sizes, there is no change in the ratios of the individual hexose phosphates. Similar changes were found in mature tubers from senescent plants (data not shown).

Table 3. Enzyme activities (nmol min±1 g±1 fresh weight) in the antisense StpPGM transgenic tubers

AGPase SuSy UGPase PGI PFK PK Enolase Aldolase

Wild-type

P-40

P-60

P-5

P-9

100 6 8 948 6 85 7641 6 341 1225 6 149 117 6 11 409 6 43 405 6 34 461 6 63

91 6 10 947 6 140 7884 6 298 1123 6 170 128 6 15 414 6 56 428 6 29 408 6 53

108 6 15 1025 6 59 7282 6 349 1298 6 130 105 6 7 412 6 51 430 6 61 514 6 67

104 6 17 1084 6 64 7494 6 396 1285 6 155 114 6 15 457 6 48 437 6 43 451 6 24

104 6 5 1079 6 57 7513 6 369 1165 6 87 101 6 3 400 6 39 399 6 51 423 6 70

Enzyme activities were determined in samples from developing tubers. The data presented are the means 6 SE of determinations on six individual plants per line. AGPase, ADP-glucose pyrophosphorylase; SuSy, sucrose synthase; UGPase, UDP-glucose pyrophosphorylase; PGI, phosphoglucose isomerase; PFK, ATP-dependent phosphofructokinase; PK, pyruvate kinase. Table 4. Metabolites (nmol g±1 fresh weight) of the glycolytic sequence in the antisense StpPGM transgenic tubers Wild-type

P-40

P-60

P-5

P-9

G6P G1P F6P 3-PGA PEP Pyr

120 6 6 12 6 1 39 6 3 73 6 5 22 6 2 10 6 1

125 6 12 11 6 1 41 6 4 78 6 7 21 6 2 10 6 1

168 6 11* 16 6 2* 52 6 5* 77 6 7 20 6 2 11 6 2

160 6 10* 15 6 1* 47 6 3* 80 6 6 24 6 1 13 6 2

163 6 17* 14 6 2 51 6 4* 82 6 4 23 6 2 11 6 2

Ratios G6P/G1P G6P/F6P

10.1 6 1.2 3.1 6 0.2

11.3 6 1.5 3.2 6 0.4

11.2 6 1.2 3.3 6 0.3

11.3 6 1.6 3.5 6 0.3

12.7 6 1.5 3.2 6 0.3

Metabolites were determined in the same samples from developing tubers used for the enzyme analysis presented in Table 3. The data are presented as the means 6 SE of determinations on six individual plants per line. G6P, glucose-6-phosphate; G1P, glucose-1phosphate; F6P, fructose-6-phosphate; UDPG, UDP-glucose; PEP, phosphoenolpyruvate; Pyr, pyruvate. *Values that were determined by the t test to be signi®cantly different from the wild-type. ã Blackwell Science Ltd, The Plant Journal, (2000), 23, 43±53

Phosphoglucomutase and starch synthesis

49

Discussion

StpPGM transcript levels in all lines. Our values for phosphoglucomutase activity in the wild-type were comparable with those previously documented (e.g. Sweetlove et al., 1996). In the transgenic lines expressing the StpPGM cDNA in the antisense orientation, a signi®cant reduction in the total phosphoglucomutase activity was acheived. Studies on the subcellular location of phosphoglucomutase activity showed that the decrease in phosphoglucomutase activity occurred predominantly, if not exclusively, in the plastid. In addition, we determined the maximal catalytic activities of a range of enzymes of carbohydrate metabolism which were found to be unaltered in the transgenic lines. Taking all this evidence into account, we argue that lines P-5, P-9, P-40 and P-60 are valid for further study.

Isolation and characterization of a cDNA encoding the potato plastidial phosphoglucomutase

Effect of a reduction in plastidial phosphoglucomutase activity on starch metabolism

A full-length 2.4 kb clone was obtained by screening a potato tuber cDNA library with radiolabelled oligonucleotides designed following analysis of consensus regions in a range of different phosphoglucomutase sequences. The StpPGM cDNA encodes a protein of molecular mass of 70 kDa, and exhibits homology over large parts of the middle and C-terminal domains of the protein to sequences from plants, yeast and bacterial sources (Boles et al., 1994; Lu and Kleckner, 1994; Manjunath et al., 1998). The N-terminal domain, however, does not align to any known phosphoglucomutase sequence and bears features characteristic of a transit peptide (Gavel and von Heijne, 1990). In order to verify that this protein was indeed targeted to the plastid, we performed import experiments in which in vitro transcription and translation products of the StpPGM cDNA were incubated with isolated pea chloroplasts. These experiments revealed that StpPGM was imported into the plastid and that the transit peptide was cleaved resulting in the formation of a 62 kDa mature peptide. The sequence homology of StpPGM cDNA in comparison to that from Spinacea oleracea is only 19%, which is surprisingly low (Penger et al., 1994). However, in the latter case, no functional evidence that the protein encoded by this cDNA is a phosphoglucomutase was presented. Therefore, this study represents the ®rst functional characterization of a plastidic phosphoglucomutase in plants.

Antisense repression of plastial phosphoglucomutase activity leads to a reduction in starch content of up to 40% on a fresh weight basis. However, the reduction in starch content is not as severe as that observed in the leaves of Arabidopsis (Caspar et al., 1985) and tobacco (Hanson and McHale, 1988) plastidial phosphoglucomutase mutants which were essentially starchless. Furthermore, the level of starch in these lines is not reduced as dramatically as in potato lines in which the activity of ADP-glucose pyrophosphorylase is inhibited (these lines exhibit reductions in starch content of up to 96% when the enzyme activity is similarly reduced; MuÈllerRoÈber et al., 1992). These data strongly suggest that the conversion of glucose-6-phosphate to glucose-1-phosphate within the plastid is involved in the starch biosynthetic pathway, and therefore that the import of glucose-6phosphate is necessary for normal rates of starch synthesis. However, there are two different possibile explanations for the fact that the restriction in starch synthesis was not greater in the antisense lines. It is possible that either up to 60% of the starch biosynthetic ¯ux is occurring via a phosphoglucomutase-independent pathway (involving the uptake of glucose-1-phosphate) or the residual plastidial PGM activity is capable of providing ADP-glucose pyrophosphorylase with enough substrate to support starch synthesis. Given that the reduction of plastidial PGM activity is only partial, and that the activity of AGPase is considerably lower than that of PGM, it is conceivable that the remaining PGM activity can support the residual rate of starch synthesis. Furthermore, as enzyme activity and pathway ¯ux are commonly logarithmically related (exempli®ed by recent studies on the contribution of AGPase to the control of starch synthesis: Geigenberger et al., 1999; Sweetlove et al., 1999), it is perhaps not completely

Determination of glycolytic intermediates We investigated the levels of intermediates of the glycolytic pathway in order to determine whether this pathway was affected by the antisense repression of the plastidial PGM (Table 3). The levels of 3-phosphoglycerate (3-PGA), phosphoenolpyruvate (PEP) and pyruvate were comparable to those previously determined in potato tubers (Burrell et al., 1994; Jelitto et al., 1992; Trethewey et al., 1998). The measurements were made on the same extracts that were used for the hexose phosphate determinations. There was a slight increase in the level of 3-PGA; however, this was only signi®cant in the case of line P-9.

Antisense repression of StpPGM in potato plants We generated transgenic potato plants transformed with a construct containing the StpPGM cDNA in the antisense orientation in order to evaluate the importance of plastidial phosphoglucomutase in the starch biosynthetic pathway. Northern analysis revealed a large decrease in steady-state ã Blackwell Science Ltd, The Plant Journal, (2000), 23, 43±53

50

Eva Tauberger et al.

surprising that an 81% reduction in plastidial PGM activity results in only a 40% reduction in starch content. Therefore, whilst we cannot exclude the possibility that the import of glucose-1-phosphate makes some contribution in the ¯ux to starch, we believe that the residual rate of starch synthesis can essentially be explained by the residual PGM activity within the amyloplasts of the transgenic lines. Thus we argue that these data provide the ®rst in vivo evidence that carbon enters the potato tuber amyloplast in the form of glucose-6-phosphate. Furthermore, they are in agreement with studies on isolated potato amyloplasts (Schott et al., 1995; Wischmann et al., 1999) and are supported by the recent cloning of a glucose-6-phosphate/phosphate transporter (Kammerer et al., 1998). Taken in combination, these ®ndings provide compelling evidence that glucose-6-phosphate is the main precursor for starch synthesis within the potato tissue. Inhibition of plastidial phosphoglucomutase activity leads to an increase in sucrose levels Concomitant with the decrease in starch content in the transgenic lines, we observed an increase in sucrose levels. The same pattern was observed when the starch biosynthetic pathway was inhibited by antisense repression of ADP-glucose pyrophosphorylase under the control of the CaMV 35S promoter (Geigenberger et al., 1999; MuÈller-RoÈber et al., 1992). Detailed biochemical studies revealed that plants exhibiting reduced ADPglucose pyrophosphorylase activities were characterized by an increase in the activation state of sucrose phosphate synthase (SPS) coupled to an increased rate of sucrose re-synthesis. In addition, the level of glucose was observed to increase both in the antisense plastidial phosphoglucomutase plants, reported here, and the antisense ADP-glucose pyrophosphorylase plants (Geigenberger et al., 1999; MuÈller-RoÈber et al., 1992). Effect of a reduction in the activity of plastidial phosphoglucomutase on the levels of glycolytic intermediates Investigation of the levels of glycolytic intermediates revealed an increase in the level of all three hexose phosphates. Given the nature of the genetic manipulation applied this was a surprising result as we expected an imbalance in the glucose-6-phosphate to glucose-1-phosphate ratio in the transgenic lines. That the ratio remains unchanged in the transgenic lines is intriguing; one possible explanation is that the level of glucose-1-phosphate in the cytosol may be vastly in excess of that in the plastid.

In leaves, SPS is activated by dephosphorylation and SPS kinase is known to be inhibited by hexose phosphates (McMichael et al., 1995), and there is evidence that potato tuber SPS is allosterically activated by glucose-6-phosphate (Reimholz et al., 1994). Due to the increased levels of all three hexose phosphates, any of these mechanisms could explain the increased sucrose content observed in the transgenic lines. Further down the glycolytic sequence there was little change in the levels of 3PGA, PEP or pyruvate (again similar to the situation observed in plants exhibiting reduced activity of ADP-glucose pyrophosphorylase: Geigenberger et al., 1999; Trethewey et al., 1999). In summary, the metabolic consequences of antisense repression of plastidial phosphoglucomutase are essentially the same as those following repression of ADPglucose pyrophosphorylase. This provides evidence for our conclusion that plastidial phosphoglucomutase and ADP-glucose pyrophosphorylase catalyse sequential reactions in the biosynthesis of starch in potato tuber amyloplasts. Experimental procedures Plant material Solanum tuberosum L. cv. Desiree was obtained from Saatzucht Lange AG (Bad Schwartau, Germany). Plants were maintained in tissue culture with a 16 h light, 8 h dark regime on MS medium (Murashige and Skoog, 1962) which contained 2% sucrose. In the greenhouse, plants were grown under the same light regime with a minimum of 250 mmol photons m±2 sec±1 at 22°C. In this paper, the term `developing tubers' is used for tubers (over 10 g fresh weight) harvested from healthy 10-week-old plants; the term `mature tubers' refers to tubers harvested from senescent plants.

Chemicals The starch determination kit, the biochemical enzymes, and all the enzymes used for the modi®cation and restriction of DNA were obtained from Boehringer Mannheim (Mannheim, Germany). All other chemicals were purchased from Sigma or Merck (Darmstadt, Germany).

Cloning of plastidial phosphoglucomutase A potato cDNA library (Kossmann et al., 1991) was screened with radiolabelled (Multiprime labelling system, Amersham Buchler, Braunschweig, Germany) oligonucleotide probes designed with respect to conserved phosphoglucomutase sequences deduced from a range of phosphoglucomutase sequences available from GenBank. Standard procedures were performed according to Sambrook et al. (1989). The following oligonucleotide probes were used:

5¢-ATCATRTTICKRTCICCRTCICCRTC-3¢ 5¢-AARAAYTTCCAICCIGTIGG-3¢ 5¢-GTICCRAAISWYTCYTCICCRCA-3¢

Positive plaques were selected, and puri®ed lZAP II clones were excised in vivo prior to DNA sequencing. A clone with a fullã Blackwell Science Ltd, The Plant Journal, (2000), 23, 43±53

Phosphoglucomutase and starch synthesis length plastidial phosphoglucomutase was then subcloned into the plasmid pBluescript SK.

Import experiments

51

pyrophosphosphorylase which were assayed as described by Sweetlove et al. (1996) and UDP-glucose pyrophosphorylase which was assayed by the protocol described by Sweetlove et al. (1999).

In vitro transcription and translation of plasmid pBluescript SK containing the phosphoglucomutase cDNA was carried out using the TNT-coupled reticulocyte lysate system (Promega, Madison, WI, USA), containing RNAase inhibitor (RNasin; Stratagene, La Jolla, CA, USA), according to the manufacturer's instructions. The reaction was incubated at 30°C for 1 h in the presence of 120 mCi 35 S-methionine (speci®c activity 1000 Ci mmol±1, Amersham Buchler) and kept at ±80°C until use. Intact chloroplasts were isolated from the leaves of pea (Pisum sativum L.) seedlings which had been deprived of light for 12 h. Chloroplast isolation was performed using the method of KloÈsgen et al. (1989). The chlorophyll concentration of the resultant chloroplast suspension was determined according to the method of Arnon (1949). Experiments on chloroplast import were carried out by incubating chloroplasts (20 mg chlorophyll) and 9 ml in vitro translation product (in a total reaction volume of 50 ml) in the reaction media detailed by Waegemann and Soll (1991). Either 50 mM or 2 mM ATP was added to the reaction and both conditions were assayed in the presence and absence of thermolysin. Following the import reaction the chloroplasts were thoroughly washed. Stromal fractions from the import experiment were denatured by boiling for 5 min in SDS sample buffer, and separated on a 10% (w/v) SDS±polyacrylamide gel using the buffer system of Laemmli (1970). The radioactively labelled protein was visualized by standard autoradiography.

The protein puri®cation procedure was based on that described by Takamiya and Fukui (1978). Developing tubers were harvested from non-senescent plants grown in the greenhouse during the spring season. Peeled and cut potato tuber tissue (4 g) was ®nely ground in a pre-cooled mortar containing liquid N2. The powder was then immediately mixed in 20 ml of extraction buffer (20 mM Tris±HCl, pH 7.5, 1 mM MgCl2, 1 mM DTT and 0.2 mM EDTA). The homogenate was ®ltered, centrifuged, re®ltered, and then applied to a DE-52 cellulose column (18 cm long, 1.6 cm in diameter). The column was washed with buffer (20 mM Tris±HCl, pH 7.5, 1 mM MgCl2, 1 mM DTT and 0.2 mM EDTA) and eluted on a linear gradient of 0±1 M KCl at a ¯ow rate of 1.0 ml min±1; 100 1 ml fractions were collected. A 20 ml aliquot of each fraction was checked for phosphoglucomutase activity as described below. Fractions containing phosphoglucomutase activity were desalted by passage through PD-10 columns (Pharmacia) and the activity was determined as before. Puri®cation from leaf extracts and from isolated chloroplasts prepared from approximately 5 g leaf tissue was carried out using exactly the same procedure as described above for tuber material. Isolated chloroplasts were prepared essentially as described by Haake et al. (1998).

Preparation of transgenic antisense lines

The t tests were performed using the algorithm embedded into Microsoft Excel (Microsoft Corporation, Seattle, WA, USA). The term signi®cant is used in the text only when the change in question was con®rmed to be signi®cant (P < 0.05) using the t test.

A 2.4 kb Asp718/XbaI fragment of the StpPGM cDNA was cloned into the vector pBinAR-Kan (Liu et al., 1990) between the CaMV 35S promoter and the ocs terminator. This construct was introduced into potato by an Agrobacterium-mediated transformation protocol (Rocha-Sosa et al., 1989). Transgenic plants were selected on kanamycin-containing medium (Dietz et al., 1995). Initial screening of around 80 lines was performed by determining the StpPGM transcipt level in the leaves of plants grown in 2 litre pots under greenhouse conditions. A second screen was then performed at the transcript and enzyme activity levels using tubers and leaves from six plants per line for the nine pre-selected lines in the greenhouse.

Extraction of RNA and Northern blot experiments Total RNA was isolated from 2 g fresh weight of tuber tissue as described by Logemann et al. (1987). Standard conditions were used for the transfer of RNA to membranes and for the subsequent hybridization (Sambrook et al., 1989).

Biochemical analysis Starch, sugars and metabolites were determined exactly as described by Trethewey et al. (1998). Recoveries of metabolites in the trichloroacetic acid extracts have been documented previously (e.g. Trethewey et al., 1998; Veramendi et al., 1999). Enzyme activities were determined as detailed by Trethewey et al. (1998) with the exceptions of sucrose synthase and ADP-glucose ã Blackwell Science Ltd, The Plant Journal, (2000), 23, 43±53

Partial puri®cation of PGM from potato tubers

Statistical analysis

Acknowledgements We are indebted to to Olaf Woiwode, Frank Huhn and Bruno Marty for excellent supervision of the plants. We are also grateful to Romy Ackermann for performing the potato transformation. A.F. and R.T. thank the Max Planck Gesellschaft for fellowships.

References Arnon, M.I. (1949) Copper enzymes in isolated chloroplasts. Phenoloxidase in Beta vulgaris. Plant Physiol. 24, 1±15. Batz, O., Scheibe, R. and Neuhaus, H.E. (1993) Identi®cation of the putative hexose-phosphate translocator of amyloplasts from cauli¯ower buds. Biochem. J. 294, 15±17. Batz, O., Maass, U., Henrichs, G. and Scheibe, R. and Neuhaus, H.E. (1994) Glucose- and ADPGlc-dependent starch synthesis in isolated cauli¯ower-bud amyloplasts. Analysis of the interaction of various potential precursors. Biochim. Biophys. Acta, 1200, 148±154. Boles, E., Liebetrau, W., Hofmann, M. and Zimmermann, F.K. (1994) A family of hexosephosphate mutases in Saccharomyces cerevisiae. Eur. J.Biochem. 220, 83±96. Burrell, M.M., Mooney, P.J., Blundy, M., Carter, D., Wilson, F., Green, J., Blundy, K.S. and ap Rees, T. (1994) Genetic manipulation of 6-phosphofructokinase in potato tubers. Planta, 194, 95±101.

52

Eva Tauberger et al.

Caspar, T., Huber, S.C. and Somerville, C. (1985) Alterations in growth, photosynthesis, and respiration in a starchless mutant of Arabidopsis thaliana (L.) de®cient in chloroplast phosphoglucomutase activity. Plant Physiol. 79, 11±17. Dietze, J., Blau, A. and Willmitzer, L. (1995) Agrobacteriummediated transformation of potato (Solanum tuberosum). In Gene Transfer to Plants XXII (Potrykus, I. and Spangenberg, G., eds). Berlin: Springer Verlag, pp. 24±29. Emanuelsson, O., Nielsen, H. and von Heijne, G. (1999) ChloroP, a neural network-based method for predicting chloroplast transit peptides and their cleavage sites. Protein Sci. 8, 978±984. Entwistle, G. and ap Rees, T. (1990) Lack of fructose-1,6bisphosphatase in a range of higher plants that store starch. Biochem. J. 271, 467±472. Gavel, Y. and von Heijne, G. (1990) A conserved cleavage-site motif in chloroplast transit peptides. FEBS Lett. 261, 455±458. Geigenberger, P., MuÈller-RoÈber, B. and Stitt, M. (1999) Contribution of adenosine 5¢-diphosphoglucose pyrophosphorylase in response to short-term water de®cit. Planta, 209, 502±528. Haake, V., Zrenner, R., Sonnewald, U. and Stitt, M. (1998) A moderate decrease of plastid aldolase activity inhibits photosynthesis, alters the levels of sugars and starch, and inhibits growth of potato plants. Plant J. 14, 147±157. Hanson, K.R. and McHale, N.A. (1988) A starchless mutant of Nicotiana silvestris containing a modi®ed plastid phosphoglucomutase. Plant Physiol. 88, 838±844. Harrison, C.J., Hedley, C.L. and Wang, T.L. (1998) Evidence that the rug3 locus of pea (Pisum sativum L.) encodes plastidial phosphoglucomutase con®rms that the imported substrate for starch synthesis in pea amyloplasts is glucose-6-phosphate. Plant J. 13, 753±762. Hatzfeld, W.-D. and Stitt, M. (1990) A study of the rate of recycling of triose phosphates in heterotrophic Chenopodium rubrum cells, potato tubers, and maize endosperm. Planta, 180, 198± 204. Hill, L.M. and Smith, A.M. (1991) Evidence that glucose-6phosphate is imported as the substrate for starch synthesis by the plastids of developing pea embryos. Planta, 185, 91±96. Jelitto, T., Sonnewald, U., Willmitzer, L., Hajirezaei, M. and Stitt, M. (1992) Inorganic pyrophosphate content and metabolites in leaves and tubers of potato and tobacco plants expressing E. coli pyrophosphatase in their cytosol: biochemical evidence that sucrose metabolism has been manipulated. Planta, 188, 238±244. Kammerer, B., Fischer, K., Hilpert, B., Schubert, S., Gutensohn, M., Weber, A. and and. Flugge, U.-I. (1998) Molecular characterisation of a carbon transporter in plastids from heterotrophic tissues: the glucose 6-phosphate/phosphate antiporter. Plant Cell, 10, 105±117. Keegstra, K. and Olsen, L.J. (1989) Chloroplast precursors and their transport across the envelope membranes. Annu. Rev. KloÈsgen, R.B., Saedler, H. and Weil, J.-H. (1989) The amyloplasttargeting transit peptide of the waxy protein of maize also mediates protein transport in vitro into chloroplasts. Mol. Gen. Genet. 217, 155±161. Kosegarten, H. and Mengel, K. (1994) Evidence for a glucose-1phosphate translocator in storage tissue amyloplasts of potato (Solanum tuberosum) suspension-cultured cells. Physiol. Plant. 91, 111±120. Kossmann, J., Visser, R.G.F., MuÈller-RoÈber, B., Willmitzer, L. and Sonnewald, U. (1991) Cloning and expression analysis of a potato cDNA that encodes branching enzyme: evidence for co-

expression of starch biosynthetic genes. Mol. Gen. Genet. 230, 39±44. Kossmann, J., MuÈller-RoÈber, B., Dyer, T.A., Raines, C.A., Sonnewald, U. and and. Willmitzer, L. (1992) Cloning and expression analysis of the plastidic fructose-1,6bisphosphatase coding sequences: circumstantial evidence for the transport of hexoses into chloroplasts. Planta, 188, 7±12. Kruger, N.J. (1997) Carbohydrate synthesis and degradation. In Plant Metabolism (Dennis, D.T., Turpin, D.H., Lefebvre, D.D., Layzell, D.B., eds). Harlow: Longman, pp. 83±104. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680± 692. Liu, X.J., Prat, S., Willmitzer, L. and Frommer, W.B. (1990) Cis regulatory elements directing tuber-speci®c and sucroseinducible expression of a chimeric class I patatin promoter± GUS-gene fusion. Mol. Gen. Genet. 223, 401±406. Logemann, J., Schell, J. and Willmitzer, L. (1987) Improved method for the isolation of RNA from plant tissues. Ann. Biochem. 163, 16±20. Lu, M. and Kleckner, N. (1994) Molecular cloning and characterisation of the pgm gene encoding phosphoglucomutase of Escherichia coli. J. Bacteriol. 176, 5847±5851. McMichael, R.W., Bachmann, M. and Huber, S.C. (1995) Spinach leaf sucrose-phosphate synthase and nitrate reductase are phosphorylated/inactivated by multiple protein kinases in vitro. Plant Physiol. 108, 1077±1082. Manjunath, S., Kenneth, L.C.H., Winkle, P.V. and Bailey-Serres, J. (1998) Molecular and biochemical characrterisation of cytosolic phosphoglucomutase in maize. Expression during development and in response to oxygen deprivation. Plant Physiol. 117, 997±1006. Martin, C. and Smith, A.M. (1995) Starch biosynthesis. Plant Cell, 7, 971±985. Murashige, T. and Skoog, F. (1962) A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol. Plant. 15, 473±497. MuÈller-RoÈber, B., Sonnewald, U. and Willmitzer, L. (1992) Antisense inhibition of the ADPglucose pyrophosphorylase in transgenic potato leads to sugar-storing tubers and in¯uences tuber formation and expression of tuber storage protein genes. EMBO J. 11, 1229±1238. Naeem, M., Tetlow, I.J. and Emes, M.J. (1997) Starch synthesis in amyloplasts puri®ed from developing potato tubers. Plant J. 11, 1095±1103. Penger, A., Pelzer-Reith, B. and Schnarrenberger, C. (1994) cDNA sequence for the plastidic phosphoglucomutase from Spinacea oleracea (L.). Plant Physiol. 105, 1439±1440. Pozueta-Romero, J., Perata, P. and Akazawa, T. (1999) Sucrose± starch conversion in heterotrophic tissues. Crit Rev. Plant Sci. 18, 489±525. ap Rees, T. and Morrell, S. (1990) Carbohydrate metabolism in developing potatoes. Am. Potato J. 67, 835±847. Reimholz, R., Geigenberger, P. and Stitt, M. (1994) Sucrosephosphate synthase is regulated via metabolites and protein phosphorylation in potato tubers, in a manner analogous to the enzyme in leaves. Planta, 192, 480±488. Rocha-Sosa, M., Sonnewald, U., Frommer, W., Stratmann, M., Schell, J. and Willmitzer, L. (1989) Both developmental and metabolic signals activate the promoter of the class I patatin gene. EMBO J. 8, 23±29. Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular ã Blackwell Science Ltd, The Plant Journal, (2000), 23, 43±53

Phosphoglucomutase and starch synthesis Cloning. A Laboratory Manual. Cold Spring Harbor: Cold Spring Harbor Laboratory Press. Schott, K., Borchert, S., MuÈller-RoÈber, B. and Heldt, H.W. (1995) Transport of inorganic phosphate and C3- and C6-sugar phosphates across the envelope membranes of potato tuber amyloplasts. Planta, 196, 647±652. Sweetlove, L.J., Burrell, M.M. and ap Rees, T. (1996) Characterisation of transgenic potato (Solanum tuberosum) tubers with increased ADPglucose pyrophosphorylase. Biochem. J. 320, 487±492. Sweetlove, L.J., MuÈller-RoÈber, B., Willmitzer, L. and Hill, S.A. (1999) The contribution of adenosine 5¢-diphosphoglucose pyrophosphorylase to the control of starch synthesis in potato tubers. Planta, 209, 330±337. Takamiya, S. and Fukui, T. (1978) Puri®cation and multiple forms of phosphoglucomutase from potato tubers. Plant Cell Physiol. 19, 759±768. Trethewey, R.N. and Smith, A.M. (2000) Transitory starch degradation. In Photosynthesis: Physiology and Metabolism (Leegood, R.C., Sharkey, T.D. and von Caemmerer, S., eds). Dordrecht: Kluwer Academic Publishers (in press). Trethewey, R.N. and Willmitzer, L. (1998) The application of transgenic technology to the study of sink metabolism in potato. In Regulation of Primary Metabolic Pathways in Plants (Kruger, N.J., Hill, S.A. and Ratcliffe, R.G., eds). Dordrecht: Kluwer Academic Publishers, pp. 239±256.

ã Blackwell Science Ltd, The Plant Journal, (2000), 23, 43±53

53

Trethewey, R.N., Geigenberger, P., Riedel, K., Hajirezaei, M.R., Sonnewald, U., Stitt, M., Riesmeier, J.W. and Willmitzer, L. (1998) Combined expression of glucokinase and invertase in potato tubers leads to a dramatic reduction in starch accumulation and a stimulation of glycolysis. Plant J. 15, 109± 118. Trethewey, R.N., Geigenberger, P., Sonnewald, U., Hennig, A., MuÈller-RoÈber, B. and Willmitzer, L. (1999) Induction of the activity of glycolytic enzymes correlates with enhanced hydrolysis of sucrose in the cytosol of transgenic potato tubers. Plant Cell Environ. 22, 71±79. Veramendi, J., Roessner, U., Renz, A., Willmitzer, L. and Trethewey, R.N. (1999) Antisense repression of hexokinase 1 leads to an overaccumulation of starch in leaves of transgenic potato plants but not signi®cant changes in tuber metabolism. Plant Physiol. 121, 1±11. Viola, R., Davies, H.V. and Chudeck, A.R. (1991) Pathways of starch and sucrose biosynthesis in developing tubers of potato (Solanum tuberosum L.) and seeds of faba bean (Vicia faba L.): elucidation by carbon-13 spectroscopy. Planta, 183, 202±208. Waegemann, K. and Soll, J. (1991) Characterisation of the protein import apparatus in isolated outer envelopes of chloroplasts. Plant J. 1, 149±158. Wischmann, B., Nielsen, T.H. and MoÈller, B.L. (1999) In vitro biosynthesis of phosphorylated starch in intact potato amyloplasts. Plant Physiol. 119, 455±462.