Transfer of potato tubers to low temperature leads after 2–4 d to a stimulation of sucrose synthesis, a decline of hexose-phosphates and a change in the kinetic properties, and the appearance of a new form of sucrose phosphate synthase (SPS). Antisense and co-suppression transformants with a 70–80% reduction in SPS expression have been used to analyse the contribution of SPS to the control of cold sweetening. The rate of sucrose synthesis in cold-stored tubers was investigated by measuring the accumulation of sugars, by injecting labelled glucose of high specific activity into intact tubers, and by providing 50 mol m–3 labelled glucose to fresh tuber slices from cold-stored tubers. A 70–80% decrease of SPS expression resulted in a reproducible but non-proportional (10–40%) decrease of soluble sugars in cold-stored tubers, and a non-proportional (about 25%) inhibition of label incorporation into sucrose, increased labelling of respiratory intermediates and carbon dioxide, and increased labelling of glucans. The maximum activity of SPS is 50-fold higher than the net rate of sugar accumulation in wild-type tubers, and decreased expression of SPS in the transformants was partly compensated for increased levels of hexose-phosphates. It is concluded that SPS expression per se does not control sugar synthesis. Rather, a comparison of the in vitro properties of SPS with the estimated in vivo concentrations of effectors shows that SPS is strongly substrate limited in vivo. Alterations in the kinetic properties of SPS, such as occur in response to low temperature, will provide a more effective way to stimulate sucrose synthesis than changes of SPS expression.
Sugars accumulate at low temperatures in many plants (Müller Thurgau, 1882; Larcher 1983) and could play a role in osmoregulation and cryoprotection (Crow et al. 1993; Guy 1990). This phenomenon has been studied intensively in potato tubers, because it also affects processing. Cold sweetening detracts from product quality because reducing sugars react with amino groups of nitrogenous compounds at high frying temperatures, to produce an unwanted brown coloration in potato crisps and chips (Van Vliet & Schriemer 1960; Burton 1989). Despite intensive research, the crucial genes and processes that regulate cold sweetening have not been elucidated.
Gllycolysis is restricted at low temperatures because of the cold-lability of phosphofructokinase (PFK) (Dixon & ap Rees 1980; Dixon, Franke & ap Rees 1981; Hammond et al. 1990). Pyrophosphate:fructose-6-phosphate phosphotransferase (PFP) also becomes insensitive to activation by fructose-2,6-bisphosphate at low temperatures (Trevanion & Kruger 1991). This restriction of glycolysis is postulated to lead to an accumulation of hexose-phosphates (Pollock & ap Rees 1975) and a stimulation of sucrose synthesis (ap Rees et al. 1988). However, accumulation of hexose-phosphates cannot be the immediate trigger for sugar accumulation in the cold, for two reasons. Most of the evidence for an accumulation of hexose-phosphates has been obtained for tubers after 6–10 h in the cold. Sugar accumulation does not start until several days later (Isherwood 1973; Claassen et al. 1993; Hill et al. 1996), by which time the hexose-phosphates have declined again (Isherwood 1976; Hill et al. 1996). Also, antisense inhibition of PFP (Hajirezaei et al. 1994) did not stimulate sugar accumulation in the cold, even though hexose-phosphates were higher than in wild-type tubers. When pyrophosphatase from Escherichia coli was expressed to decrease PPi in the cytosol, sugar accumulation was stimulated, even though hexose-phosphates were lower than in wild-type tubers (Jelitto et al. 1992).
Increased invertase activity in cold-stored tubers (Pressey & Shaw 1966; Pressey 1969; Davies, Jeffries & Scobie 1989) might favour sugar accumulation by removing sucrose. Comparison of different cultivars shows that there is a strong correlation between invertase activity and the relative amounts of hexoses and sucrose in cold-stored tubers (Zrenner et al. 1996), and antisense inhibition of invertase led to an accumulation of sucrose and redued amounts of hexoses (Zrenner et al. 1996). However, there is no correlation between invertase activity and the total amount of sugar accumulated in different cultivars (Richardson et al. 1990; Zrenner et al. 1996). Moreover, decreased expression of invertase did not decrease the total amount of sugar accumulated in the cold (Zrenner et al. 1996).
Less attention has been paid to the possibility that low temperature may stimulate enzymes that are directly involved in the synthesis of sucrose. It is known that low temperature leads to increased SPS activity in leaves of cold-tolerant cultivars in several species (Guy et al. 1992; Holaday et al. 1992; Hurrry et al. 1995). Earlier studies indicated that SPS activity does not increase when potato tubers are cooled (Pollock & ap Rees 1975; Sowokinos 1990). It has been confirmed that the maximum activity of SPS and the total amount of SPS protein do not change after transferring Desirée tubers to 4 °C (Hill et al. 1996). However, there is a marked change in the kinetic properties of SPS (Hill et al. 1996) and a novel form of SPS with a slightly higher apparent molecular weight appears when tubers are cooled (Reimholz et al., unpublished data). These changes occur after 2–4 d at 4 °C, and are accompanied by a 2-fold decline of the hexose-phosphates, and a 3–4-fold stimulation of sucrose synthesis (Hill et al. 1996). These changes provide correlative evidence that SPS is involved in the regulation of sucrose accumulation at low temperature. In the present study, transgenic potato tubers with decreased expression of SPS were used to investigate directly whether SPS controls sugar accumulation in cold-stored tubers.
MATERIALS AND METHODS
Solanum tuberosum cv. Desirée was obtained from Vereinigte Saatzucht eG (Ebstorf, Germany). Transformed plants together with wild-type plants were grown in a greenhouse from regenerated primary transformants out of sterile culture or from seed potatoes derived from them.
cDNA library screening for SPS clones; constructs, and transformation
A cDNA library prepared from potato leaf poly(A) + RNA in λZAP II (Koßmann et al. 1992) was screened using polyclonal antibodies directed against the purified spinach SPS protein (Sonnewald et al. 1993) applying standard protocols (Sambrook, Fritsch & Maniatis 1989). Positive plaques were detected using the peroxidase system (Amersham Buchler, Braunschweig, Germany). Two putative clones were further investigated by sequencing. SPS-P4 (submitted to the EMBL; accession number X73477) is a full-length clone of 3622 bases including an open-reading frame that encodes a protein of 1053 residues with 56·3% identity at the amino acid level to maize leaf SPS (Bruneau et al. 1991) and 76·1% identity to spinach leaf SPS (Sonnewald et al. 1993; Klein et al. 1993). SPS-P1 is a 3741 bp partial clone with an open-reading frame encoding a polypeptide of 847 amino acids but otherwise with 96·6% homology to the polypeptide encoded by SPS-P4. Both contained the first methionine residue and the conserved N-terminal sequence of maize and spinach SPS (Bruneau et al. 1991; Sonnewald et al. 1993).
The full length cDNA clone SPS-P4 was cleaved with Xbal and Sall to give a 3600 bp restriction fragment that included the entire 3159 bp open-reading frame, and the fragment cloned in antisense orientation into the plasmid pBinAR (see Fig. 1). The partial clone SPS-P1 was cleaved with EcoRV and the resulting 2000 bp 5’ fragment including a 957 bp non-coding region and the N-terminal 45% of the coding region of the SPS gene was cloned in the sense orientation into the Smal site of the plasmid pBinAR under the control of the 35S CaMV promoter and the octopine synthase polyadenylation signal (see Von Schaewen et al. 1990). The resulting binary vectors, termed pU-SPS-5 and pU-SPS-1, respectively, were used for transformation of Agrobacterium strain C58C1:pGV2260 (Höfgen & Willmitzer 1988). Plant transformation was performed as in Rocha-Sosa et al. (1989).
Sucrose-phosphate synthase extraction and assay
Plant material was frozen in liquid nitrogen, ground in a precooled mortar and homogenized with extraction medium at a 1:3 ratio (fresh weight/extraction volume). The extraction medium contained 50 mol m–3 morpholinoethanesulfonic acid (MOPS), pH 7·4 (KOH), 12mol m–3 MgCl2, 1 mol m–3 ethyleneglycol-bis(β-aminoethyl ether)- N,N,N’,N’-tetraacetic acid (EGTA), 1mol m–3 ethylenediamine-N,N,N’,N’-tetraacetic acid (EDTA), 2 mol m–3 dithiothreitol (DTT), 1 mol m–3 benzamidine, 1 mol m–3ɛ-amino-caproic acid and 1·5 mol m–3 phenylmethylsulphonylfluoride (PMSF). Extracts (max. 500 mm3) were desalted using columns filled with 4 cm3 of G25-(medium)-material (Pharmacia, Freiburg, Germany). The SPS reaction was monitored, and the fructosyl moiety quantified using the anthrone test (Kerr & Huber 1987). Samples were incubated for 15 min at 25 °C in 70 mm3 buffer containing 50 mol m–3 MOPS pH 7·4, 5 mol m–3 MgCl2 with 4 mol m–3 fructose-1,6-biphosphate (Fru1,6bisP), 20 mol m–3 Glc6P and 3 mol m–3 uridine 5’ diphosphoglucose (UDPGlc) for assay of maximal activity (Vmax conditions). Alternatively the assay contained 2 mol m–3 Fru6P, 10 mol m–3 glucose-6-phosphate (Glc6P), 5 mol m–3 inorganic phosphate (Pi) and 6 mol m–3 UDPGlc (Vsel conditions) (Reimholz, Geigenberger & Stitt 1994). The reaction was stopped by adding 70 mm3 of 30% KOH (w/v) and heating at 95 °C for 10 min. After adding 1 cm3 of 0·14% (w/v) anthrone reagent (in 80% (v/v) H2SO4) absorbance was measured at 620 nm. A standard curve with 0–200 nmol sucrose was used to calculate absolute amounts of sucrose-6-P made during incubations.
Tubers were harvested form plants grown in the greenhouse for 3 months and allowed to develop full senescence. Harvested tubers were stored for several days or weeks (see figure legends for details) at 20 °C before transfer to 4 °C.
Preparation and assay of samples for metabolites
To measure carbohydrates, freshly cut slices (10 mm diameter, 2 mm thick) from tubers were extracted twice with 80% ethanol (v/v) at 80 °C for 30 min (1 cm3 80% ethanol per disc). The supernatant was used to quantify soluble sugars spectrophotometrically as in Stitt et al. (1989). The residue was washed again with 80% ethanol, dried and starch determined as glucose released after solubilizing starch granules by heating with 0·2 kmol m–3 KOH (95 °C, 3 h) and incubation with amyloglucosidase (Stitt et al. 1989).
To measure phosphorylated metabolites, tissue slices (10 mm diameter, 2 mm thick) were prepared from intact tubers and frozen in liquid nitrogen, taking care that the slices fell into liquid nitrogen immediately after being cut (Hajirezaei et al. 1994). Extracts were prepared in trichloroacetic acid (Jelitto et al. 1992) and assayed for metabolites as in Stitt et al. (1989). The recovery of a small, representative amount of each metabolite through the extraction, storage and assay procedures has been documented (Hajirezaei & Stitt 1991; Jelitto et al. 1992).
Labelling with 14C-glucose in intact tubers
A 1 mm diameter canal was bored through the centre of the tuber and filled with 20–30mm3 of 0·3mol m–3 U14C-glucose (16·8 GBq mmol–1), both ends sealed with Vaseline and the tuber incubated for 6 h at 4 °C (see Dixon & ap Rees 1980; Hajirezaei et al. 1994). A cylinder (10 mm diameter) concentric to the original bore hole was then removed, frozen at 20 °C, and later extracted successively at 80 °C in 80%, 80%, 50% and 50% ethanol, the extracts combined, dried under vacuum and re-dissolved in 1·5 cm3 water.
Labelling with U14C-glucose and tuber slices
To measure fluxes to sucrose and hexoses, tissue slices (10 mm diameter, 2 mm thick) were cut at 4 °C, washed three times in incubation buffer, incubated at 4 °C for 6 h with 50 mol m–3 U14C-glucose (2·9 MBq mmol–1) and then incubated for up to 18 h in the absence of glucose. The incubation medium (volume 6·5 cm3 for 17 discs) always included 20 mol m–3 2-(N-morpholino)-ethane sulphonic acid (MES)-KOH pH 6·5, and was shaken at 150 r.p.m. in a closed 25 cm3 vessel, which contained a small tube holding 500 mm3 50 mol m–3 KOH to absorb evolved 14CO2. Discs were harvested at the end of the pulse, and after 4, 8 and 18 h incubation without glucose, blotted dry and stored at –20 °C until extraction as above.
To obtain highly labelled samples for analysis of the glucan fraction, a modified procedure was used. Discs were cut as above, washed, incubated (two discs/200 mm3) in 2 mol m–314C-glucose (925 Mbq/mmol) and 10 mol m–3 MES-KOH (pH 6·5) for 6 h, washed three times in 50 cm–3 unlabelled buffer, blotted dry, and immediately extracted at 80 °C twice successively in 2 cm3 80% ethanol and twice successively in 2 cm3 50% ethanol, the extracts pooled, dried overnight under a stream of pressuriszed air at 40 °C and redissolved in 1 cm3 deionized water.
Fractionation of radioactive samples
The insoluble fraction was obtained by homogenizing the residue from the ethanol extraction in 1 cm3 water. The soluble fraction was separated by ion-exchange chromatography as in Hajirezaei et al. (1994). For analysis of glucans, the neutral fraction was reconcentrated by lyophylization at –20 °C, redissolved in 300 mm3 water, and further analysed by thin-layer chromatography on cellulose plates (Schleicher & Schüll, Dassel, Germany) using a ethylacetate:pyridine:water (100:10:45/v:v:v) chromatography buffer and developing them six to eight times in succession to improve the separation of the sugars as described (Geigenberger & Stitt 1993; Hajirezaei et al. 1994), or on silica plates (Silicagel 60, Merck, Darmstadt) using a isopropanol:butanol:water (12:3:4/v:v:v) solvent and developing them twice to improve resolution. Between 10 and 15 mm3 extract containing c. 2 kBq was applied to each track. The distribution of radioactivity was visualized by autoradiography (Hyperpaper 35S sequencing, Amersham) using 10 d exposure, or were scanned and quantified using a 14C-scanner (Tracemaster 20, EG & G Berthold, Bad Wildbad, Germany).
For qualitative analysis of the labelled glucans, native samples were compared with samples incubated for 2 h at 20 °C after diluting them 1:1 with buffer (10 mol m–3 citrate-KOH pH 5·0 for β-amylase, amyloglucosidase and pullanase; 10 mol m–3 MES-KOH plus 0·2 mol m–3 Ca2+ for α-amylase) containing 0·5 U α-amylase, 0·5 U β-amylase, 0·5 U amyloglucosidase or 0·1 U pullanase per 10 mm3 buffer.
Extracts were centrifuged and protein concentrations measured using the Biorad-staining reagent (Biorad, Richmond, USA) and bovine serum albumin (BSA) for calibration.
Generation of plants with decreased expression of SPS
Two closely related cDNA clones (97–99% identity of amino acid sequence) were isolated by screening a potato leaf cDNA library (see Methods). One was a full-length cDNA (SPS-P4; EMBL accession number X73477) encoding a polypeptide of 1053 (SPS-P4) amino acids and a predicted molecular weight of 118·3 kDa, with 56% and 76% homology to the previously cloned SPS from maize (Bruneau et al. 1991) and spinach (Sonnewald et al. 1993), respectively. A 3·6 kb fragment containing the entire open-reading frame of SPS-P4 linked in antisense configuration to the 35S CaMV promoter (Fig. 1) was used to generate transformants with antisense inhibition of expression (termed line 5 transformants). A 2·0 kb fragment of the closely related partial clone SPS-P1 including a 957 bp 5’-non-translated region and 45% of the coding sequence was linked to the 35S CaMV promoter in sense configuration (Fig. 1) and used to generate transformants with decreased SPS because of co-suppression (termed line 1 transformants).
Screening of SPS transcript in leaves showed a marked reduction of the SPS mRNA steady-state level in several independent antisense (5-59, 5-47, 5-15) and co-suppression (1-74, 1-67) lines (Fig. 1b). To screen for changes in SPS activity, we used an assay that contained saturating concentrations of hexose-phosphates (Vmax assay). The transformants showed a 40–60% decrease of the overall SPS activity in the source leaves (Fig. 1c). It is not unusual for the encoded activity or protein to show a smaller inhibition than the transcript. Repeated transformation with co-suppression and antisense constructs failed to yield transformant lines with less than 40% of wild-type SPS activity in their source leaves.
The transcript for SPS was also much decreased in growing and harvested tubers from the transformants (data not shown). SPS activity was decreased strongly in warm-stored (Fig. 2) tubers. The activities of sucrose synthase, alkaline and acid invertase was similar in wild-type and transformed tubers (data not shown).
Potato plants contain several forms of SPS (Reimholz et al. 1997) that are affected differerentially in these transformants. In leaves, the antisense and co-suppression transformants show a marked reduction in two major SPS forms (at 125 and 127 kDa) whereas a third major form (at 135 kDa) is not reduced (Reimholz et al. 1997). This may explain why repeated transformation did not yield transformants with a larger depression of the overall SPS activity in the leaves. Warm-stored tubers contain one major SPS form with a molecular weight of about 125 kDa and transfer to low temperature results in a selective increase of the form at 127 kDa (Reimholz et al. 1997). Both these forms are decreased in the transformants (Reimholz et al. 1997). The 135 kDa SPS form is present only at very low levels in tubers (Reimholz et al. 1997). This may explain why the decrease of SPS activity in a given transformant line was larger in tubers than leaves (compare Fig. 2 and Fig. 1c).
Rate of sugar accumulation in intact tubers stored at 4°C
In repeated harvests, no significant differences were found between the transformants and wild-type plants with respect to growth, tuber yield (data not shown) or composition of the harvested tubers (see below). After harvest the tubers were stored initially at 20 °C, and then transferred to 4 °C or held at 20 °C. The inhibition of SPS activity in cold-stored tubers was similar to that seen in warm-stored material. In a typical experiment, SPS activity was reduced from 400 ± 23 nmol gFW–1 min–1 in wild-type tubers to 167 ± 41, 155 ± 44, 137 ± 1 and 112 ± 7 in 1-74, 5-15, 1-67 and 5-59, respectively (mean ± SE, n = five tubers; see also the table and figure legends for activity measurements carried out in parallel with the physiological investigations).
Three methods were used to measure the rate of sucrose synthesis after transfer to low temperature. In the first approach we investigated the accumulation of sugar in intact tubers at low temperature. Five separate experiments with different batches of tubers were carried out over 2 years (Fig. 3). The tubers were stored at 5–7 °C in the first experiment and 3–5 °C in subsequent experiments. The first three experiments investigated sugar accumulation during the initial 3 weeks, and the last two experiments investigated sugar accumulation after 9 weeks at low temperature. The changes of sugars in wild-type tubers in these experiments resemble those seen previously (Isherwood 1973; Hill et al. 1996). Sucrose accumulation starts after a lag of about 4 d, and hexose accumulation starts even later. In this later phase, sucrose is converted into reducing sugars, and the net rate of accumulation of sugar slows down. Sugars still accumulate in the transformants, but in most experiments and for most transformants the sugars are slightly lower than in wild-type tubers. The only exceptions are line 1-67 in experiment II, and the initial rate of accumulation in 1-74 in experiment V. In several cases, the decrease is slightly more marked towards the end of the cold incubation than in the initial phase, but further experiments would be needed to confirm this trend.
The results of the five experiments are combined in Table 1. For each line and experiment, the net sugar accumulation over the entire cold incubation is expressed as a percentage of the sugar accumulation in the wild-type tubers in that batch. SPS activity is also expressed as a percentage of the wild-type activity in each batch, and averaged across all the experiments carried out with these transformants. Tuber lines with on average 34%, 22%, 23% and 19% of wild-type SPS activity accumulated, on average, 88%, 93%, 83% and 59% as much sugar as wild-type tubers, respectively. A 3–5-fold decrease in SPS expression therefore led to a small and non-proportional inhibition of sugar accumulation.
Table 1. . Accumulation of sugars in potato tubers stored at 4°C. The net increase of sucrose, glucose and fructose is estimated from the difference between day 0 and the final time point (days 21–62) for each experiment in Fig. 3. The results are normalized to the wild-type value in that experiment
Metabolization of U14C-glucose in intact tubers
In a second approach, we investigated the metabolism of labelled glucose in intact tubers. To minimize side effects produced by wounding and addition of exogenous carbohydrate, a small (1 mm diameter) hole was bored through an otherwise intact tuber and filled with a solution of 0·3 mol m–3 U14C-glucose of high specific activity (Table 2). After 6 h at 4 °C, a concentric core was removed from around the original bore hole for analysis. The amount of U14C-glucose added (2 nmol glucose gFW–1, legend to Table 2) is negligible compared to the endogenous glucose pools (Fig. 3) or endogenous turnover of carbohydrate (the net accumulation of sugars in cold-stored tubers is 200–320 nmol hexose gFW–1 h–1, calculated from Fig. 3).
Table 2. . Labelling pattern after injecting 0·3 mol m–3 (U14C)-glucose into intact potato tubers at 4°C. The tubers had been harvested from 3-month-old plants grown in pots in the greenhouse, stored at 20°C for 10d, and then transferred to 4°C for 15 d. A 10–25mm3 aliquot of high-specific activity (10·8 GBq mmol–1) 0·3 mol m–3 (U14C)-glucose was injected into a fine 1mm borehole in an otherwise intact tuber, which was then incubated at 4°C for 6h before harvesting a 10mm diameter tissue core symmetrical to the borehole and analysing the labelling pattern. The radioactivity recovered in the core (25–30kBq per core) is equivalent to 40–50% of the introduced radioactivity, and represents an input of about 2nmol glucose gFW–1. Values are means±SE (for three separate tubers)
Between 55 and 72% of the added U14C-glucose was metabolized in wild-type and transformed tubers. The percentage metabolized was actually slightly higher in the transformants with decreased SPS activity. This could be attributable to slightly faster metabolism of glucose, or to less dilution of the added U14C-glucose by internal unlabelled glucose pools.
In the wild-type tubers, almost half of the absorbed label is converted to sucrose (see also Hill et al. 1996). Labelling of sucrose in tubers from transformant 5-15 (which has the smallest inhibition of SPS expression, see above) resembled that in wild-type tubers. Transformants 1-74 and 5-59 showed significantly reduced labelling of sucrose. The relative rates of sucrose synthesis can be estimated as 14C carbon in sucrose plus 2 ×14C carbon in fructose (assuming that the label in fructose derives from hydrolysis of sucrose). The estimated rate of sucrose synthesis is not significantly decreased in transformant 5-15, and is decreased by 24% and 25% in lines 1-74 and 5-59.
Labelling of cationic (amino acids) and anionic compounds was increased in low-SPS tubers, compared with wild-type tubers (Table 2). At 4 °C, the anionic fraction includes considerable label in phosphorylated intermediates, and organic acids (Pollock & ap Rees 1975; Hill et al. 1996). The transformants (especially 5-59) also incorporated a small amount of label into neutral compounds that remained at the origin or ran between the origin and sucrose during TLC separation of the neutral fraction (Fig. 4c). There is little 14C carbon at this region in wild-type tubers (Fig. 4a).
Metabolization of U14C-glucose by tuber slices
In a third approach, freshly cut tissue slices from cold-stored tubers were incubated with 50 mol m–3 U14C-glucose for 6 h at 4 °C, and then incubated in the absence of glucose for up to 18 h at 4 °C (Fig. 5).
The rate of glucose uptake (data not shown) was similar in wild-type slices and in the transformants. The glucose was metabolized slightly faster in the wild-type tubers (Fig. 5). In slices from wild-type tubers, more than 50% of the absorbed U14C-glucose was converted to sucrose during the 6 h pulse (Fig. 5), at a rate of about 0·2 μmol hexose gFW–1 h–1. This is similar to the net rate of sucrose accumulation in intact tubers (0·2–0·32 μmol hexose gFW–1 h–1,calculated from Fig. 3 and Table 1). During the first 4 h of the ensuing incubation without glucose, label in sucrose rose further to 70% of the absorbed label. Later, the label in sucrose decreased slightly, and label appeared in fructose (Fig. 5). Label in glucose, which had declined during the first 8 h of the incubation without glucose, also rose again later (Fig. 5). The reciprocal changes of label in sucrose and the reducing sugars are consistent with slow hydrolysis of sucrose by invertase. Label in the anionic fraction fell during the incubation without glucose, and that in cations and CO2 rose. Labelling of insoluble compounds was low throughout the experiment (Fig. 5).
A different labelling pattern was found in slices from tubers with decreased expression of SPS (Fig. 5). During the 6 h pulse only 13–15% of the absorbed label was converted to sucrose, compared with more than 50% in wild-type tubers. This experiment was carried out with two co-suppression lines. A similar inhibition of sucrose synthesis during a pulse experiment was obtained in the antisense lines 5-47 and 5-59 (data not shown). Unexpectedly, the transformants incorporated label into sucrose at high rates in the first part of the ensuing incubation without glucose. The rate of label incorporation into sucrose in 1-67 and 1-74 tubers during the first 4 h of the incubation without glucose was equivalent to 0·15 and 0·11 μmol hexose gFW–1 h–1. This is faster than label incorporation in the transformed material during the pulse (see below for explanation). Label incorporation into sucrose continued, although at a slower rate, throughout the remainder of the experiment. At the end of the experiment, the label in sucrose represented 49% and 56% of the label metabolised in 1-74 and 1-67 tubers, respectively. This is 23–28% less than in the wild-type material at the end of the experiment. To understand why the labelling kinetics of sucrose were so different in the wild-type and transformant tubers, the movement of label through other pools in the tubers was investigated.
Labelling of the anionic fraction was much higher in the transformants than in wild-type discs during the pulse (17–20% compared with 7%). Label decreased in the anions and rose in the cations and CO2 during the following incubation without glucose (Fig. 5). At the end of the pulse, labelling of anions was still higher in the transformants than in wild-type discs, and labelling of cations and CO2 was also 2-fold higher in the transformants. There was also more label in the insoluble fraction in the transformants (7–9%, compared with 2% in wild-type discs, Fig. 5).
The neutral fraction collected from the ion-exchange columns was routinely analysed by TLC on cellulose plates to separate sucrose from other sugars. Unexpectedly, in the extracts from the transformants a large part of the neutral label remained at the origin. Two new peaks appeared: a small peak that just moved into the plate, and a larger peak that ran slightly more slowly than sucrose (Fig. 6a–c). On the basis of Rf values, these could correspond to maltotriose and maltose, respectively (see below for more data). A similar labelling pattern was found in a separate experiment with tubers from the antisense line 5-59 (shown in Fig. 6d). Incubation of the extracts with α(1-4),(1-6) amyloglucosidase before chromatography resulted in a complete loss of label from the origin and the two slow-running positions, and a corresponding increase in label co-eluting with glucose (data not shown, see also below). The label in this putative glucan fraction accounted for 20–30% of the metabolized label, compared with less than 1% in wild-type material at the end of the pulse. The label in the putative glucan fraction decreased to low levels during the incubation without glucose (Fig. 5). These results indicated that in cold-stored tubers with decreased SPS activity the glucan pool turns over rapidly when they are stored at low temperature. This labelling of glucans was absent in labelling experiments carried out at 20 °C with slices from warm-stored tubers (data not shown).
To establish the identity of this putative glucan fraction, tuber slices were incubated with high-specific-activity glucose to increase label incorporation, and the neutral compounds were separated on silica plates to improve separation of high-molecular-weight oligosaccharides (Fig. 7). Whereas only traces of maltose and maltotriose and no higher-weight oligosaccharides could be seen in the wild type, there was heavy labelling of a series of oligosaccharides of increasing molecular weight in the transformant (Fig. 7a). There was also heavy labelling of larger compounds that remained at the origin. These compounds were almost completely converted to glucose after incubation with α(1-4),(1-6)amyloglucosidase (Fig. 7b, lane 6). Incubation with α-amylase converted most of the higher-molecular-weight compounds into maltotriose, maltotetraose, maltopentaose, and maltohexaose, but left a small amount of label at the origin (Fig. 7b, lane 2). Incubation with β-amylase converted the higher-molecular-weight compounds into maltose, but left a considerable amount of label at the origin (Fig. 7b, lane 3). Inclusion of pullanase in the incubations with either α-amylase (lane 4) or β-amylase (lane 5) resulted in a marked reduction in the amount of label remaining at the origin. Similar results were obtained with three separate batches of extracted glucans (data not shown) These results show that these compounds are predominantly α(1-4) glucans, with a very low frequency of α(1-6) branching in the higher-molecular-weight material that remains at the origin.
The rate of sucrose synthesis in tuber slices at 4 °C could, in principle, be estimated in three different ways from the data in Fig. 5. (1) Comparison of the label entering sucrose during the pulse indicates that sucrose synthesis is inhibited by up to 75% in the transformants. However, this approach probably overestimates the inhibition, for the following reason. Sucrose is labelled more slowly during the 6 h pulse than during the first 4 h of the incubation without glucose in the transformants, and this delay is accompanied by the accumulation and subsequent release of 14C carbon from the glucan pool. These results could be explained if U14C-glucose, after entering the tissue slices of the transformants, equilibrates with an unlabelled pool of glucans. This exchange would decrease the flow of 14C carbon into sucrose during the pulse, with the result that the synthesis of sucrose is underestimated in the transformants compared to the wild-type value. (2) By comparing the maximum rate of 14C carbon-incorporation into sucrose during the 6 h pulse for the wild-type discs (0·2 μmol hexose gFW–1 h–1) and during the first 4 h of the incubation without glucose in the transformants (0·15–0·11 μmol hexose gFW–1 h–1 in 1-67 and 1-74 tubers): this comparison indicates a 25% and 45% inhibition of sucrose synthesis in transformants 1-67 and 1-74, respectively. (3) By comparing label distribution at the end of the incubation without glucose: the 14C carbon moving into the sucrose pool (calculated as 14C carbon in sucrose plus 2×14C carbon in fructose to compensate for subsequent hydrolysis) is decreased by 23% and 28% in transformants 1-67 and 1-74, respectively.
Changes of metabolite levels
Potato tuber SPS is activated by Glc6P and inhibited by Pi (Reimholz et al. 1994). When tubers from transformants were stored in the cold, they contained up to 60% more Fru6P and Glc6P, and 35% more UDPGlc than did the wild-type tubers (Fig. 8). They also contained slightly higher 3-phosphoglycerate (3PGA) and ATP (Fig. 8). Similar increases of metabolites were seen when warm-stored tubers were analysed (data not shown).
Potato-tuber SPS is regulated by protein phosphorylation (Reimholz et al. 1994) leading to a change in kinetic properties that allows activity in an assay in the presence of limiting substrate levels and the inhibitor Pi (Siegl, MacKintosh & Stitt 1990; Reimholz et al. 1994; Huber & Huber 1996). SPS activity was assayed (1) in the presence of saturating concentrations of hexose-phosphates (Vmax assay) and (2) in the presence of lower concentrations of hexose-phosphates and the inhibitor Pi (Vsel assay) (Fig. 9). There was no change of Vmax activity between warm- and cold-stored tubers (Fig. 9a–d). This confirms the results from Western Blot analysis (Hill et al. 1996; Reimholz et al. 1997). There were marked changes of activity when SPS was assayed in the presence of limiting substrates plus Pi (Vsel assay). Vsel activity increased when wild-type tubers were transferred to low temperature (Fig. 9a; see also Hill et al. 1996). As a result the ratio of Vsel:Vmax activity in wild-type tubers was low in the warm and increased after transfer to low temperature (Fig. 9e). In the transformants, the activity in the Vsel assay was already quite high (relative to Vmax) at high temperature (Figs 9b–d), and there was only a small additional rise Vsel (Figs 9b–d) and the Vsel:Vmax ratio (Fig. 9e) after transfer to low temperature.
Reduced expression of SPS in tubers
SPS protein and activity in potato tubers have been reduced by 65–80% by antisense inhibition and co-suppression in several independent transformant lines (Fig. 1 and Reimholz et al. 1997). This material has been used to investigate how SPS contributes to the regulation of cold sweetening. A potential problem arises in interpreting physiological experiments, because potato plants contain several forms of SPS, and their expression is not affected in the same way by the constructs that we have used (Reimholz et al. 1997). However, this should not interfere with the interpretation of experiments with stored tubers. Most SPS protein in warm-stored tubers is attributable to one major band at 125 kDa, and the accumulation of sugar and the alteration of the kinetic properties of SPS that occur after 2–4 d in the cold (Hill et al. 1996) are associated with a selective increase of a further band at 127 kDa (Reimholz et al. 1997). The 125 kDa and 127 kDa forms are both markedly reduced in our antisense and co-suppression lines (Reimholz et al. 1997). The higher-molecular-weight form of SPS at 135 kDa (which is not repressed by our constructs) is present only at low levels in stored tubers, and does not alter in response to low temperature (Reimholz et al. 1997).
Overall SPS activity is not limiting for the rate of cold sweetening
The rate of sucrose synthesis in cold-stored tubers was measured in three different ways. Each approach could be subject to error: (1) Estimates based on net accumulation of sugars (Table 2, Fig. 3) will underestimate the absolute rate of sucrose synthesis if there is rapid recycling of sucrose or hexose sugars. The estimates can nevertheless be used to assess the contribution of SPS to the control of sucrose synthesis, provided that the rate of recycling is not high and does not change greatly between wild-type and transformed tubers. The latter is a reasonable assumption, as the activities of sucrose-degrading enzymes were not altered in the transformants. (2) The distribution of label after injecting tracer amounts of U14C-glucose into whole tubers (Table 2) provides relative values for fluxes. These will give an erroneous picture of the real fluxes if the tracer 14C is diluted to a different extent by internal metabolite pools in the wild-type and transformant tubers. This is probably not a major source of error, because the internal pools of hexose sugars do not vary greatly between genotypes (Fig. 3). (3) Estimates based on the fate of 50 mol m–3 U14C-glucose in tissue slices (Fig. 5) could be in error because the exogenous carbohydrate may modify metabolism, and because of changes that could result from wounding. The rates of glucose uptake and metabolism in the discs are in the same range as the fluxes of carbon in intact tubers (see Results). Metabolism in the slices is nevertheless altered; in particular, in the transformants, there is a large label incorporation into glucans during the pulse. This labelling was transient, and label moved out of the glucans into end-products during a subsequent incubation without glucose. The distribution of 14C carbon at the end of the experiment was therefore used to estimate the rate of sucrose synthesis in the disc experiments.
The results from the three sets of flux measurements are summarized in Fig. 10. For each approach, the fluxes in the transformants are normalized on the flux in the wild-type tubers, and are plotted against SPS activity, which is also normalized against the wild-type value. All three approaches give a similar answer, which provides support for their reliability. Transformants with a 70% reduction of SPS activity only show a slight inhibition of sucrose synthesis. A reduction of 75–80% led to a 10–35% decrease of sucrose synthesis, depending on the method used to estimate the flux. These results show that the rate of cold sweetening in wild-type tubers is not strongly controlled by the overall SPS activity or (see Reimholz et al. 1997) the overall amount of SPS protein.
Comparison of in vivo enzyme kinetics with in vivo fluxes and metabolite concentrations also indicates that wild-type tubers contain excess SPS activity. The maximum in vitro activity of SPS in wild-type tubers is between 18 and 27 μmol sucrose gFW–1 h–1 at 25 °C, which would be equivalent to 4·5–6·8 μmol sucrose gFW–1 h–1 at 4 °C, assuming a Q10 of about 2 (Pollock & ap Rees (1975). This is 30- to 50-fold above the estimated rate of sucrose accumulation in intact tubers after transfer to low temperature (0·1–0·16 μmol sucrose gFW–1 h–1, see Fig. 3), or the rate of sucrose synthesis from external 50 mol m–3 glucose in tuber discs from cold-sweetening tubers (about 0·1 μmol sucrose gFW–1 h–1, Fig. 5). In transformants with 80% inhibition of SPS expression, SPS is starting to restrict the rate of sucrose synthesis in cold-sweetening tubers (see Fig. 10). In these lines the estimated maximum activity of SPS at 4 °C (0·7–1·0 μmol sucrose gFW–1 h–1) is still 5–10 times higher than the rate of sugar accumulation in intact tubers (0·050–0·075 μmol sucrose gFW–1 h–1, Fig. 3) and the rate of sucrose synthesis from external 50 mol m–3 glucose (0·08–0·12 μmol sucrose gFW–1 h–1; Fig. 5).These comparisons show that (1) SPS activity in wild-type tubers is much higher than the rate of sucrose synthesis during cold-sweetening, and (2) SPS can achieve only a fraction of its maximum activity in vivo.
Comparison of the kinetic properties of partially purified SPS from cold-stored potato tubers (Reimholz et al. 1994) with the levels of metabolites in cold-stored tubers (Table 3) shows that SPS is strongly substrate limited in vivo, in particular with respect to UDPGlc. When SPS is assayed in the presence of saturating concentrations of its activator Glc6P, the estimated cytosolic concentrations of Fru6P and UDPGlc are similar and about 10-fold lower than the Km values for these substrates, respectively. When Pi (a competitive inhibitor to Glc6P; Doehlert & Huber 1983; Reimholz et al. 1994) is added at similar concentrations to Glc6P, the estimated cytosolic concentrations are 6- and 25-fold lower than the Km values.
Table 3. . Comparison of estimated in vivo substrate concentrations and the kinetic parameters of SPS. The Fru6P and UDPGlc concentrations are estimated from the data in Fig. 8, assuming that the cytosol represents 5% of the tuber fresh weight, and that half the Fru6P and all the UDPGlc are located in the cytosol. The Km values for potato tuber SPS are calculated from Reimholz et al. (1994)
Metabolic compensation for decreased expression of SPS
Potato-tuber SPS has highly co-operative kinetics for the substrates (UDPGlc and Fru6P) and the allosteric activator (Glc6P) (Reimholz et al. 1994). The levels of the substrates and Glc6P are 40–60% higher in the transformants than in the wild type (Fig. 8), which will strongly activate the remaining SPS protein. Significantly, changes in Glc6P lead to large changes in the affinity for UDPGlc (Reimholz et al. 1994) which (see above) is especially limiting in vivo.
Cold-induced changes in the apparent kinetic properties of SPS (Fig. 9) might also compensate for the decreased expression of SPS in the transformants. Regulation of leaf SPS by protein phosphorylation leads to changes of the ratio of Vsel/Vmax activity (Huber & Huber 1992, 1996). Potato tuber SPS contains the phosphorylation site identified in spinach leaf SPS (McMichael et al. 1993). Moreover, SPS activity is modified when tuber slices are incubated with okadaic acid (Reimholz et al. 1994). However, an alternative explanation for the results in 9Fig. 9e would be that the 135 kDa form of SPS has different kinetic properties, and that the shift in the kinetic properties in the extracts from the transformants occurs because the repression of the lower-molecular-weight forms means that the minor form makes a larger contribution to the remaining SPS activity (see Reimholz et al. 1997). Purification and comparison of the different SPS forms will be necessary to distinguish between these explanations.
Further effects of decreased expression of SPS on tuber metabolism
The decreased rate of sucrose synthesis in transformants is accompanied by increased levels of glycolytic intermediates, increased labelling of CO and other products of glycolysis, and increased incorporation of label into maltose, maltotriose, a range of short-chain glucans, and larger glucans with a small amount of α(1–6) branching. Labelling of glucans is low but significant when high-specific-activity glucose is injected into intact tubers, and considerable when 50 mol m–3 U14C-glucose is added to tuber slices. The label is lost again during a subsequent incubation without glucose, showing that the glucan pool undergoes rapid turnover. One explanation for the labelling of glucans in the low SPS tubers would be that high concentration of hexose-phosphates stimulates the reverse (synthetic) reaction of α-glucan phosphorylase, leading to label incorporation into a transient glucan pool. An alternative explanation would be that ADPglucose pyrophosphorylase is stimulated by the high levels of phophorylated intermediates in the transformants. The increased labelling in discs in 50 mol m–3 glucose may be because hexose phosphates are especially high in these conditions (M. Geiger, data not shown).
Role of SPS in the regulation of sugar accumulation at low temperatures
These results are of interest because they contribute to the understanding of processes that regulate sugar accumulation in the cold in wild-type tubers, and to the design of strategies to reduce cold-sweetening in tubers. First, the overall amount of SPS is not important for the control of sucrose synthesis at low temperatures in tubers. Small changes of overall SPS activity do not inhibit sucrose synthesis, and a 4–5-fold reduction in SPS activity leads only to a small and non-proportional inhibition of sugar accumulation in cold-stored tubers of the cultivar Desirée, because metabolic adjustments partly compensate for the decreased expression of SPS. The combination of decreased SPS expression with additional changes to prevent an accumulation of substrates of SPS could provide a more successful strategy for producing a stronger inhibition of cold sweetening. Second, potato-tuber SPS has highly co-operative kinetic properties (Reimholz et al. 1994) and SPS is strongly restricted in vivo by the concentrations of its substrates and activators. Changes of the kinetic properties of SPS will therefore provide a more effective mechanism to alter the rate of sucrose synthesis than changes of the overall amount of SPS protein. This underlines the importance of cold-induced changes in the kinetic properties of SPS (Hill et al. 1996; Reimholz et al. 1997) for the regulation of cold sweetening. Third, it is becoming evident from our studies that many genes will affect sucrose accumulation in cold-stored potato tubers. The momentary rate of sucrose synthesis will depend on the properties of SPS itself, on the proteins that are involved in the cold-induced shift in the apparent molecular weight of SPS and the accompanying alteration of its kinetic properties, and on genes that influence the levels of the hexose phosphates and UDPglucose. The amount of sugars that accumulate will also depend on genes that affect the recycling of sucrose and reducing sugars (see discussion in Hill et al. 1996).
This research was supported by the DFG (SFB 199; M. S., R. R.) and the Bundesministerium für Forschung und Technologie (K.-P. K., U. S., L. H.).