To test the hypothesis that the up-regulation of sucrose biosynthesis during cold acclimation is essential for the development of freezing tolerance, the acclimation responses of wild-type (WT) Arabidopsis thaliana (Heynh.) were compared with transgenic plants over-expressing sucrose phosphate synthase (over-sps) or with antisense repression of either cytosolic fructose-1,6-bisphosphatase (antifbp) or sucrose phosphate synthase (antisps). Plants were grown at 23 °C and then shifted to 5 °C. The leaves shifted to 5 °C for 10 d and the new leaves that developed at 5 °C were compared with control leaves on plants at 23 °C. Plants over-expressing sucrose phosphate synthase showed improved photosynthesis and increased flux of fixed carbon into sucrose when shifted to 5 °C, whereas both antisense lines showed reduced flux into soluble sugars relative to WT. The improved photosynthetic performance by the over-sps plants shifted to 5 °C was associated with an increase in freezing tolerance relative to WT (−9.1 and −7.2 °C, respectively). In contrast, both antisense lines showed impaired development of freezing tolerance (− 5.2 and −5.8 °C for antifbp and antisps, respectively) when shifted to 5 °C. In the new leaves developed at 5 °C the recovery of photosynthesis as typically seen in WT was strongly inhibited in both antisense lines and this inhibition was associated with a further failure of both antisense lines to cold acclimate. Thus, functional sucrose biosynthesis at low temperature in the over-sps plants reduced the inhibition of photosynthesis, maintained the mobilization of carbohydrates from source leaves to sinks and increased the rate at which freezing tolerance developed. Modification of sucrose metabolism therefore represents an additional approach that will have benefits both for the development of freezing tolerance and over-wintering, and for the supply of exportable carbohydrate to support growth at low temperatures.
When cold-tolerant plants are exposed to low temperatures (below 5 °C) a series of events is initiated that, over a period of time varying from days to weeks, results in these plants acclimating to the lower growth temperature and becoming more freezing tolerant. This process is generally referred to as either cold acclimation or cold hardening. As cold acclimation and the development of freezing tolerance in cold-tolerant plants are delayed processes, there are two separate problems that need to be addressed when attempting to understand freezing tolerance with the view of modifying this trait by breeding or genetic manipulation. One problem is the sensitivity of non-hardened (NH) plants or sensitive organs to damage from mild frost. To address this we need to modify the basic frost sensitivity of non-hardened tissues or frost-sensitive organs, such as the flowers. The second problem is to understand how plants develop and maintain deep freezing tolerance. Addressing the second problem requires an understanding of how metabolism as a whole is regulated in response to prolonged exposure to low temperatures. Understanding these longer-term responses will provide insight into how plants can not only survive individual freeze/thaw events but also sustain freezing tolerance through multiple freeze/thaw cycles and maintain basal metabolism during over-wintering.
To test the hypothesis that the up-regulation of the sucrose biosynthetic pathway during cold acclimation is an essential element for the development of freezing tolerance, we made transgenic Arabidopsis plants with altered rates of sucrose synthesis and compared these to WT plants after exposure to low temperature. We used three types of transgenic plants of the same ecotype but with specific modifications to sucrose metabolism. First, plants with reduced cytosolic FBPase activity, reduced rates of sucrose synthesis but a compensatory increase of starch synthesis (Strand et al. 2000); second, transgenic plants with reduced sucrose phosphate synthase (SPS) activity, reduced rates of sucrose synthesis but no increase of starch synthesis (Strand et al. 2000); third, transgenic plants with an increased expression of SPS and an increased capacity for sucrose synthesis (Signora et al. 1998). The following experiments demonstrate that using reverse genetics to alter the rates of sucrose synthesis can either mimic (sense transgenics) or attenuate (antisense transgenics) the cold-acclimation responses of Arabidopsis that enhance the partitioning of newly fixed carbon into soluble sugars, and that these changes in photosynthetic performance and soluble sugar production strongly influence the development of freezing tolerance at low temperature.
MATERIALS AND METHODS
Arabidopsis thaliana L. (Heynh.) ecotype Colombia and three transgenic lines were selected as representative lines from multiple lines we have characterized previously (Signora et al. 1998; Strand et al. 2000). For the construction of the antisense lines cDNA inserts of the EST clones 109I5T7 and 169K9T7 were ligated in an antisense orientation under the control of the 35S CaMV promoter into the binary vector pBI120 (Jefferson, Kavanagh & Bevan 1987). The Arabidopsis EST 109I5T7 is a 1810 bp partial clone encoding the C-terminal of SPS (EC 188.8.131.52). The Arabidopsis EST clone 169K9T7 encodes the entire 1060 bp structural protein of cytosolic fructose-1,6-bisphosphatase (cFBPase; EC 184.108.40.206). From the three antisense repression of sucrose phosphate synthase (antisps) and the three antisense repression of cytosolic fructose-1,6-bisphosphatase (antifbp) lines studied previously (Strand et al. 2000), antisps line 6 and antifbp line 12 were selected for these experiments. Arabidopsis plants with elevated amounts of SPS (over-sps) were made using the pCGN3812 construct, containing the maize SPS cDNA under the control of the rbcs promoter of tobacco (Worrell et al. 1991; Signora et al. 1998). From the lines studied previously (Signora et al. 1998), line 5 was selected for these experiments.
Seeds were germinated under controlled-environment conditions: 150 µmol photons m−2 s−1; day/night temperature regime 23/18 °C; and photoperiod 8 h. After 49 d, when the leaves had developed into fully mature source leaves the plants were shifted to a day/night temperature regime 5/5 °C. The photoperiod was still 8 h and the light 150 µmol photons m−2 s−1. After 1 and 10 d at 5 °C, source leaves that had completed expansion before transfer to low temperature were harvested to analyse acclimation in pre-existing leaves. After 40 d at 5 °C, source leaves that had developed at 5 °C were sampled to investigate acclimation processes requiring modification during leaf development.
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblotting
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was carried out according to Laemmli (1970). Proteins were transferred electrophoretically to a polyvinylidene difluoride membrane and visualized using an ECL chemiluminescent kit (Amersham Pharmacia Biotech, Little Chalfont, Bucks, UK). The SPS antibody was raised against a conserved fragment from the potato protein (Reimholz et al. 1997) and the cFBPase antibody was raised against the full-length protein from Arabidopsis (Strand et al. 2000). Western blots were quantified using spot density measurements (AlphaImager TM 1200; B and L Systems, Maarssen, The Netherlands).
Leaf material was frozen in ambient light using a freeze clamp chilled to the temperature of liquid N2. The frozen material was ground to a fine powder at the temperature of liquid N2 in a mortar. Enzymes were extracted and measured as in Strand et al. (1997, 1999).
Chlorophyll, protein content and specific leaf weight
Chlorophyll was determined in 80% buffered acetone (Porra, Thompson & Kriedemann 1989) and total protein according to Bradford (1976) (Bio-Rad Laboratories, Hercules CA, USA). Specific leaf dry weight was determined after drying leaf discs to constant weight at 80 °C.
CO2 exchange and carbon partitioning
Gas exchange was measured with an open flow gas exchange system on attached leaves (Model LI-6202; Li-Cor Inc., Lincoln. NE, USA). The 14C-partitioning was investigated by incubating leaf discs in an oxygen electrode (LD-2; Hansatech, Kings Lynn, Norfolk, UK) at 600 µmol m−2 s−1 20 min at 23 °C, and for 40 min at 5 °C with 5% CO2 containing 4 µCi 14CO2. The 14C incorporation in the soluble fractions and starch was analysed as in Kruckenberg et al. (1989).
Soluble sugars and proline
At various times during the photoperiod, leaf material was harvested into liquid N2. Sucrose, glucose, fructose and starch were measured in the soluble and residual fractions of ethanol–water extracts (Stitt et al. 1989). Proline was determined in the supernatants according to Bates, Waldren & Teare 1973).
A 2 cm2 piece of washed leaf material was put in a glass tube containing 200 µL of high-performance liquid chromatography (HPLC)-grade water. The tubes were put in an ethanol bath (Hetofrig; Heto, Birkrød, Denmark). At −2 °C the temperature was kept constant for 1 h and a steel needle that had been cooled in liquid nitrogen was used to initiate ice formation. The samples were then frozen to different temperatures at a rate of 2 °C h−1. For every chosen temperature the samples were removed and placed on ice to thaw. Control leaves were kept on ice for the same time. When the leaf samples were thawed, an additional 1.3 mL of HPLC-grade water was added to the samples and the tubes were put on a shaker at room temperatures for approximately 15 h. The electrolyte leakage was determined after freezing using a conductivity cell (CDM210; Radiometer, Copenhagen, Denmark). To determine the total ion content in the leaves the samples were frozen in liquid N2. The samples were shaken again for at least 12 h and the total ion content was determined. Freeze damage was expressed as percentage electrolyte leakage before and after freezing in liquid N2, and was corrected for the electrolyte leakage from the non-frozen control leaves.
RNA isolation and reverse transcriptase-polymerase chain reaction
Total RNA was isolated by using the Trizol® reagent according to the manufacturers protocol (Life Technologies, Frederik, MD, USA). The total RNA samples were DNase treated prior to reverse transcriptase (RT)-polmerase chain reaction (PCR) reactions. SuperScript® A one-step RT-PCR system kit (Life Technologies) was used for the RT-PCR reactions. Ten microlitres of the PCR reaction mix was run on a 1% agarose gel. The following primers were synthesized for the purpose: COR15a, TTTCTCA ACGCAAGAAGTCGTT and TAGAAATTACAACAG ACTCATC; COR6.6, CAA CAAGAATGCCTTCCAAG and TCCAAACGTAGTA CATCTAAAGGG; COR47, AGTGAAGGAGAACAA GATTA and GCATGATAA CCTGGAAGCTT; COR78, ATCAGAAGCCAGGACA ATTT and TCGCCGGAAA TTTATCCTCT. The primers for ACTIN were synthesized according to Ha et al. (1999).
One-way analysis of variance (anova) with Tukey–Kramer multiple comparison post tests were performed using GraphPad InStat (version 3.00; GraphPad Software, San Diego, CA, USA) to test for significant differences between wild-type (WT) and the different transgenic lines. The lethal freezing temperatures (TEL50s) were calculated from sigmoidal functions fit to the electrolyte leakage data using Microcal Origin (version 6; Microcal Software, Northampton, MA, USA).
Western blot and enzyme activity measurements
Western blot analysis (Fig. 1a & c) and enzyme assays (Fig. 1b & d) for cFBPase and SPS from WT Arabidopsis leaves showed strong increases in the amount of protein and enzyme activity following exposure to low temperature. The activation state of SPS in these samples collected 2 h into the photoperiod increased from 37% in NH leaves to 45% in 10 d leaves and then returned to 38% in leaves that developed at 5 °C. These results are similar to those reported previously for Arabidopsis (Strand et al. 1997, 1999).
Under warm growth conditions the antifbp plants contained approximately 12% of the WT cFBPase protein, and less than 40% of the apparent cFBPase activity (Fig. 1a & b). The inhibition of cFBPase activity underestimates the decrease of cFBPase protein because even though assay conditions were used that favoured cFBPase, some plastid FBPase was also detected (Strand et al. 2000). Nevertheless, the data show that in the antifbp plants the cold-induced increase in the cFBPase protein and activity was largely blocked (Fig. 1a & b). The antifbp plants contained slightly less SPS protein and enzyme activities than WT (Fig. 1c & d) and SPS activation was also lower (21, 34 and 27% in NH, 10 d and 5 °C developed leaves, respectively).
The antisps plants contained approximately 20% of the WT SPS protein and 25% of the WT SPS activity under normal warm growth conditions (Fig. 1c & d). The increase in SPS protein and activity following exposure to low temperature was strongly inhibited in the antisps plants. Furthermore, whereas NH antisps plants had a higher activation state relative to WT (50%), this did not increase further in leaves shifted to 5 °C, confirming the antisense repression of the cold-induced increase in SPS activity in these leaves.
The plants over-expressing SPS showed more than eight times higher activity compared with WT at 23 °C (Fig. 1d). This increase in activity was confirmed by the results from Western blotting (Fig. 1c). The increase in SPS activity and protein following cold exposure was comparable to the increase in activity from the endogenous SPS gene. The activation state of SPS in the over-sps leaves was 30, 40 and 51% for NH, 10 d and 5 °C cold-developed leaves, respectively. This response is different from that shown by WT, the over-sps plants maintaining a higher activation state at low temperature, reflecting the different effect of temperature on the activation of the maize enzyme.
Total protein, chlorophyll and specific leaf fresh weight
In the WT and all three transgenic lines, leaf development at 5 °C resulted in a three-fold increase in specific dry weight per leaf area and a reduction in water content from approximately 90 to 85% (Table 1). Chlorophyll (Chl) content was not consistently affected by the low temperature exposure, although there was generally a slight decrease in Chl per unit fresh weight (FW) in all lines except the antisps plants. In the antisps plants there was a recovery from the low amounts of Chl in warm-grown leaves to amounts similar to the other lines in cold-developed leaves. Total protein per unit leaf FW increased two-fold following leaf development at 5 °C. The amount of protein in the leaves was lower in both antisense lines under all three different sample conditions but it was reduced in particular in the cold-developed leaves of the antifbp plants (Table 1).
Table 1. Physical characteristics of control leaves (23 °C), 10-day-shifted leaves (10 d) and cold-developed (Dev.) leaves
Values represent the mean (± SD) for four different leaves collected from different plants.
Specific leaf weight (mg DW cm−2)
1.5 ± 0.1
2.2 ± 0.1
4.5 ± 0.1
1.5 ± 0.1
2.3 ± 0.1
4.5 ± 0.1
1.4 ± 0.1
2.1 ± 0.2
4.4 ± 0.1
1.4 ± 0.1
2.0 ± 0.1
4.4 ± 0.1
Water content (%)
91 ± 0.4
89 ± 0.9
84 ± 0.5
91 ± 0.3
88 ± 0.5
85 ± 0.7
92 ± 0.3
89 ± 0.8
84 ± 0.5
91 ± 0.4
89 ± 0.4
84 ± 0.6
Chl (mg g−1 FW)
1.58 ± 0.03
1.46 ± 0.10
1.43 ± 0.10
1.51 ± 0.04
1.45 ± 0.10
1.43 ± 0.04
1.24 ± 0.06
1.13 ± 0.06
1.59 ± 0.16
1.62 ± 0.06
1.37 ± 0.04
1.44 ± 0.15
Protein (mg g−1 FW)
12.1 ± 0.4
13.3 ± 0.5
25.0 ± 1.0
9.6 ± 0.6
11.9 ± 2.6
20.6 ± 2.4
6.4 ± 0.9
8.6 ± 2.0
21.3 ± 2.4
6.8 ± 0.4
8.1 ± 0.9
15.6 ± 2.9
Proline FW (µmol g−1)
1.0 ± 0.2
7.0 ± 0.4
12.5 ± 2.2
1.3 ± 0.1
6.6 ± 1.6
9.2 ± 1.7
1.4 ± 0.2
7.7 ± 1.4
7.3 ± 2.0
0.8 ± 0.1
6.3 ± 0.4
12.0 ± 2.4
Photosynthesis was measured on attached leaves using an open gas exchange system. Figure 2a–c show light response curves at ambient CO2 (350 µmol mol−1) and ambient temperature (Fig. 2a and 23 °C; Fig. 2b and 5 °C; Fig. 2c and 5 °C). At 23 °C and growth irradiance (150 µmol photons m−2 s−1) photosynthesis decreased from 5.3 µmol CO2 m−2 s−1 in WT plants to 4.0 and 3.8 µmol CO2 m−2 s−1 in the antisps and antifbp plants, respectively (Fig. 2a). The WT and the over-sps showed similar light response curves at 23 °C. Photosynthesis was inhibited down to similar rates in the 10-day-shifted leaves from WT and the antisense plants (2.2, 2.3 and 2.0 µmol m−2 s−1 for WT, antisps and antifbp, respectively; Fig. 2b). However, the photosynthetic rate was much less inhibited in the plants over-expressing SPS and these plants maintained a photosynthetic rate of 3.8 µmol m−2 s−1 at 5 °C (Fig. 2b). Following leaf development at 5 °C, photosynthesis recovered strongly in WT and the over-sps plants to a rate at 5 °C that was similar, at the growth irradiance, to the rate of the control leaves measured at 23 °C. The light response curves at 5 °C were similar for WT and the plants over-expressing SPS (Fig. 2c). Both antisense lines, on the other hand, showed only minimal recovery of photosynthesis at 5 °C (Fig. 2c). The limited capacity to recover photosynthesis at 5 °C in both antisense lines confirms the importance of the up-regulation of the cytosolic pathway for photosynthetic acclimation to low temperatures.
Partitioning of newly fixed carbon
Rates of soluble sugar and starch synthesis were investigated by supplying saturating (5%) 14CO2, for 20 min at 23 °C, and for 40 min at 5 °C, to leaf discs in saturating light (600 µmol m−2 s−1). In 23 °C plants the flux of newly fixed 14C into soluble sugars increased significantly (P < 0.05) in the over-sps plants and decreased significantly in both the antisps (P < 0.01) and antifbp (P < 0.001) antisense lines (Fig. 3a). When the plants were shifted to 5 °C for 10 d, all lines showed a strong reduction in flux of newly fixed 14C into soluble sugars, as could be expected from the reductions shown for photosynthesis (Fig. 2). However, the over-sps plants maintained a significantly (P < 0.01) higher flux of newly fixed carbon into the soluble sugars (Fig. 3a), correlating with the ability of these plants to maintain higher photosynthetic rates at 5 °C. After new leaves developed at 5 °C, all lines increased the rate of incorporation of newly fixed 14C into the soluble sugars to rates exceeding the flux in the respective 23 °C leaves. However, even after the development of new leaves at 5 °C, both the antisps (P < 0.05) and antifbp (P < 0.001) antisense lines showed significantly lower fluxes of fixed carbon into soluble sugars (Fig. 3a).
When we compare the ratio between 14C partitioned into the soluble fraction (soluble sugars plus organic acids and phosphorylated intermediates) and the insoluble fraction (predominantly starch), warm-grown control WT and antisps leaves allocated approximately 55% of their newly fixed carbon into the soluble fraction (Fig. 3b). In contrast, the over-sps plants partitioned significantly (P < 0.001) more, and the antifbp line partitioned significantly (P < 0.001) less, into the soluble fraction, with the antifbp line allocating as much as 60% of the newly fixed carbon into the insoluble starch fraction (Fig. 3b). After the plants were shifted to 5 °C for 10 d, the pattern of partitioning of newly fixed carbon between starch and soluble carbohydrates remained similar to the warm-grown treatment despite the reduction in overall flux at 5 °C. However, in WT leaves that developed at 5 °C, the partitioning of newly fixed carbon into starch decreased to less than 30% and the ratio between soluble sugars and starch increased from 1.3 to 2.2 (Fig. 3b). This change in carbon partitioning following the development of new leaves at 5 °C has been reported previously in WT Arabidopsis leaves (Strand et al. 1999). Both the over-sps and the antisps plants also showed increased partitioning into soluble sugars, with the over-sps plants maintaining a consistent and significant (P < 0.01) increase relative to WT. In contrast, leaf development at 5 °C shifted the partitioning of newly fixed carbon only slightly towards sucrose synthesis in the antifbp plants and partitioning remained significantly (P < 0.001) lower in the antifbp plants than in WT and the transgenic SPS lines (Fig. 3b). These changes in the ability to modify partitioning of newly fixed carbon in response to low growth temperature are consistent with the effects of the introduced transgenes as shown previously in multiple transgenic lines at warm growth temperatures (Signora et al. 1998; Strand et al. 2000).
Diurnal accumulation of fixed carbon into different carbohydrate pools
Samples for carbohydrate analysis were collected at the end of the dark period and at intervals during the 8 h light period. Leaf carbohydrate levels at these different times reflect the balance between synthesis and mobilization. Under normal warm growth conditions, WT Arabidopsis preferentially accumulates starch during the day, with a ratio between soluble sugars and starch being approximately 0.3 at the end of the photoperiod. The starch that accumulates during the photoperiod is then almost fully mobilized during the following night (Fig. 4b). In the antifbp plants, the amount of Suc that accumulated during the day was 50% lower than in WT leaves, whereas the starch pool that accumulated was 1.6-fold larger (Fig. 4a & b) [see also (Strand et al. 2000)]. Increased accumulation of starch is to be expected in plants with reduced rates of Suc synthesis and it has been reported previously in potato and in Flavaria with reduced cFBPase activity (Sharkey et al. 1992; Zrenner et al. 1996). In the antisps plants the amount of Suc was also 50% lower at the end of the day in comparison with WT and, unexpectedly, the antisps plants also accumulated slightly less starch than WT (Fig. 4a & b) (see also Strand et al. 2000). In the antisps plants, total leaf carbohydrate was lower at the end of the photoperiod than in WT plants or the antifbp plants due to this simultaneous decrease of sugars and starch (Fig. 4e). The plants over-expressing SPS accumulated slightly more Suc but similar amounts of starch at the end of the day relative to WT (Fig. 4a & b).
When the plants were shifted to 5 °C the amount of Suc that accumulated increased slightly but the amount of starch at the end of the photoperiod was largely unchanged (Fig. 4f & g). The total amount of carbohydrate that accumulated was similar for WT and all the transgenic lines (Fig. 4j). However, unlike control 23 °C-grown leaves, the carbohydrate pools in the 10-day-shifted leaves, particularly the large starch pool, were not fully mobilized during the night. Thus, the accumulation of carbohydrates in the leaves shifted to low temperature was due to limited mobilization and export and not to increased synthesis. Interestingly, the over-sps plants showed a slight increase in the amount of Suc that accumulated and a similar amount of starch, despite the much higher in situ photosynthetic rate (Fig. 2b). The over-sps plants also showed more complete nocturnal mobilization of the starch pools (Fig. 4g). These data indicate that the over-sps plants are much better at maintaining carbon export from their source leaves during exposure to low temperature, suggesting a positive interaction between Suc production and the capacity for mobilization in this transgenic line. Similar results were obtained when the over-sps plants were grown at high CO2 where, in contrast to WT, the over-sps plants did not accumulate carbohydrates or show reduced photosynthetic capacity (Signora et al. 1998). These data suggest that the over-sps plants are better able to supply the developing sink tissues with carbon during the acclimation phase, something that may be important for the development of freezing tolerance during the expansion of a new rosette.
Cold-developed WT and over-sps plants showed a strong increase in the Suc pool and a substantial part of this pool of Suc is not mobilized during the night (Fig. 4k), consistent with this additional Suc playing a cryoprotective role. Cold-developed leaves from WT and over-sps plants also showed a return to mobilization of the starch pool overnight (Fig. 4l). Following the growth and development of new leaves at 5 °C, the antisense plants showed reduced pools of Suc in comparison with WT and the over-sps plants (Fig. 4k), and minimal mobilization of the starch pool during the night (Fig. 4l). Furthermore, the antisense plants accumulated very little total carbohydrate during the day (Fig. 4o), which correlated with the limited capacity of these antisense lines to recover photosynthetic capacity.
Under normal warm growth conditions there were only small changes in the freezing tolerance of the different transgenic lines (Fig. 5a). However, following a shift to 5 °C the over-sps plants significantly (P < 0.001) increased TEL50 relative to WT (−9.1 versus −7.2 °C). In contrast, both antisense lines developed significantly (P < 0.001) less freezing tolerance after cold exposure for 10 d, only showing a TEL50 of −5.8 and −5.2 °C for the antisps and antifbp lines, respectively (Fig. 5b). After new leaves developed at 5 °C, the over-sps plants again developed significantly (P < 0.001) greater freezing tolerance than WT (TEL50 of −11.4 °C). The development of new leaves in the cold improved the freezing tolerance of the antisense lines but even after leaf development both antisense lines showed less freezing tolerance than WT (antisps TEL50 of −9.1, P < 0.01; antifbp TEL50 of − 7.6, P < 0.001) (Fig. 5c).
To test the relation between altered partitioning of photosynthetically fixed carbon into Suc in the different transgenic lines and freezing tolerance, we plotted the changes in freezing tolerance against the sucrose and soluble sugar content present at the time the leaves were harvested for the freezing tests (Fig. 6). Regardless of whether freezing tolerance is plotted against total soluble sugars (Fig. 6a) or sucrose alone (Fig. 6b) we find a strong correlation between the TEL50 estimated from the electrolyte leakage curves and soluble sugar content. Furthermore, this relation holds not only for WT and all transgenic lines but also for the different growth regimes. These data confirm that modifying the activities of the sucrose biosynthetic enzymes, either by enhancing the pathway by over-expressing SPS or by causing a bottleneck in the pathway with either antisense cFBPase or antisense SPS plants, results in changes in the accumulation of sucrose during cold acclimation that significantly enhances or limits the development of freezing tolerance in Arabidopsis leaves, respectively. Clearly, the cold-induced up-regulation of the cytosolic pathway is important for the development of freezing tolerance and an increased capacity of the cytosolic pathway for sucrose synthesis improves both the rate of cold acclimation and also the eventual freezing tolerance of fully cold-acclimated Arabidopsis leaves.
COR expression and proline accumulation
One concern with experiments of this type is that the position in which the transgene was inserted within the Arabidopsis genome will have pleiotropic effects on other low-temperature acclimation responses. To assess this we analysed, as additional factors implicated in cold acclimation and the development of freezing tolerance, the expression of the various COR genes and the accumulation of proline.
The COR proteins have been shown to play an important role in the development of freezing tolerance (Jaglo-Ottosen et al. 1998; Gilmour et al. 2000; Jaglo et al. 2001). When the expression of the COR genes, COR6.6, COR15a, COR47/RD22 and COR78/RD29A/LTI78 was investigated we found that all four were strongly induced after only 1 d exposure to 5 °C and the expression levels remained high in WT and all three transgenic lines after 10 d at 5 °C (Fig. 7). Furthermore, as we have shown previously (Hurry et al. 2000), the amount of transcript for all four COR genes declined in the WT and transgenic lines when new leaves developed at 5 °C (Fig. 7).
Proline accumulation has also been shown to be a key plant response to a number of abiotic stresses, including low temperature (Rudolph & Crowe 1985; Nanjo et al. 1999b). We also chose to assay proline as an indicator of possible flow-on effects to general plant metabolism of altering sucrose synthesis. Samples for proline analysis were collected at the end of the day. Proline was present at low levels in all warm-grown plants, it increased markedly in leaves shifted to 5 °C for 10 d and the highest amount was found in cold-developed leaves, in which there was an over 10-fold increase in the amount of proline compared with the control warm-grown leaves (Table 1). Proline increased somewhat less following cold development in the antisps plants but the accumulation of proline in response to low temperature was otherwise similar to WT in all transgenic lines.
In Arabidopsis, the recovery of photosynthesis and the expression of full freezing tolerance by leaves that develop at low temperatures are strongly correlated with a reprogramming of carbon metabolism, involving a change in carbon partitioning in favour of sucrose synthesis (Strand et al. 1997, 1999; Hurry et al. 2000). In the experiments reported here we test the hypothesis that the up-regulation of the sucrose biosynthetic pathway during cold acclimation is an essential element for the development of freezing tolerance. We use transgenic plants of the same ecotype but with specific modifications to sucrose metabolism. The different transgenic lines have either an increased expression of SPS and an increased capacity for sucrose synthesis (Signora et al. 1998) or decreased sucrose synthesis because of either reduced cytosolic FBPase activity or reduced SPS activity (Strand et al. 2000).
When we compare the responses of warm-grown WT and transgenic Arabidopsis plants shifted to low temperature for 10 d, photosynthesis was significantly less inhibited at 5 °C in plants over-expressing SPS than it was in WT plants. The elevated photosynthetic rate in plants over-expressing SPS also translated into a significant (two-fold) increase in the flux of newly fixed carbon into soluble sugars and a significant increase in the soluble sugar/starch ratio (Fig. 3). Furthermore, in the new leaves that developed at 5 °C, the recovery of photosynthesis that has been shown in WT plants (Strand et al. 1997, 1999) and the over-sps line was blocked by the antisense repression of either cFBPase or SPS. Similarly, the developmentally induced increase in flux of newly fixed carbon into soluble sugars was significantly reduced in both antisense lines, with the antifbp plants showing both the strongest reduction in soluble sugar synthesis and a significantly lower ratio of soluble sugar to starch synthesis. These data strongly support the hypothesis, developed from studies of both dicotyledonous (Guy et al. 1992; Hurry et al. 1995; Strand et al. 1997) and monocotyledonous (Hurry et al. 1994; Savitch, Gray & Huner 1997) plants that the recovery of photosynthetic capacity at low temperatures is strongly dependent on the induction of the sucrose biosynthetic pathway.
The capacity for soluble sugar synthesis in WT and the three different transgenic lines correlated well with their respective abilities to cold harden (Fig. 6). In these experiments WT Arabidopsis plants increased their freezing tolerance down to temperatures around −7 °C after exposure to 5 °C for 10 d, whereas the over-sps plants developed significantly (P < 0.001) increased freezing tolerance down to around −9 °C. In contrast, both the antifbp and antisps plants, which suffered from a reduced supply of soluble carbohydrates for metabolism when they began to acclimate to 5 °C, could only reach a freezing tolerance of approximately −5 °C after 10 d in the cold (Fig. 5). Furthermore, the development of freezing tolerance in new leaves that expanded at 5 °C was significantly enhanced in the over-sps plants, which maintained consistently higher partitioning of carbon into soluble compounds (soluble sugars plus organic acids and phosphorylated intermediates), but was significantly impaired in both antisense lines, relative to WT. The reduced ability to cold harden was shown most strongly by the antifbp plants, in which the flux of newly fixed carbon into soluble sugars and the shift in partitioning towards sucrose synthesis was most strongly inhibited. These data suggest that the acclimation of photosynthesis and the shift in partitioning toward sucrose biosynthesis is probably important for providing energy for other critical mechanisms of freezing tolerance, such as the synthesis of specific stress-tolerance proteins and lipid turn-over, in addition to supplying cryoprotective or osmotically active sugars. The data from the transgenic line with increased SPS activity suggests that this energy supply aspect may be particularly important in developing sink tissues, and that not only sucrose synthesis but sucrose transport mechanisms may be important for the long-term maintenance of freezing tolerance over the winter.
As one would expect from a quantitative trait such as freezing tolerance, cold acclimation is strongly inhibited in the antisense plants but it is not eliminated. As we outline above, the accumulation of the COR proteins (Jaglo-Ottosen et al. 1998; Gilmour et al. 2000; Jaglo et al. 2001) and of other compatible solutes such as proline (Rudolph & Crowe 1985; Nanjo et al. 1999b) are known to be important for the development of freezing tolerance. In the experiments we report here the four COR genes, COR6.6, COR15a, COR47/RD22 and COR78/RD29A/LTI78, were all strongly induced after only 24 h exposure to 5 °C and the expression levels remained high after 10 d in the cold (Fig. 7). The similar changes in expression of the cold-induced COR transcripts in WT and the transgenic plants, and the similar accumulation patterns for proline (Table 1) shows that these important low-temperature responses have not been inhibited in any of our transformants and that the differing ability to increase freezing tolerance was therefore related to the ability of the different transgenic lines to modulate carbon metabolism. This was especially true of both antisense lines, indicating that although low sugar supply does not prevent the low temperature-induced increase of COR transcripts it does interfere with the development of freezing tolerance. These data strongly support our initial hypothesis that the acclimation of carbon metabolism is an essential prerequisite for the full expression of freezing tolerance by cold acclimating Arabidopsis. How low sugars interact with other low-temperature responses is not known but it is clear that understanding how the different mechanisms interact and modulate the acclimation potential of the plant will be essential if we are to develop effective strategies for improving the freezing tolerance of crop plants (Stitt & Hurry 2002).
In this light, it is interesting to note that although the low-temperature response of the COR genes was not affected in the transgenic lines, the expression of COR6.6 and COR47/RD22 was induced in both the antifbp and the antisps plants under normal warm growth conditions (Fig. 7). These data indicate that there is a possible interaction between the metabolic changes in the antisense transformants and COR gene expression under warm growth conditions. The fact that both antisense lines show this response indicates that it is unlikely to be due to positional effects arising from the insertion of the transgenes. Little is known about the function of these two proteins at low temperatures but both COR6.6 and COR47/RD22 are also induced by drought (Wang et al. 1995; Iwasaki et al. 1997; Liu et al. 1998). Our results suggest that they may also be induced in response to altered osmotic potential in the antisense lines, even under warm growth conditions, and that their expression levels may be elevated in the antisense lines in response to the low production of soluble sugars.
Taken together, data from these different transgenic lines and different growth treatments yielded a strong relation between the accumulated pools of soluble sugars and freezing tolerance (Fig. 6) that is independent of any changes to the induction of cold-stress proteins (Fig. 7) or the accumulation of compatible solutes such as proline (Table 1). This relation was strongest for sucrose (Fig. 6b) and had two characteristics making it particularly robust. First, the data for the sense and antisense transformants bracket the WT data in both low-temperature treatments; demonstrating that the cold acclimation responses we observe are consistent with the roles of these enzymes in sucrose biosynthesis and leaf sucrose content and could not be the result of pleiotropic positional effects. Second, the data from both low-temperature treatments overlap rather than form distinct groupings; demonstrating that sucrose content and electrolyte leakage are dependent variables in these experiments. Thus, these experiments demonstrate that it is possible to either significantly enhance or reduce freezing tolerance in transgenic Arabidopsis by modifying the flux of carbon into sucrose and that the change in freezing tolerance correlates with the effect of the introduced transgenes on the sucrose content of the leaves. Modification of sucrose metabolism therefore represents an additional approach that will have benefits both for the development of freezing tolerance and over-wintering, and for the supply of exportable carbohydrate to support continuing growth at low temperatures.
This work was supported by grants from the Swedish Council for Forestry and Agricultural Research (V.H), the Centre for Forest Biotechnology and Chemistry (V.H., P.Ga.) and from the Swedish Technical Research Council (P.Gu., P.Ga.).
Received 20 May 2002; received in revised form 17 September 2002; accepted for publication 20 September 2002