Journal of Plant Research

, Volume 119, Issue 2, pp 115–123

Identification of sugar-modulated genes and evidence for in vivo sugar sensing in Arabidopsis


  • Silvia Gonzali
    • Department of Crop Plant BiologyUniversity of Pisa
  • Elena Loreti
    • Institute of Biology and Agricultural BiotechnologyCNR
  • Cinzia Solfanelli
    • Department of Crop Plant BiologyUniversity of Pisa
  • Giacomo Novi
    • Department of Crop Plant BiologyUniversity of Pisa
  • Amedeo Alpi
    • Department of Crop Plant BiologyUniversity of Pisa
    • Sant’Anna School of Advanced Studies
Regular Paper

DOI: 10.1007/s10265-005-0251-1

Cite this article as:
Gonzali, S., Loreti, E., Solfanelli, C. et al. J Plant Res (2006) 119: 115. doi:10.1007/s10265-005-0251-1


Sugar status regulates mechanisms controlling growth and development of plants. We studied the effects of sucrose at a genome-wide level in dark-grown 4-day-old Arabidopsis thaliana seedlings, identifying 797 genes strongly responsive to sucrose. Starting from the microarray analysis data, four up-regulated (At5g41670, At1g20950, At1g61800, and At2g28900) and four down-regulated (DIN6, At4g37220, At1g28330, and At1g74670) genes were chosen for further characterisation and as sugar sensing markers for in vivo analysis. The sugar modulation pattern of all eight genes was confirmed by real time RT-PCR analysis, revealing different concentration thresholds for sugar modulation. Finally, sugar-regulation of gene expression was demonstrated in vivo by using the starchless pgm mutant, which is unable to produce transitory starch. Sucrose-inducible genes are upregulated in pgm leaves at the end of a light treatment, when soluble sugars levels are higher than in the wild type. Conversely, sucrose-repressible genes show a higher expression at the end of the dark period in the mutant, when the levels of sugars in the leaf are lower. The results obtained indicate that the transcriptional response to exogenous sucrose allows the identification of genes displaying a pattern of expression in leaves compatible with their sugar-modulation in vivo.


Arabidopsis thalianaPhosphoglucomutase mutantSucroseSugar sensing


The ability of plants to sense sugars plays an important role in carbon partitioning and allocation in source and sink tissues. These processes are modulated as a consequence of the sugar status of the plant, and sugar signals act both at the transcriptional and translational levels in tight coordination with light and other environmental stimuli (Koch 1996; Roitsch 1999; Coruzzi and Zhou 2001). Besides sugar regulation of metabolic activities, sugar sensing and signalling are involved in the control of growth and development throughout the life of the plant, from germination to floral transition and senescence (Gibson 2005). Furthermore, sugar modulation of gene expression appears to be involved in responses to many biotic and abiotic stresses, often cross-talking with hormones (Gazzarrini and McCourt 2003; Gibson 2004). At least three different sugar-sensing systems have been proposed to operate in plants. One is a sucrose-specific signalling mechanism, the others are involved in hexose perception, which can be sensed via systems that are hexokinase-independent, not requiring hexose metabolism, or hexokinase-dependent (see Smeekens 2000; Loreti et al. 2001; Rolland et al. 2002; Halford and Paul 2003; Gibson 2005).

Several different approaches have been used to uncover and study sugar-regulated genes. Many such approaches employed increasing concentrations (often in the range of 60–180 mM) of either sucrose or glucose, exogenously fed to plants, to test the resulting changes in gene expression (Koch et al. 1992; Mita et al. 1995; Dijkwel et al. 1996; Umemura et al. 1997; Oliveira and Coruzzi 1999). Less frequently, lower (1–50 mM) concentrations of sugars were used (Jang and Sheen 1994; Loreti et al. 2000; Wingler et al. 2000; Villadsen and Smith 2004).

DNA microarray analysis was recently used to evaluate the effect of sugars (alone or interacting with other signals) on gene expression in Arabidopsis thaliana seedlings. Price et al. (2004) and Thum et al. (2004) showed that sugars modulate a broad range of genes involved in all the main cellular processes, from carbohydrate and nitrogen metabolism to signal transduction, metabolite transport and stress responses.

Little is known about the in vivo occurrence of sugar sensing, but Thimm et al. (2004) recently described the response of Arabidopsis to sugar starvation, claiming that exogenous sugars can revert the gene modulation resulting from an extended night period. This suggests that genes that are modulated by exogenous sugars are similarly modulated in vivo by the sugar status of the plant tissues. Lloyd and Zakhleniuk (2004) studied the genomic response to sugar accumulation in leaves of the mutant pho3, where sugar levels are higher than in wild type because of a defective copy of the SUC2 gene, supplying evidence of some consequent primary and secondary metabolic responses.

In the present paper, the effects of sucrose were studied at a genome-wide level in dark-grown 4-day-old Arabidopsis seedlings. Starting from microarray analysis data, a set of sucrose-modulated genes was chosen and their expression in the presence of sucrose and other sugars was thoroughly characterised by real time RT-PCR analysis. Finally, sugar-regulation of gene expression was demonstrated in vivo using the starchless pgmArabidopsis mutant (Caspar et al. 1985).

Materials and methods

Plant materials

Arabidopsis thaliana ecotype Columbia glabra (gl1-1) was used in this study. Seeds were sterilised with diluted bleach (10 min incubation in 1.7% sodium hypochlorite, rinsing and washing thoroughly in sterile water), and incubated in 2.5 ml liquid growing medium [Murashige-Skoog (MS) half-strength solution] in six-well plates. Plates were incubated in the dark at 4°C for 2 days and finally transferred to 23°C for 4 days before the sugar treatments. Two independent, replicated sucrose treatment experiments were performed. Each independent experiment consisted of four replicated seedlings cultures, pooled after RNA extraction.

Seeds of the phosphoglucomutase (pgm) Arabidopsis mutant were obtained from the Nottingham Arabidopsis Stock Centre, NASC (Scholl et al. 2000). Experiments performed using the pgm mutant used Arabidopsis ecotype Columbia 0 (Col 0) as the wild type.

RNA isolation, cRNA synthesis, and hybridisation to Affymetrix GeneChips

Total RNA was extracted from seedling samples using an Ambion RNAqueous extraction kit (Ambion, Austin, TX). RNA quality was assessed by agarose gel electrophoresis and spectrophotometry. RNA was processed for use on Affymetrix Arabidopsis ATH1 GeneChip arrays (Affymetrix, Santa Clara, CA) as previously described (Loreti et al. 2005). Hybridisation, washing, staining, and scanning procedures were performed by Biopolo (University of Milano Bicocca, Italy) as described in the Affymetrix technical manual. Expression analysis via the Affymetrix Microarray Suite software (version 5.0) was performed with standard parameters. Two independent, replicated experiments were performed for the sucrose treatment, and the output of the Affymetrix Microarray Suite software for each independent experiment was subjected to further analysis using Microsoft Excel (Microsoft, Redmond, WA). Signal values (indicating the relative abundance of a particular transcript) and detection call values (indicating the probability that a particular transcript is present) were generated by Microarray Analysis Suite 5.0 software. Probe pair sets (genes) called “Absent” in both the “control” and “treated” samples were removed from subsequent analyses. Furthermore, genes with “Absent” for the detection value in the baseline data and “Decrease” for the change call were excluded from the list. Similarly, genes with “Absent” for the detection call in the experimental data and “Increase” for the change value were also excluded from the list. Differences in transcript abundance, expressed as fold change, were calculated using the Microarray Analysis Suite 5.0 software change algorithm. Fold change was assumed to be correct only if the corresponding “change call” indicated a significant change (“I” increase; “D” decrease, generated by Microarray Analysis Suite 5.0 software). Expression data were filtered to select only those genes showing a coinciding change call in the two biological replicates samples for each experimental condition. Comparing the datasets by using the Microarray Analysis Suite 5.0 software algorithm identified genes significantly affected by exogenous sucrose. Only genes showing a significant change in both biological replicates were selected.

Real time RT-PCR

RNA was extracted from 4-day-old seedlings grown on MS 0.5× solution (control) or in the same medium supplemented with different sugars at the concentrations indicated in the figure legends. Total RNA, extracted with an RNAqueous kit (Ambion) according to the manufacturer’s instructions, was subjected to a DNase treatment using the TURBO DNA free Kit (Ambion). Two micrograms of each sample was reverse transcribed into cDNA with a High capacity cDNA archive kit (Applied Biosystems, Foster City, CA). Real time PCR amplification was carried out with the ABI Prism 7000 Sequence Detection System (Applied Biosystems), using the primers described in Supplemental Table S1. UBQ10 was used as an endogenous control. TaqMan probes specific for each gene were used. Probe sequences are reported in Supplemental Table S1. PCR reactions were carried out using 50 ng cDNA and TaqMan Universal PCR Master Mix (Applied Biosystems) following the manufacturer’s protocol. Relative quantitation of each single gene expression was performed using the comparative CT method as described in the ABI PRISM 7700 Sequence Detection System User Bulletin #2 (Applied Biosystems).

Sugar analysis

Samples were rapidly frozen in liquid nitrogen and ground to a powder. Samples were then extracted and assayed by coupled enzymatic assay methods measuring the increase in A340 as described by Guglielminetti et al. (1995).

Results and discussion

Analysis of Arabidopsis transcriptome in young seedlings grown in the presence of sucrose

Research into sugar sensing has taken great advantage of the use of exogenous sugars to identify physiological processes affected by carbohydrate levels. The present available knowledge about gene modulation by exogenous sugars is vast, but still suffers from the lack of evidence about the in vivo expression pattern of genes identified as sugar-modulated using an in vitro approach, usually based on the exogenous feeding of sugars to young seedlings.

As a starting point, we analysed the effects of 90 mM sucrose on the transcriptome of 4-day-old A. thaliana seedlings grown in the dark using the Affymetrix whole-genome ATH1 Genechip. A large number of genes were consistently induced or repressed in two biological replicates, with 5,156 probe sets (a total of over 22,746) showing a statistically significant change in the sucrose-treated samples (Fig. 1a). Treatment with 90 mM sucrose may also result in the modulation of osmotic-responsive genes. Therefore, in order to subtract such genes from the list of genes modulated by sucrose, we filtered our expression data by removing all genes that showed a response to 90 mM mannitol higher than 55% of that to sucrose in a parallel microarray experiment. Removing the mannitol-responsive genes results in a list of 3,257 putative sucrose-modulated genes (Fig. 1a, yellow + green + red dots; see Supplemental Table S2). Furthermore, genes showing a response lower than a 3-fold change were also removed, resulting in a list of 797 genes strongly responsive to sucrose (Fig. 1a, yellow + red dots; Supplemental Table S2).
Fig. 1

a Scatter plot relating modulation by sucrose to modulation by mannitol (both used at 90 mM). Each dot identifies a gene, and its position is related to the effect of sucrose and mannitol on its expression. Data are expressed as log ratio changes in expression. In order to subtract osmotic responsive genes from the list of genes modulated by sucrose, we filtered our expression data by removing all genes showing a response to 90 mM mannitol higher than 55% of that to sucrose (based on the log ratio). Removing the mannitol-responsive genes results in 3,257 putative sucrose-modulated genes (yellow + green + red dots). Furthermore, genes showing a response lower than a 3-fold change were also removed, leaving 797 genes strongly responsive to sucrose (yellow + red dots). Genes selected for subsequent analyses using real time RT-PCR are identified as red dots. b Analysis of the data using MapMan software (Thimm et al. 2004). Genes involved in the pathways of amino acid, lipid, starch, and sucrose metabolism as well as glycolysis, photosynthesis and the pentose-P pathway were identified and their changes in expression (fold-change) plotted. Each bubble identifies a metabolic cluster, and its position in the scatter plot is related to the proportion of genes in the cluster showing induction or repression by sucrose. The bubble radius is proportional to the total number of genes in the metabolic cluster, regardless of their responsiveness to sucrose

Analysis of the data using MapMan software (Thimm et al. 2004) allows clustering of the modulated genes according to their putative function. We analysed the effects of sucrose on some aspects of plant primary metabolism to obtain clues as to the consequences of exogenous sucrose treatment (Fig. 1b). Pathways involved in amino acid synthesis, starch synthesis/degradation and glycolysis are substantially induced by sucrose, while amino acid and lipid degradation as well as the photosynthesis light reactions appear to be sucrose-repressed. The effects on the pentose-P pathway are less well defined, with a comparable percentage of genes induced as well as repressed by sucrose (Fig. 1b). Interestingly, Thimm et al. (2004) evaluated the effects of sugar starvation on genes involved in plant primary metabolism, observing that low leaf sugar levels result in induction of lipid and amino acid breakdown, and repression of starch and amino acid synthesis and glycolysis. Thus, the overall effect of sugar starvation in vivo (Thimm et al. 2004) is indirectly confirmed by the results obtained by us in vitro (Fig. 1b).

Choice of some candidate sucrose-induced or repressed genes

The microarray experiment allowed identification of a very high number of genes that exhibited a strong response to exogenous sucrose at the mRNA level. To verify a possible coherent response to endogenous fluctuations in sucrose levels, we decided to select a limited number of sucrose-modulated genes for complete molecular characterisation. Eight genes (Fig. 1a, red dots) were thus selected as candidates from the 797 genes identified as sucrose-responsive (Fig. 1a, yellow + red dots). Besides high induction or repression levels, two choice criteria were principally employed: (1) genes codifying for proteins involved in different metabolic processes (sugar metabolism, amino acid metabolism, stress responses, etc.), and (2) possibly acting in different cell compartments (cytosol, chloroplast, mitochondrion, etc.). Six of the eight genes have not yet been characterised. We also included in our analysis two genes already identified as sugar-responsive but only partially characterised. The four sucrose-induced genes are At5g41670, At1g20950, At1g61800, and At2g28900. Of these, only At1g61800, codifying for a putative chloroplast glucose-6-phosphate/phosphate translocator, has been previously studied; its expression was found to be induced in leaves of the mutant pho3, according to its high endogenous sucrose levels (Lloyd and Zakhleniuk 2004). This gene is usually expressed only in heterotrophic tissues. For this reason, Lloyd and Zakhleniuk (2004) hypothesised that a change in the nature of metabolite exchange between the plastid and the cytosol occurs in the pho3 mutant as a consequence of high sucrose levels. Concerning the other three sucrose-induced genes, both At5g41670 (a mitochondrial 6-phosphogluconate dehydrogenase family protein) and At1g20950 (a putative pyrophosphate-fructose-6-phosphate 1-phosphotransferase) are involved in glucose metabolism (pentose-phosphate pathway and glycolysis, respectively), a pathway activated by high sugar levels (see also Fig. 1b). The product of At2g28900 is a mitochondrial import inner membrane translocase subunit, which should be involved in protein transport; its induction concurs with a general increase in respiratory activity by high sucrose.

The glutamine-dependent asparagine synthetase 1 (ASN1) gene (At3g47340, Lam et al. 1994) is one of the four sucrose-repressed genes examined. Its expression is repressed by light and by sucrose in both light- and dark-grown plants (Lam et al. 1995, 1998). At3g47340 also corresponds to DIN6, a dark-induced and sugar-repressed gene (Fujiki et al. 2001). Expression of this gene correlates with sugar starvation, a situation in which asparagine represents a very important metabolite for nitrogen storage or transport (Fujiki et al. 2001). The other sucrose-repressed genes chosen were At4g37220 (a putative stress responsive protein, similar to cold acclimatisation proteins targeted to membranes, Breton et al. 2003), At1g28330 (a putative dormancy/auxin-associated protein, also known as DRM1), and At1g74670 (a putative gibberellin-responsive protein). Their negative regulation by high sucrose is less obvious, and their tentative annotation does not allow speculation about their actual function; however, these genes were chosen because they represent useful markers for a general analysis of sugar sensing in vivo (e.g. they were not chosen on the basis of preexisting knowledge about their possible involvement in sugar-modulated pathways).

Analysis of gene expression in the presence of different sugars

With the aim of confirming and further characterising the sugar-regulated expression of the selected genes, real time RT-PCR analysis was carried out using RNA extracted from 4 day-old Arabidopsis seedlings grown in the presence of increasing concentrations of sucrose, glucose, fructose, and turanose. Turanose is a non-metabolisable analogue of sucrose that has already been used to dissect the sucrose-signalling pathways in plants (Loreti et al. 2000; Sinha et al. 2002). Use of turanose, together with glucose and fructose, could allow detection of specific sucrose sensing for the genes analysed. Mannitol was included as an osmotic control, but, as expected, was not able to influence expression of the genes analysed (Figs. 2, 3), confirming that their sugar-specific modulation was not dependent on osmotic effects. Only when the mannitol concentration was increased to 25 mM was expression of At4g37220 slightly increased (Fig. 3). However, this being one of the sugar-repressed genes, the result further indicates that the sugar effects were not due to mere osmotic effects.
Fig. 2

Real time RT-PCR analysis of expression of At5g41670, At1g20950, At1g61800 and At2g28900 in dark-grown 4-day-old A. thaliana seedlings treated for 6 h with increasing concentrations (0–200 mM) of sucrose, glucose, fructose, turanose and mannitol. Expression levels are indicated relative to the untreated sample (0 mM; expression level =1). Each value is the mean ±SD of three independent measurements

Fig. 3

Real time RT-PCR analysis expression of DIN6 (At3g47340), At4g37220, DRM1 (At1g28330) and At1g74670 in dark-grown 4-day-old A. thaliana seedlings treated for 6 h with increasing concentrations (0–200 mM) of sucrose, glucose, fructose, turanose and mannitol. Expression levels are indicated relative to the untreated sample (0 mM; expression level =1). Each value is the mean ±SD of three independent measurements

Turanose turned out to be ineffective in modulating expression of sucrose-induced genes (Fig. 2). This could reflect a lack of specificity for sucrose-sensing or the existence of a sensing machinery located inside the cell, where turanose cannot be transported (Loreti et al. 2000). With regard to the effects of the other three sugars tested, the genes could be divided in two different groups. At5g41670 and At1g20950 showed low sugar specificity, being induced by sucrose, glucose, and fructose, but high sensitivity to low concentrations of sugars (Fig. 2). In contrast, At1g61800 and At2g28900 were principally induced by sucrose and only moderately by glucose, but were less responsive to low sucrose concentrations (Fig. 2). Concerning sucrose-repressed genes, a different behaviour was observed. DIN6, At4g37220 and DRM1 were almost completely repressed by sucrose or glucose at concentrations as low as 5–10 mM (Fig. 3). Fructose was effective at higher concentrations (25–50 mM) and turanose was able to slightly repress DIN6 expression in a concentration-dependent manner, although it was ineffective on At4g37220 and DRM1 expression (Fig. 3). At1g74670 was shown to be repressed by 50–200 mM sucrose, while it was slightly induced by 5 mM sucrose or glucose, and by 5–10 mM fructose (Fig. 3). Again, turanose had no effect on the expression of this gene (Fig. 3).

On the whole, five out of eight genes responded to concentrations of sucrose as low as 5 mM (At5g41670, At1g20950, DIN6, At4g37220, DRM1), while At1g61800, At2g28900 and At1g74670 required 50 mM sucrose to be modulated (see Figs. 2, 3). Sucrose and glucose appear to be the most effective modulators of gene expression, although fructose can modulate At5g41670 and At1g20950. It is tempting to speculate that distinct sensing mechanisms may operate, possibly with a low sugar-concentration sensor as well as a high sugar-concentration sensor. Interestingly, all five genes regulated by low sucrose (At5g41670, At1g20950, DIN6, At4g37220, DRM1) are also modulated by glucose and fructose, suggesting that the low-concentration sensor could also be a low-specificity sensor.

Time-course of gene expression

A time-course of induction or repression of the selected genes was performed using 5 and 50 mM sucrose or glucose, these being the two most effective sugars, as already discussed, and 5 and 50 mM being the two main threshold concentrations. Fifty millimolar sucrose or glucose induced both At5g41670 and At1g20950 in a time-dependent manner, while the level of induction was, as expected, lower in the presence of 5 mM sucrose or glucose (Fig. 4, left panels). Sucrose at 50 mM was required for maximum induction of At1g61800 and At2g28900, whose mRNA levels increased during the 8 h of treatment (Fig. 4, left panels). Glucose at 50 mM induced expression of these two genes during the first 4 h of treatment, but their mRNA levels declined in the subsequent 4 h (Fig. 4, left panels). Neither sucrose nor glucose at 5 mM increased mRNA levels during the 8 h experiment (Fig. 4, left panels), confirming the lower sugar responsiveness of these genes. Concerning sugar-repressed genes, DIN6 and DRM1 mRNA levels displayed very similar trends regardless of the sugar employed, with a strong and almost complete repression within the first 2–4 h of treatment (Fig. 4, right panels). At4g37220 showed a slower sugar-dependent repression, which reached a maximum within the 4th and 8th hour of treatment (Fig. 4, right panels). Finally, At1g74670 showed a more complex pattern of regulation. Despite being categorised as “sugar-repressed”, 5 mM sucrose or glucose was confirmed to be able to induce the gene (Fig. 3), although the effect of glucose was transient (Fig. 4, right panels).
Fig. 4

Time-course of expression of the eight selected genes in the presence of sucrose or glucose at 5 or 50 mM. Gene expression was determined by real time RT-PCR analysis in dark-grown 4-day-old Arabidopsis seedlings treated for 0–8 h with one of the two sugars. Expression levels are indicated in relative units assuming the value of gene expression at t0=1. Each value is the mean ±SD of three independent measurements

The time-course experiment (Fig. 4) reveals that genes that behave similarly in terms of sugar concentration response (e.g. DIN6 and At4g37220) may differ as far as the timing of the response is concerned (Fig. 4). This could reflect, for example, a different mechanism of regulation, with some genes that can be modulated directly by sugars being therefore earlier in their response, while others may be modulated indirectly. Further work is certainly needed to ascertain the existence of multiple sugar sensing mechanisms, the existence of which have been postulated (Smeekens 2000; Loreti et al. 2001; Rolland et al. 2002; Halford and Paul 2003; Gibson 2005) and, at least in part, suggested by the results presented here.

Analysis of in vivo gene expression in wild type and pgm mutant plants

Sugar responsive genes should be regulated in vivo according to the different amounts of soluble sugars present in leaf tissue during the day. A starchless mutant of A. thaliana was used to obtain leaf samples containing different sugar concentrations compared to wild type (Fig. 5). The pgm mutant is defective in the plastidial form of phosphoglucomutase (Caspar et al. 1985), an enzyme required to convert glucose-6-phosphate into glucose-1-phosphate, which is then converted into ADP-glucose, the glucose donor for starch synthesis. The pgm mutation thus hampers the synthesis of starch. Therefore, higher levels of soluble sugars accumulate in the leaves and stems of the mutant grown in the light (Fig. 5, 5 and 10 h light treatment; Caspar et al. 1985; Gibon et al. 2004), while free sugars rapidly deplete during dark periods, probably due to very high rates of respiration, leading pgm leaves towards sugar starvation at the end of the night (Fig. 5, t0; Gibon et al. 2004). For this reason, we analysed the expression level of the eight selected genes in wild type and pgm leaves at the end of a dark period and after different light treatments. All eight selected genes were shown to be expressed in the leaf tissues analysed. All the sugar-inducible genes were expressed more highly in the pgm mutant 10 h after the onset of light treatment (Fig. 6), in accordance with the higher level of total soluble sugars (Fig. 5). Conversely, a very low sugar level is present at the end of night in the pgm mutant. This should result in higher expression of sugar-repressible genes at time 0 h (Fig. 6), which corresponds to the end of night. Indeed all the sugar-repressible genes, with the single exception of At1g74670, were expressed more highly in pgm leaves than in wild type leaves at time 0 h (Fig. 6). During light treatment, the expression levels of DIN6, At4g37220 and DRM1 in the pgm mutant rapidly declined, reaching the lowest values already after the first 5 h of light (Fig. 6), in accordance with the increase in soluble sugar content. In wild-type leaves, the expression level of DIN6 and At4g37220 was almost unchanged, while DRM1 was very slightly repressed during the first 5 h (Fig. 6). It is therefore possible that the sugar content at time 0 h was already high enough to maintain all these genes in a repressed state, and that further increases in the leaf sugar content were not sensed. The expression level of At1g74670 in the wild type at time 0 h was higher than that in the pgm mutant, and declined rapidly during light treatment (a likely consequence of the increasing level of sugars in the leaves), while in the pgm mutant a repression during the 5–10 h of light is preceded by a transient induction of expression (Fig. 6). Although this pattern was more difficult to predict than the others, the data already collected on sugar-specificity (Fig. 3) as well as the time-course experiment (Fig. 4) reveal that At1g74670 expression is enhanced by low sugar concentrations, while only high sucrose levels (50–200 mM) can repress accumulation of its mRNA.
Fig. 5

Content of glucose, fructose and sucrose measured in wild type and pgm leaves at the end of a dark period (0) and after 5 and 10 h of a light treatment (90 μm photons m−2). Each value is the mean of three independent measurements. SD is calculated on the total amount of sugars

Fig. 6

Expression levels of the eight selected genes analysed by real time RT-PCR in wild-type (open circles) and pgm (closed circles) leaves after 0, 5, and 10 h of light (90 μm photons m−2). Expression levels are indicated in relative units assuming the value of gene expression of the wild-type sample at t0=1. Each value is the mean ±SD of three independent measurements

Analysis of all eight selected genes showed that the differential expression exhibited by wild-type and pgm leaves are on the whole consistent with the differences in sugar levels. This suggests that genes selected as sugar-responsive using an in vitro sucrose-feeding protocol are also responsive to in vivo changes in leaf carbohydrate content. Therefore, sugar-modulation represents a major mechanism of regulation for the eight candidate genes in vivo. However, their restricted number does not allow us to draw more general considerations. For this reason, we evaluated whether this small-scale validation is consistent with whole-genome changes in transcript level as a consequence of sugar starvation, taking advantage of the microarray dataset published by Thimm et al. (2004) comparing the transcriptome of wild-type A. thaliana to that of the pgm mutant. Several genes were up or down-regulated in pgm mutant leaves at the end of night (Thimm et al. 2004). Assuming that the 797 sugar-modulated genes we identified (Fig. 1a, yellow dots) are responsive to the endogenous level of sugars in A. thaliana leaves, they should be expressed differently in the wild type compared to the pgm mutant. To verify this hypothesis, we plotted the dataset from Thimm et al. (2004), identifying the sugar modulated genes on the basis of our selection criteria (Fig. 1a, yellow dots). The resulting scatter plot (Fig. 7a) shows that most of the red dots (identifying sugar-repressible genes, see list in Supplemental Table S2) are located near the y axis, while blue dots (identifying sugar-inducible genes, see list in Table S2) are found near the x axis (Fig. 7a). This is what was expected, since genes located nearby the y axis (Fig. 7a) are upregulated in the pgm mutant and thus are likely to be sugar-repressible (the pgm mutant has a low sugar content by the end of night, Thimm et al. 2004; Fig. 5). The genes located near the x axis (Fig. 7a) are instead upregulated in the wild type, and thus are likely to be sugar-inducible (the wild-type has a higher sugar content by the end of night, Thimm et al. 2004; Fig. 5). Interestingly, the same pattern of distribution of blue and red dots is observed when the dataset from the wild-type/end of night is plotted against the wild-type/end of extended night (8 h-longer night, Thimm et al. 2004). Again, the red dots located near the y axis (Fig. 7b) identify genes that we categorised as sugar-repressible and that in vivo are more highly expressed in leaves that experienced a longer night, i.e. with a lower sugar content (Thimm et al. 2004). The opposite applies to the blue dots, identifying sugar-inducible genes (see Supplemental Table S2), which show a lower expression in vivo in sugar-starved leaves that experienced a longer night (Thimm et al. 2004).
Fig. 7a,b

Colour-coding of sucrose-modulated genes identified in the present work using the dataset from Thimm et al. (2004). a Scatter plot showing a comparison of gene expression in samples from wild-type A. thaliana leaves with samples from pgm mutant leaves, both at the end of the dark period (see Thimm et al. 2004 for experimental details). Sucrose-modulated genes identified as such on the basis of our experiments and selection criteria (Fig. 1a, yellow dots) are marked in red and blue. Red dots Sugar-repressible genes, blue dots sugar-inducible genes. b Scatter plot showing a comparison of gene expression in samples from wild-type A. thaliana leaves at the end of night with samples from leaves taken at the end of an extended dark period (see Thimm et al. 2004 for experimental details). Sucrose-modulated genes identified as such on the basis of our experiments and selection criteria (Fig. 1a, yellow dots) are marked in red and blue. Red dots Sugar-repressible genes, blue dots sugar-inducible genes


In the present paper we describe the identification of a large set of sucrose-modulated genes using a transcriptome approach. Real time RT-PCR characterisation carried out on a restricted number of selected genes, confirmed the microarray data and demonstrated that not only sucrose, but in some cases also hexoses, although to only a minor extent, can play a role in the regulation of mRNA accumulation of specific genes. We further demonstrated that the eight selected genes, as well as the whole set of sucrose-modulated genes, can respond to fluctuations in sugar levels taking place in vivo. We conclude that the transcriptional response to exogenous sucrose allows identification of genes that display a pattern of expression in leaves suggesting that they are actually sugar modulated. Multiple sugar sensing mechanisms may operate in vivo, and further work is certainly required to identify the sensing and signalling pathways operating in A. thaliana leaves subjected to fluctuations in their sugar content.

Supplementary material

Table S1

10265_2005_251_ESM_supp1.pdf (250 kb)
(PDF 250 KB)

Table S2

10265_2005_251_ESM_supp2.xls (16.4 mb)
(Excel 16.7 MB)

Copyright information

© The Botanical Society of Japan and Springer-Verlag 2006