Overexpression of a proton-coupled vacuolar glucose exporter impairs freezing tolerance and seed germination



  • Arabidopsis vacuoles harbor, besides sugar transporter of the TMT-type, an early response to dehydration like 6 (ERDL6) protein involved in glucose export into the cytosol. However, the mode of transport of ERDL6 and the plant's feedback to overexpression of its activity on essential properties such as, for example, seed germination or freezing tolerance, remain unexplored.
  • Using patch-clamp studies on vacuoles expressing AtERDL6 we demonstrated directly that this carrier operates as a proton-driven glucose exporter.
  • Overexpression of BvIMP, the closest sugar beet (Beta vulgaris) homolog to AtERDL6, in Arabidopsis leads surprisingly to impaired seed germination under both conditions, sugar application and low environmental temperatures, but not under standard conditions.
  • Upon cold treatment, BvIMP overexpressor plants accumulated lower quantities of monosaccharides than the wild-type, a response in line with the reduced frost tolerance of the transgenic Arabidopsis plants, and the fact that cold temperatures inhibits BvIMP transcription in sugar beet leaves. With these findings we show that the tight control of vacuolar sugar import and export is a key requisite for cold tolerance and seed germination of plants.


In plants, sugars function as precursors and reserves for cellular energy, as carbon sources for synthesis of most metabolic intermediates, as building blocks for starch and cellulose biosynthesis, as well as osmolytes preventing cellular defects caused by cold, drought or salt stress (Wanner & Junttila, 1999; Heldt, 2005). In agronomically relevant plant species such as sugar beet and sugarcane, sucrose can accumulate up to 20% of fresh weight. In storage tissues from some other agronomic species, such as wine grapes, glucose represents the major sugar type, while high concentrations of fructose (if not in form of polymeric fructans) are typically absent from storage tissues (John, 1992; ap Rees, 1994). Due to these important functions it is not surprising that sugar concentrations in plant cells are sensed and govern gene expression (Rolland et al., 2002).

During the day triose phosphates produced in chloroplasts are exported into the cytosol and serve as precursors for sucrose (Kruckeberg et al., 1989; Neuhaus et al., 1990; Flügge, 1999), which is either exported via the phloem (Ruiz-Medrano et al., 2001) or stored in the central vacuole (Martinoia et al., 2012). Thereby, mesophyll cell vacuoles act as an intermediate storage compartment for sugars (Martinoia et al., 2007). Especially under cold conditions, sucrose, glucose and fructose represent the main sugars in leaf-mesophyll vacuoles (Schulze et al., 2012) which they enter via specific transport proteins (Neuhaus, 2007). Vacuolar sucrose import occurs by facilitated diffusion (Kaiser & Heber, 1984) or via sucrose/proton antiporters energized by the proton-motive force (McRae et al., 2002; Martinoia et al., 2007). Only recently, a candidate protein for vacuolar sucrose loading has been identified at the molecular level. The tonoplast-located carriers TMT1 and 2 appear to serve dual functions as they can mediate the uptake of glucose and fructose (Wormit et al., 2006), as well as sucrose (Schulz et al., 2011). In planta the role of TMT-mediated vacuolar monosaccharide import is apparent because Arabidopsis mutants with increased TMT activity produce a higher seed biomass (Wingenter et al., 2010). Apart from the proton-motive force, TMT activity is regulated by a mitogen-activated protein-triple kinase named VIK1 (Wingenter et al., 2011).

In order to allow a dynamic sugar homeostasis, vacuolar sugar exporters are also required. SUC4-type sucrose carriers such as AtSUC4 from Arabidopsis catalyze proton-coupled sucrose export (Schulz et al., 2011). Very recently, a member of the large SWEET family of sugar porters (Chen et al., 2010), namely SWEET17, has been identified as the first and so far sole tonoplast-located fructose-equilibrating transporter (Chardon et al., 2013). In addition to this, two independent vacuolar glucose exporters, ESL1 and ERDL6, have been identified (Kiyosue et al., 1998; Yamada et al., 2010; Poschet et al., 2011). ESL1 (early response to dehydration six like 1) is mainly expressed in pericycle and xylem parenchyma cells, and the corresponding mRNA accumulates under conditions of drought stress (Yamada et al., 2010). Arabidopsis plants lacking ERDL6 (early responsive to dehydration six like 6) show increased concentrations of glucose in leaves and 10% increase in seed weight (Poschet et al., 2011), indicating that controlled glucose export via this carrier is critical for seed development. The sugar-beet carrier BvIMP is a closest homolog to ERDL6 (Chiou & Bush, 1996; Poschet et al., 2011) and overexpression of BvIMP in Arabidopsis affects glucose concentrations under cold conditions (Poschet et al., 2011).

However, so far neither the transport mode of ERDL6 nor the gross impact of stimulated vacuolar glucose export on plant properties is known. In order to learn more, we studied the transport characteristics of AtERDL6 using the patch-clamp technique on isolated vacuoles (Schulz et al., 2011). In addition, we overexpressed BvIMP in Arabidopsis (see Poschet et al., 2011; to prevent cosuppression often seen when homologous genes are overexpressed) to study impacts on germination pattern and freezing tolerance, representing processes highly affected by cellular sugar concentrations (Hanson & Smeekens, 2009; Schulze et al., 2012).

Materials and Methods

Plant material and growth conditions

Arabidopsis wild-type (WT) (Arabidopsis thaliana (L.) Heynh., ecotype Col-0), the two BvIMP overexpression lines BvIMP6 and BvIMP8 (Poschet et al., 2011) and Beta vulgaris L. cv belladonna (KWS Saatgut AG, Einbeck, Germany) were grown on potting soil (ED-73, Patzer GmbH, http://www.einheitserde.de) in a growth chamber under short-day conditions (10 h : 14 h, light : dark) at 22°C or 4°C (cold conditions), and 125 μmol quanta m−2 s−1. For germination experiments, seeds were surface sterilized and incubated for 24 h in the dark at 4°C for stratification. Germination rates were determined on 0.5× MS (pH 5.7, KOH) plates, supplemented with or without sugars from triplicate assays with at least 20 seeds per plant line.

Subcellular localization studies on BvIMP

For the subcellular localization of BvIMP in Arabidopsis we generated a BvIMP-GFP fusion construct using the GATEWAY specific destination vector pK7FWG2.0 (Karimi et al., 2002). For this the cDNA of BvIMP was amplified by PCR using the gene-specific primers BvIMP_gw_fwd (5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTAATGAGTTCAGATTCAGAA-3′) and BvIMP_gw_rev-stop (5′-GGGGACCACTTTGTACAAGAAAGCTGGGTAGGCTCTTCTGAAGGACCA-3′) harboring the attB1 and attB2 sites, then cloned via BP reaction into pDONRZEO (Invitrogen) and via LR reaction into the plasmid pK7FWG2.0. Transformation of Arabidopsis mesophyll protoplasts was performed exactly as described in Yoo et al. (2007). A Leica TCS SP5 II confocal laser scanning microscope (http://www.leica-microsystems.com) was used for imaging. All pictures were taken using a Leica HCX PL APO 63·/1.20 w motCORR CS objective with a VIS-Argon laser suitable for GFP constructs (488 nm/495–520 nm).

Gene expression analysis

RNA of Arabidopsis WT, BvIMP overexpression lines or sugar beet plants was extracted from frozen leaves or leaf discs using the NucleoSpin® RNA Plant Kit (Machery-Nagel, Düren, Germany) according to the manufacturer's advice. The iScript cDNA Synthesis Kit (Bio-Rad, Munich, Germany) was used for synthesis of cDNA. Quantitative real-time PCR for evaluation of the expression patterns of CAB1 and BvIMP was performed as described earlier (Wormit et al., 2006). Primers for quantification of BvIMP were: BvIMP_RT_f (5′-AGAATTTGAGGTTTCCTTGC-3′) and BvIMP_RT_r (5′-GCAACAGATCTCTTGATTTC-3′). At4 g26410 was used as a reference gene for transcript normalization (Czechowski et al., 2005), while for normalization of the transcript level in sugar beet ubiquitin-conjugating enzyme no. 9 was used with the following primers: BvUBC9fwd (5′-AAGGAGCAGTGGAGTCCTGC-3′) and BvUBC9rev (5′-TGTGTCCAGCTCCTTGCGG-3′).

Metabolite extraction and quantifications

For isolation of sugars, 100 mg plant material was frozen in liquid-nitrogen and ground, added to 1 ml water, mixed thoroughly and heated for 15 min at 95°C. After centrifugation for 10 min at 18 000 g the supernatant was used for metabolite quantification by ion chromatography on a HPLC-RCX-30 7 μm 4.6 × 250 mm column (Hamilton, http://www.hamiltoncomp.com), using an 871 IC compact device (Metrohm-Switzerland, Herisau, Switzerland) followed by amperometric quantification. NaOH (0.15 M) was used as the mobile phase with a flow rate of 0.5 ml min−1 and pressure was set to 9.5 MPa at 27°C.

Electrical conductivity measurements

The electrical conductivities of frozen leaf tissues from WT, BvIMP overexpressing and tmt1-2 knockout lines were assessed as described in Ristic & Ashworth (1993) with minor modifications as followed: three-week-old plants were acclimated to cold (4°C) for 4 d and fully expanded leaves were taken from 10 plants, and placed in glass tubes (1 leaf per tube) containing 2 ml deionized water. The tubes were transferred to a freezing bath set at 2°C for 1 h followed by a reduction of 2°C h−1 to −6°C. Freezing of the water around the leaf tissue was initiated at −2°C by the addition of ice chips to each tube. After the last step, tubes were taken from the bath and thawed overnight by gently shaking in a cold room (4°C). After thawing, 2 ml deionized water was added to the tubes and gently shaken for 1 h at room temperature. The total electrical conductivity was quantified at room temperature using a WTW LF521-conductivity meter (Wissenschaftlich-Technische Werkstätten GmbH, http://www.wtw.de), after boiling for 2 h and shaking overnight at 4°C.

Patch-clamp recording on isolated plant vacuoles

Arabidopsis thaliana tmt1-2 and aterdl6/tmt1-2 mutant plants were cultivated on soil in a growth chamber at 8 : 16 h, day : night regime, 22 : 16°C, day : night temperature and a light intensity of 125 μmol m−2 s−1. Isolation of leaf-mesophyll protoplasts and subsequent release of vacuoles of 5- to 8-wk-old plants were performed as described previously (Beyhl et al., 2009). According to the convention for electrical measurements on endomembranes (Bertl et al., 1992), patch-clamp experiments were accomplished in the whole-vacuole configuration as described by Schulz-Lessdorf & Hedrich (1995) and Ivashikina & Hedrich (2005). To gain patch pipettes with a resistance in the range 1.5–3 MΩ, Kimax-51 glass capillaries (Kimble Products, Vineland, NY, USA) were processed. Macroscopic currents were recorded at a sample acquisition rate of 2 ms with an EPC-7 patch-clamp amplifier (HEKA Electronics, Lambrecht, Germany) and low-pass filtered at 30 Hz. After digitization of the recorded current traces via an ITC-16 computer interface (Instrutech Corp., Elmont, NY, USA), the data were stored on a computer.

Offline analysis of the acquired data was carried out with the software programs Pulse (HEKA Electronics, http://www.heka.com/) and IGORPro (Wave Metrics Inc., Lake Oswego, OR, USA). Current amplitudes of individual vacuoles were normalized to the whole-vacuolar membrane capacitance (Cm) to allow quantitative comparison of macroscopic current magnitudes among the different organelles. The whole vacuole configuration was established using symmetric solute compositions within the pipette (vacuolar lumen) and bath solution (cytosol) containing 200 mM KCl, 2 mM MgCl2 and 1 mM CaCl2. The pH of each solution was adjusted to 7.5 with 10 mM HEPES-Tris. For macroscopic resolution of the glucose/proton symport activity, the pH of the bath medium was exchanged to pH 5.5 (10 mM MES-Tris) before the glucose stimulation of the vacuoles clamped to 0 mV. For glucose treatment a concentration of 50 mM was used which was manually applied via an application system as described previously (Schulz et al., 2011).


Vacuolar BvIMP overexpression affects sugar signaling

When characterizing the BvIMP overexpressor lines 35S:BvIMP6 and 35S:BvIMP8 grown under liquid-culture conditions, we found that these mutants exhibit reduced glucose concentrations (Poschet et al., 2011). This fact indicates that vacuolar glucose storage capacity and homeostasis has been impaired in these plants.

In order to examine whether overexpression of BvIMP alters subcellular sugar distribution, we analyzed the expression of the sugar-sensitive reporter gene CAB1. Transcription of CAB1, encoding the chlorophyllab binding protein 1, is negatively controlled by the cytosolic sugar concentration (Koch, 1996; Wingenter et al., 2010). When we incubated leaf discs from 5-wk-old Arabidopsis plants in 1% glucose or fructose, respectively, CAB1 mRNA was reduced in the BvIMP overexpression lines when compared to wild-type (WT) plants. Transcript levels within lines 6 and 8 were 1600 and 1400, while CAB1 transcripts in the WT were up to 1900 relative to the house-keeping control gene. Upon application of 1% glucose or fructose, respectively, CAB1 transcripts decrease to 800–900 in BvIMP overexpressor and 1100–1200 in corresponding WT leaf discs (Fig. 1). These results can be taken as the first hint of a stimulated vacuolar sugar export in mesophyll cells from both BvIMP overexpressor lines, when compared to WT leaf cells.

Figure 1.

Accumulation of CAB1 mRNA in Arabidopsis thaliana wild-type (WT) and BvIMP overexpressors in response to external sugars. Leaf discs were cut from leaves of 5-wk-old Arabidopsis wild-type (gray bars), BvIMP overexpressor line 6 (white bar with black dots) and line 8 (black bars with white dots) and transferred to Petri dishes containing 3 mM MES buffer (pH 5.7/KOH) with 0.5× MS medium (control, ctrl) or supplemented with glucose (glc) or fructose (fru). Dishes were shaken for 16 h in the dark on a rotary shaker (40 rpm). Transcript levels of the CAB1 gene were analyzed by qRT-PCR using At4 g26410 as reference gene for transcript normalization. (Mean values of three biological replicates ± SE given; P-values according to Student's t-test: *, < 0.05; ***, < 0.005).

ERDL6 operates as proton-driven glucose exporter

The CAB1 mRNA response in nonsugar-fed or sugar-fed leaf discs suggests that BvIMP overexpressor plants exhibit higher cytoplasmic sugar concentrations. This observation raises the question of whether increased concentrations of the sugar-beet AtERDL6 homolog BvIMP induce the vacuolar store to release glucose. Thus, to test the likely tonoplast localization of BvIMP we constructed a BvIMP-GFP fusion protein and transformed isolated Arabidopsis protoplasts (Supporting Information Fig. S1). Following lysis of GFP-positive protoplasts, green-fluorescing vacuoles were released (Fig. S1a–d). This observation unequivocally identified BvIMP as a tonoplast-located transport protein.

In order to explore the transport mode, we had previously expressed a number of ERDL6-like genes in baker's yeast as GFP fusions. While most of them exclusively localized to the vacuolar membrane (M. Büttner, unpublished data), the ERDL10-GFP expressing strain pMO10 displayed a partial mistargeting of the fusion protein to the plasma membrane (Fig. S2a,b). In turn, this localization of the ERDL10 transporter led to a partial growth complementation on glucose (Fig. S2c) of the Saccharomyces cerevisiae mutant EBY.VW4000, lacking all endogenous hexose transporters and therefore unable to grow on glucose (Wieczorke et al., 1999). Furthermore, we could demonstrate glucose uptake activity for the ERDL10-GFP expressing strain pMO10, which was repressible by the proton uncoupler CCCP. The latter observation suggests that ERDL10 mediates an active transport of glucose driven by a proton gradient (Fig. S2d).

However, to provide direct evidence for a proton-coupled transport of the ERDL6-type transporters we exploited a recently established electrophysiological approach allowing the characterization of sugar carriers in Arabidopsis mesophyll vacuoles (Schulz et al., 2011; Schneider et al., 2012). Using this method we elucidated the transport mode of ERDL6 representing the closest Arabidopsis homolog to the sugar beet transporter BvIMP (Poschet et al., 2011). In corresponding patch-clamp experiments we compared glucose-induced currents in vacuole mutants lacking the two major monosaccharide transporters TMT1 and TMT2 together (tmt1-2; Wormit et al., 2006), and triple mutants additionally lacking ERDL6 (aterdl6/tmt1-2). These experiments were performed in the so-called ‘whole-vacuole configuration’ to increase the resolution of the system by monitoring the assembled action of all ERDL6 carrier proteins, active in the vacuole in question.

The transport capacity of proton-driven sugar carriers depends on the membrane gradient of the sugar and the proton. To avoid the interference of H+/glucose currents with background proton-coupled transporters, such as the CLCs (Monachello et al., 2009), the pH gradient during the experiments remained unchanged, instead glucose was applied to the cytosolic side of the vacuolar membrane and served as an inductor of glucose-coupled proton transport.

In order to resolve ERDL6-mediated vacuolar efflux of glucose and H+ our experiments were performed in the presence of a pH gradient with pH 5.5 in the cytosol (bath medium) and pH 7.5 in the vacuolar lumen (Fig. 2a, upper panel). When 50 mM glucose was supplied to the cytosolic side of the vacuolar membrane (in the absence of luminal sugar), tmt1-2 vacuoles showed an upward deflection in the current baseline of c. 1.6 pA/pF (Fig. 2b,c). This macroscopic current response (Fig. 2b,c) reflects a glucose-triggered proton flux from the cytosol into the vacuolar lumen pointing to underlying glucose–proton symport activity. In comparison to the tmt1-2 double mutant, vacuoles prepared from the triple mutant aterdl6/tmt1-2 showed a significant decrease in the glucose-induced current response by c. 50%, indicating that ERDL6 contributes substantially to the proton-coupled glucose symport activity in isolated vacuoles (Fig. 2b,c). Thus, these results suggest that ERDL6 functions as an electrogenic transporter that, under in vivo conditions, mediates glucose export from the vacuole in symport with protons.

Figure 2.

Proton-coupled glucose transport catalyzed by Arabidopsis thaliana ERDL6. (a) Cartoon shows the experimental conditions, with pH 7.5 in the vacuolar lumen (pipette) and pH 5.5 in the cytosol (bath) in the absence (left) and presence (right) of glucose in the cytosol. (b) Typical macroscopic current responses of Attmt1-2 (left) and Aterdl6/tmt1-2 (right) knockout mutants to cytosolic glucose application. The period of glucose treatment is indicated by gray bars. (c) Glucose-induced changes in current density determined from individual vacuoles of Attmt1-2 (black bar, = 9) and Aterdl6/tmt1-2 (white bar, = 10) plants (means ± SE). The applied glucose concentration in (a–c) was 50 mM.

BvIMP overexpression impairs germination

In order to test whether increased vacuolar monosaccharide export also affects germination efficiency in the absence of additional sugar supply, we allowed WT seeds and those of both 35S:BvIMP lines to germinate. Under control conditions (no additional sugar) the same germination efficiency of nearly 100% completed within 2 d was found for all three plant lines tested (Fig. 3a). However, in the presence of 2.5% glucose the germination efficiency decreased and WT seeds required c. 6 d to reach a maximal germination efficiency of c. 92% (Fig. 3b). Interestingly, seeds from the two 35S:BvIMP lines showed an even stronger delay in their germination efficiency. This delay can, for example, be seen at day 4, where germination of both 35S:BvIMP lines reached only c. 60% of that observed on corresponding WT seeds (Fig. 3b). A similar difference in the germination of BvIMP overexpressors was obtained when seeds were challenged with fructose: in the presence of 1.5% fructose, at day 2, seeds from 35S:BvIMP6 and 35S:BvIMP8 plants gained only 35% of the germination efficiency exhibited by WTseeds (Fig. 3c).

Figure 3.

Germination pattern of Arabidopsis thaliana wild-type (WT) and BvIMP overexpressor seeds in the presence of different sugars. (a) Control; (b) 2.5% glucose; (c) 1.5% fructose. At least 20 seeds from WT (gray diamonds) and BvIMP overexpressors 6 (black squares) and 8 (black triangles) were surface-sterilized and after stratification for 24 h at 4°C spread on Petri-dishes with 0.5× MS medium supplemented with, or without sugars. Germination was followed over the course of 8 d and given in percentage of total seeds. Shown are mean values of three independent experiments. (d) Seeds were incubated for stratification at 4°C for 1 d and transferred for 16 h to 22°C to initiate mobilization of sugars and again transferred for 3 wk to 4°C for germination. (e) Two-leaf stadium was quantified after 5 wk at 4°C (means ± SE; ***, < 0.005)

We checked the altered germination properties under more natural conditions by allowing seeds to germinate at 4°C for up to 35 d (Fig. 3d,e). At 4°C within this time period 90% of WT seedlings developed to the two-leaf stage (Fig. 3d,e), while both BvIMP overexpressor lines developed with a markedly reduced developmental rate. After 35 d at 4°C 58% of seeds from BvIMP line 6 and just c. 45% of the seeds from BvIMP line 8 developed to the two-leaf stage (Fig. 3d,e).

BvIMP overexpressors show altered carbohydrate metabolism

In order to test in sugar beet plants whether vacuolar sugar export is adjusted during cold conditions or in the presence of sugars, we analyzed the expression of the BvIMP gene in response to low temperature or after administration of various sugars. For this approach BvIMP mRNA was quantified from either plants exposed to 4°C or from leaf discs incubated with various exogenous sugars. Leaf samples from plants previously cultured for 4 wk were exposed for 16 h to 4°C, or left for 16 h at 22°C (controls). As a result of this acclimation of sugar beet to cold, the BvIMP mRNA dropped to c. 20% of the level at 22°C (Fig. 4a). A similar but somewhat weaker downregulation of BvIMP mRNA to 55% and 66% of the water control was observed when leaf discs where incubated with glucose or sucrose, and fructose, respectively (Fig. 4b).

Figure 4.

Accumulation of BvIMP mRNA in sugar beet (Beta vulgaris) leaves in response to different stimuli. (a) Response to cold; (b) application of different sugars. Seven leaf-discs of 4-wk-old sugar beet plants were cut out (0.7 cm Ø) and transferred to Petri dishes containing 3 mM MES buffer (pH 5.7 / KOH) with 0.5× MS medium without (control) or with sugar (glc, glucose; frc, fructose; suc, sucrose). Dishes were placed on a rotary shaker and shaken for 16 h in the dark (40 rpm). The dishes for the response to cold were kept for 16 h in a cold chamber at 4°C. Data represent mean values ± SE of at least three individual experiments from different plants.

It is widely accepted that plants accumulate sugars in response to cold temperatures (Alberdi & Corcuera, 1991; Wanner & Junttila, 1999). To monitor the cold (4°C) responses of WT and BvIMP lines, we quantified, glucose, fructose and sucrose for 72 h following stimulus onset. For comparison, we included the tmt1-2 double mutant, known for its limitation in cold-induced vacuolar accumulation of sugar (Wormit et al., 2006). In WT leaves glucose accumulated to a concentration of 7.0 μmol g−1 FM and rose slowly to 8.5 μmol g−1 FM within 10 h of illumination. Interestingly, the sugar concentrations showed a tendency to oscillation. For example, during the first night at cold temperature glucose concentrations increased to 23 μmol g−1 FM, decreased to 20 μmol g−1 FM in the subsequent night phase and increased again to c. 30 μmol g−1 FM in the next light phase (Fig. 5a). Within the last night phase glucose rose to 35 μmol g−1 FM (Fig. 5a). In leaves of the two BvIMP overexpressor lines the glucose concentrations were initially significantly lower than in the corresponding WT, namely 3.3 and 5.0 μmol g−1 FM, respectively (Fig. 5a). Similar to the trend in WT plants BvIMP overexpressor plants accumulate glucose during the first 72 h of cold incubation. However, in absolute numbers glucose accumulation was always below the concentrations in WT plants. After 24 h of cold exposure mutant lines 6 and 8 contained 20 and 12 μmol glucose g−1 FM, respectively, while at 48 h both lines contained c. 22 μmol glucose g−1 FM. At the end of the cold period analyzed (72 h), line 6 contained 24 μmol glucose and line 8 contained only 19 μmol glucose g−1 FM (Fig. 5a). In contrast to WT and BvIMP overexpressors, tmt1-2 plants hardly accumulated glucose in the cold (Fig. 5a).

Figure 5.

Sugar and starch accumulation in Arabidopsis thaliana leaves during 72 h of cold incubation. (a) Glucose, (b) fructose, (c) sucrose and (d) starch (in C6). Plants of wild-type (WT, gray diamonds), BvIMP overexpressors 6 (black squares) and 8 (black triangles) and tmt1-2 double knock-out mutant (red circles) were grown under standard conditions for 6 wk and 10 h : 14 h, light : dark regimen and transferred at the end of the light period to 4°C for 72 h. At this point samples were harvested in a diurnal rhythm at the end of the night and at the end of the day. Each data point represents mean values of three biological replicates ± SE. (P-values according to Student's t test: *, < 0.05; ***, < 0.005). tmt1-2 double knock-out mutants were as described in Wormit et al. (2006).

Fructose concentrations in the different plants changed similar to those with glucose (Fig. 5b). Fructose concentrations in BvIMP overexpressor lines 6 and 8 were similar to those in WT plants at the beginning of cold treatment, 0.8 and 0.7 μmol g−1 FM, respectively, and increased after 72 h of cold treatment to c. 20 μmol g−1 FM, representing 75% of the corresponding WT concentration (Fig. 5b). As seen for glucose, tmt1-2 plants are also unable to accumulate fructose upon transfer into the cold. Even after 72 h at 4°C, tmt1-2 leaves contained only 2.4 μmol g−1 FM fructose, c. 15% of that present in WT leaves (Fig. 5b).

In contrast to glucose and fructose, the cold-induced changes in concentrations of sucrose in WT plants and both BvIMP overexpressor lines were similar. The sucrose concentration saturated 24 h after stimulus onset with a tendency to oscillate in a day–night-dependent manner (Fig. 5c). tmt1-2 plants, which are unable to accumulate monosaccharides at low temperatures (Fig. 5a,b), showed slightly higher sucrose concentrations than corresponding WT and BvIMP overexpressor lines (Fig. 5c). Their maximal sucrose concentration of c. 18 μmol g−1 FM was reached after 48 h in the cold and declined slightly to 13 μmol g−1 FM after 72 h (Fig. 5c).

BvIMP overexpressors show impaired freezing tolerance

Sugar accumulation in the cold is highly correlated with the ability to tolerate freezing temperatures (Wanner & Junttila, 1999). Thus the following question is raised: does altered sugar accumulation (Fig. 5) and altered sugar compartmentation (Fig. 1) in BvIMP overexpressor plants causally affect freezing tolerance? To answer this, we monitored the integrity of cells challenged by low temperatures. As a measure for membrane damage, we followed electrolyte leakage of leaves after exposure to freezing temperatures (Ristic & Ashworth, 1993; Zuther et al., 2004).

After 4 d acclimation at 4°C, leaf samples were taken from intact plants and incubated at a freezing temperature of −6°C. As a result, WT leaves leaked 24% of their total electrolytes, while BvIMP overexpressor lines 6 and 8 leaked 32%, and tmt1-2 plants released 38% of their charged solutes (Fig. 6).

Figure 6.

Electrical conductivity of various Arabidopsis thaliana lines. Leaves from wild-type (WT), BvIMP overexpression lines 6 and 8 and tmt1-2 double knock-out mutants were prepared after 4 d of acclimatization to cold, and subsequently incubated at −6°C. ( 8 ± SE; P-values according to Student's t-test: *, < 0.05; ***, < 0.005).


For dynamic sugar storage in plant vacuoles both import and export of corresponding solutes has to be catalyzed. Although sucrose import into the vacuole has been shown to occur via both facilitated transport and proton-driven import, the corresponding sucrose-specific carrier proteins have not yet been identified on the molecular level (Neuhaus, 2007). By contrast, sucrose export is most likely mediated by SUC4-type sucrose/proton symporters (synonym to SUT2) which have been identified and functionally analyzed in various species (Eom et al., 2011; Schulz et al., 2011; Schneider et al., 2012).

The demonstration that Arabidopsis mutants lacking the vacuolar monosaccharide transporter TMT do not accumulate glucose and fructose upon transfer into the cold (Fig. 5a,b, and Wormit et al., 2006) indicated that this carrier fulfils an important function for subcellular sugar homeostasis. This conclusion gained further support by electrophysiological evidence for a proton/sugar antiport catalyzed by TMT proteins (Schulz et al., 2011), by the observation that TMT1 gene expression is strongly induced by cold (Wormit et al., 2006), that TMT activity is regulated by a mitogen-activated protein triple kinase (Wingenter et al., 2011), and that increased TMT activity led to altered intracellular glucose compartmentation, that is, increased glucose concentrations in the vacuole and less glucose in the cytosol (Wingenter et al., 2010). However, for a dynamic import/export of monosaccharides their vacuolar release via TMT proteins is impossible. This is because the proton-motive force across the vacuolar membrane (Krebs et al., 2010; Schumacher & Krebs, 2010) prevents TMT-mediated sugar export into the cytosol by this proton/sugar antiporter. Accordingly, the observation that the carrier ERDL6 resides in the tonoplast and that erdl6 loss-of-function mutants exhibit increased total concentrations of glucose, but unchanged concentrations of fructose and sucrose, gave rise to the conclusion that ERDL6 is a vacuolar glucose exporter (Poschet et al., 2011).

A detailed phylogenetic analysis revealed that BvIMP is the closest sugar beet homolog to ERDL6 from Arabidopsis (Poschet et al., 2011). BvIMP was previously found in sugar beet tonoplasts by use of a peptide-specific antibody (Chiou & Bush, 1996) and this location was confirmed by us via use of GFP-fusion protein (Fig. S1a–d).

The recently developed electrophysiological technique of a ‘whole-vacuole configuration’ allowed us to identify electrogenic transport of both mono and disaccharides across the tonoplast (Wingenter et al., 2010; Schulz et al., 2011; Schneider et al., 2012). Here we employed this technique to identify the modus operandi of ERDL6, the Arabidopsis homolog to BvIMP. According to the data gained, ERDL6 transports glucose in symport with protons (Fig. 2b,c) which implies an unidirectional glucose export from the vacuolar lumen into the cytoplasm under cellular conditions. Given that the Arabidopsis homolog ERDL10 also showed reduced transport activity when expressed in yeast and treated with the uncoupler CCCP (Fig. S2d), that knock-out mutants lacking ERDL6 accumulate glucose, and not fructose or sucrose (Poschet et al., 2011), and that cold temperatures induce a strong increase of the cellular glucose concentrations in the two BvIMP overexpression lines (Fig. 5a), we assume that members of the ERDL protein family are mainly involved in vacuolar glucose homeostasis, acting as proton coupled exporters (Fig. 2). The observation that sugar feeding leads to decreased amounts of BvIMP mRNA (Fig. 4b) further substantiates the conclusion that ERDL6 and BvIMP act as vacuolar glucose exporters, because sugar application leads to high cytosolic sugar concentrations due to the pumping activity of plasma membrane-located STP type carriers (Büttner & Sauer, 2000) strongly increased.

We have shown before that overexpression of the vacuolar monosaccharide transporter TMT1 leads to a decreased cytosolic/vacuolar monosaccharide ratio and that the corresponding relative changes of CAB1 mRNA can be used as a molecular readout for this process (Wingenter et al., 2010). This type of molecular readout is possible, because the expression of CAB1 is governed by cytosolic glucose-sensing capabilities (Koch, 1996; Moore et al., 2003). We observed that under both standard growth conditions and sugar-feeding conditions CAB1 mRNA is lower in BvIMP overexpressor lines when compared to corresponding changes in WT plants (Fig. 1). Thus, the CAB1 expression in both BvIMP overexpressor lines in response to sugars fully concurs with the assumption that BvIMP acts as a vacuolar exporter for glucose (Fig. 2a,b). During glucose or fructose feeding TMT1 imports these sugars into the vacuole and in WT cells a futile cycle of glucose import and export is prevented by the sugar-induced downregulation of the AtERDL6 – or BvIMP – gene expression (Poschet et al., 2011; Fig. 1b).

It was surprising to see that seeds from BvIMP overexpressor lines showed impaired germination patterns when supplied with additional glucose or fructose, while germination in the absence of sugars was unaltered (Fig. 3a–c). Interestingly, Arabidopsis mutants lacking the homologous protein (ERDL6) activity also exhibited impaired seed germination (Poschet et al., 2011). The latter observation in conjunction with our findings here (Fig. 3a–c) indicate that any sort of modification of vacuolar monosaccharide transport, either export or import capacity, negatively affects seed-germination efficiency. This surprising effect might be explained, given that conditions of blocked vacuolar sugar export might limit cellular respiration while, in contrast, under conditions of stimulated vacuolar sugar export expression of genes required for developmental processes is repressed (Rolland et al., 2002). Both processes would result in an arrest of seed development (Rolland et al., 2002). Both processes would result in an arrest of seed development. The observation that germination of BvIMP overexpressor seeds at low temperatures is strongly hampered when compared to WT seeds (Fig. 3d,e) lifts the observation of a defect of early plant development at an environmental level. It is well known that cold temperatures cause accumulation of sugars in developing seedlings from both mono- and dicotyledonous species (Sasaki et al., 1996; Marsado et al., 2000). Thus, it seems likely that excess cytosolic monosaccharides in BvIMP overexpressor plants are sensed and act repressively on early plant development, as also seen under conditions of external sugar feeding (Rolland et al., 2002; Dekkers et al. 2004).

Glucose and fructose contribute to the protection of membrane-associated cellular compounds, particularly proteins and polar lipids. It is assumed that the presence of several highly hydroxylated sugar moieties optimize the membrane phase-transition properties and stabilize protein structures to resist osmotic, heat or cold stress conditions (Leslie et al., 1995; Carninci et al., 1998). All higher plants analyzed so far accumulate high concentrations of soluble sugars during acclimatization to cold- or freezing conditions (Alberdi & Corcuera, 1991; Wanner & Junttila, 1999; Zuther et al., 2004; Wormit et al., 2006). We demonstrated in a recent analysis that significant portions of the cold-induced sugars – that is, glucose, fructose and sucrose - accumulate in the vacuole (Wingenter et al., 2011; Schulze et al., 2012). Thus, the question arises whether a decrease of vacuolar monosaccharide concentrations affects freezing tolerance of higher plants. Indeed, measurements of the electrical conductivity revealed that both BvIMP overexpressor lines and the tmt1-2 double loss-of-function line are significantly less tolerant of freezing temperatures (Fig. 6). These observations are fully in line with the demonstration that vacuolar monosaccharide concentrations increase several-fold during cold acclimatization (Schulze et al., 2012) and thus most likely actively contribute to the prevention of ice crystal formation. In this respect it is also worth mentioning that the vacuolar monosaccharide-importer gene TMT is cold induced (Wormit et al., 2006) while the BvIMP gene and its Arabidopsis homolog ERDL6 are cold repressed (Fig. 1a and Poschet et al., 2011; respectively). Obviously, the reciprocal response of both genes to cold temperatures contributes to the massive accumulation of monosaccharides in the vacuole during cold adaptation (Schulze et al., 2012), which is important for maximizing freezing tolerance (Fig. 6).

Our observations reported above clearly point to a requirement for strict fine tuning of vacuolar sugar import and export. Given the recent discovery of a mitogen-activated protein kinase required for full activation of the TMT1 protein (Wingenter et al., 2011) it will be challenging to further decipher all modes of regulation allowing the plant to adapt vacuolar sugar homeostasis to phases of sugar storage and sugar consumption.


We would like to thank the Deutsche Forschungsgemeinschaft and the FOR1061 for financial support.