Light signalling mediated by phytochrome plays an important role in cold-induced gene expression through the C-repeat/dehydration responsive element (C/DRE) in Arabidopsis thaliana

Authors


*For correspondence (fax +82-62-972-5085; e-mail jmkim@ksc.kumho.co.kr).
†These authors contributed equally to this work.

Summary

Low temperature induces a number of genes that encode the proteins promoting tolerance to freezing, mediated by ABA-dependent and ABA–independent pathways in plants. The cis-acting element called C/DRE is known to respond to low temperature independently of ABA action. To investigate the signalling and network of ABA-independent pathways, the transgenic Arabidopsis plants were generated containing several copies of the C/DRE derived from cor15a gene with a minimal promoter fused to a GUS reporter gene. The transgenic plants containing four copies of the C/DRE (4C/DRE-GUS) showed responsiveness to cold and drought treatments and were used for characterization of cold signalling and cross-talk. Cold-induced GUS expression was inhibited by okadaic acid at 1 nm, indicating that protein phosphatase 2A might act as a positive regulator. Light was shown to activate cold- and drought-induced GUS expression. Photo-reversibility of the GUS mRNA by red and far-red light with concomitant cold treatment suggests a role of phytochrome as a photoreceptor in mediating light signalling to activate the cold-induced gene expression through the C/DRE. Furthermore, GUS expression analysis in phyA or phyB or phyAphyB mutant backgrounds showed that phytochrome B is a primary photoreceptor responsible for the activation of cold-stress signalling in response to light. Light enhanced the induction kinetics of CBF1, 2, and 3 encoding the cognate transcription factors, and cor15a, in a consecutive manner compared to the dark condition in the cold, suggesting that the connection point between cold and light signalling mediated by phytochrome is at a higher step than the expression of CBF genes.

Introduction

Plants respond to a variety of environmental stresses including cold, drought, and high salinity, by changing various aspects of physiological processes to increase their survival rates (Bray, 1997; Hughes and Dunn, 1996; Shinozaki and Yamaguchi-Shinozaki, 1997, 1999; Thomashow, 1998, 1999). A large body of experiments suggests that changes in gene expression play a major role in increasing the ability of the plants to tolerate environmental stresses. Numerous genes are induced at the transcriptional level in response to various stresses (Shinozaki and Yamaguchi-Shinozaki, 1997, 1999; Thomashow, 1998, 1999). The products encoded by these genes can be classified into two groups (Shinozaki and Yamaguchi-Shinozaki, 1999). The first group includes proteins that probably function in protecting the plant cells against stress such as late embryogenesis abundant proteins, antifreeze proteins, osmotin, and mRNA binding proteins. The other group includes proteins that function as regulatory proteins involved in gene expression and signal transduction.

The expression analysis of the stress-inducible genes in ABA-deficient (aba) and ABA-insensitive (abi) mutants and others have established that the signal transduction from the perception of the stress to the expression of the stress-inducible genes is mediated by ABA-dependent and ABA-independent pathways (Gilmour and Thomashow, 1991; Grill and Himmelbach, 1998; Leung and Giraudat, 1998; Nordin et al., 1991, 1993; Shinozaki and Yamaguchi-Shinozaki, 1997). ABA-responsive elements (ABRE) and their trans-acting factors regulating the expression of the genes mediated by ABA-dependent pathway have been extensively analysed (Leung and Giraudat, 1998), and several components involved in the signal transduction pathways have been identified by molecular genetic analysis (Finkelstein and Lynch, 2000; Giraudat et al., 1992; Gosti et al., 1999; Leung et al., 1994; Meyer et al., 1994). In ABA-independent pathway, the C/DRE containing a core sequence of –CCGAC- has been shown to be essential for the transcriptional activation in response to cold, drought, and/or high salt treatments (Yamaguchi-Shinozaki and Shinozaki, 1994). The transcription factors that bind to the C/DRE have been isolated, CBF(C/DRE binding factor)s and DREB(DRE binding protein)s, both contain a DNA binding motif found in EREBP1 and AP2 transcriptional activators (Liu et al., 1998; Stockinger et al., 1997). Over-expression of CBF1(= DREB1B) and DREB1A(= CBF3) in transgenic Arabidopsis enhanced freezing and drought tolerance with a concomitant expression of target genes (Jaglo-Ottosen et al., 1998; Kasuga et al., 1999; Liu et al., 1998).

To investigate stress signal transduction pathways, a number of Arabidopsis mutants showing altered expression of stress-inducible genes have been isolated using transgenic plants transformed with a chimeric construct containing the firefly luciferase reporter gene driven by the C/DRE and ABRE-containing rd29A promoter (Ishitani et al., 1997, 1998; Xiong et al., 1999a). Based on the characteristics of these mutants, a working model was proposed in which ABA-dependent and ABA-independent pathways interact and converge to activate stress genes. Using the same transgenic plants containing the luciferase reporter, interactive effects of temperature, osmotic stress and ABA were investigated, indicating that both positive and negative interactions exist among these stress factors in regulating gene expression (Xiong et al., 1999b). Although this approach using the whole promoter of a stress-inducible gene revealed possible interactions between stress signalling, it was difficult to dissect the signalling events converging into individual cis-acting elements and their interaction, due to the complicated interactive effects of the known (C/DRE and ABRE) and unknown cis-acting elements. Expression of the luciferase reporter gene under various stress conditions could result from the synergistic, inhibitory, or additive effects of the several cis-acting elements existing in the whole promoter. To overcome this problem, we made multiple copies of the synthetic C/DRE, whose sequence was derived from the cor15a promoter fused to a GUS reporter gene with a minimal promoter, and generated transgenic Arabidopsis plants containing these chimeric constructs. With these transgenic plants, we found that light can activate cold-stress signalling converging into the C/DRE. Treatment of 3-(3,4-dichlorophenyl)-1,1-dimethyl urea, an inhibitor of electron transport chain, to 4C/DRE-GUS during cold treatment did not significantly reduce GUS expression, suggesting that products of photosynthesis are not involved in light-stimulated induction of the GUS mRNA. We further showed that phytochrome B is a primary photoreceptor mediating light signalling to activate cold-induced gene expression at a higher step of the CBF transcription factors through the C/DRE.

Results

Four copies of the C/DRE is sufficient for cold- and drought-induced gene expression

The mRNAs of cor genes accumulate upon cold treatment in both the ABA-insensitive mutant abi1 and ABA-deficient mutant aba-1 of Arabidopsis as it does in the wild type, suggesting that the ABA-independent pathway plays a major role in mediating cold stress to cor gene expression (Gilmour and Thomashow, 1991). Among cor genes, cor15a was extensively investigated in terms of stress gene regulation. We therefore used the C/DRE of cor15a as a molecular probe to study the ABA-independent cold-signal transduction pathway and possible cross-talk with other signalling pathways. The various cis-acting elements of the promoter region of cor15a are depicted in Figure 1a (Baker et al., 1994). It contains three ABREs at positions −71, − 132, and −308 and two C/DREs at positions −185 and −362 (Figure 1a). Two additional motifs are the potential G-box core element (CACGTG) located at position −132 and the G-box-related sequence (TACGTG) at position −71. There are three additional repeat sequences named A, B, and D throughout the promoter region. Cor15a was shown to respond to cold, drought and ABA (Baker et al., 1994).

Figure 1.

Diagrams of the cor15a promoter and the chimeric constructs used for generating Arabidopsis reporter plants.

(a) Schematic view of cor15a promoter showing the stress-responsive cis-acting elements and the repeating sequences. ABRE is positioned at −71, −132, and −308 and C/DRE positioned at −185 and −362. The G-box core element is located at position −132 and the G-box-related sequence at position −71, which overlap with the ABRE. A, B, and D are the repeating sequences whose functions are unknown.

(b) Structure of the chimeric constructs used for Arabidopsis transformation. Abbreviations: RB, right border; N-P, nopaline synthase promoter; NPTII, neomycin phosphotransferase II; N-T, nopaline synthase terminator; 35SmP, minimal promoter encompassing −90 to +4 nucleotides from CaMV 35S promoter; GUS, β-glucuronidase; LB, left border.

(c) The various chimeric constructs used for generating Arabidopsis reporter plants. MP-GUS, a plant transformation vector containing a minimal promoter fused to GUS reporter gene. The rest of the vector parts are the same as in (b). 2C/DRE-GUS, two copies of the C/DRE fused to GUS reporter gene. 4C/DRE-GUS, four copies of the C/DRE fused to GUS reporter gene.

We first determined whether the C/DRE alone is sufficient for cold-induced gene expression. C/DRE was synthesized containing –CCGAC- core sequence with the 10 nucleotides of the flanking sequences at both the 5′- and 3′-ends (Figure 1b), and inserted into the plant transformation vector containing GUS gene and TATA-box-containing minimal promoter, encompassing −90 to + 4 nucleotides of CaMV 35S promoter, yielding the construct without the C/DRE or containing 2 or 4 copies of the C/DRE (Figure 1c). We then transformed Arabidopsis plants with each construct, generating MP-GUS, 2C/DRE-GUS, and 4C/DRE-GUS plants. Homozygous plants of light-grown 10-day-old-seedlings were treated at 3°C for 2 days with white fluorescent light and visualized by histochemical GUS assay.

Figure 2 shows typical GUS staining pattern of the transgenic plants. MP-GUS plants did not show any detectable GUS expression upon cold treatment (Figure 2a,b). In 2C/DRE-GUS plants, GUS expression was detected in the petiole, but it was not enhanced by cold treatment at 3°C (Figure 2c,d). In contrast, 4C/DRE-GUS plants showed GUS expression over the whole leaf after cold treatment (Figure 2f) while GUS expression was detected only in the petiole when incubated at room temperature (23°C) (Figure 2e). GUS expression was also enhanced when 4C/DRE-GUS plants were subjected to drought treatment (Figure 2k), although the level of the reporter activity was lower than cold treatment. While it cannot be ruled out that the condition for inducing GUS expression upon drought treatment may not be optimized, it is more likely that proper drought responsiveness may require more flanking sequences than the element that we used for the C/DRE or other interactive element(s) nearby. CaMV 35S-GUS plants did not show any difference in GUS expression between the treatments at room and cold temperature (Figure 2g,h) and drought stress (Figure 2m). These results show that four copies of the C/DRE are sufficient for cold- and drought stress-induced gene expression and that they exhibit similar tissue specificity as the expression of cor15a gene (Baker et al., 1994), except for GUS expression in the petiole. We then tested whether the C/DRE responds to ABA or NaCl, using 4C/DRE-GUS plants. As shown in Figure 2o,p, and Figure 3, the C/DRE did not respond to ABA at 100 µm or NaCl at 300 mm concentration, which is the maximum concentration for inducing the stress genes (Xiong, L et al., 1999), demonstrating that the C/DRE specifically responds to cold and drought stresses. GUS expression in CaMV 35S-GUS was not affected by NaCl or ABA treatment (Figure 2q,r).

Figure 2.

Histochemical GUS staining of transgenic Arabidopsis containing various constructs of the C/DRE.

Transgenic Arabidopsis plants containing various constructs were grown for 10 days at 16 h photoperiod, subjected to various treatments for 2 days, and the promoter activity was assayed by visualizing GUS expression histochemically. At least 10 seedlings from several homozygous lines were used for the assay and the typical examples of GUS staining were shown. (a and b), MP-GUS plants; (c and d), 2C/DRE-GUS plants; (e, f, j, k, l, o, and p), 4C/DRE plants; (g, h, i, m, n, q, and r), CaMV35S-GUS plants.

Plants were incubated at 23°C (a, c, e, g, k, m, o, p, q, and r) or at 3°C (b, d, f, and h) with light or at 3°C without light (i and j). Plants were incubated under drought stress with (k and m) or without (l and n) light. Plants were incubated with 100 µm ABA (o and q) and 300 mm NaCl (p and r), respectively.

Figure 3.

RT–PCR analysis of temperature, DCMU, NaCl, and ABA on the expression of GUS gene fused to four copies of the C/DRE cis-acting element.

4C/DRE-GUS plants grown for 10 days after germination at 16 h photoperiod were incubated at 3°C with light under appropriate conditions (lanes 1–5).

Total RNAs from each treatment were isolated, and the accumulation of the GUS mRNA was determined by RT–PCR assay. The actin7 mRNA was assayed for control. 100 µm of ABA or 300 mm NaCl or 10 µm of DCMU was used for each treatment.

Light is necessary for cold- or drought-induced gene expression mediated by the C/DRE

Light was shown to be required for enhanced freezing tolerance induced by cold temperature in a variety of plants including Arabidopsis (Gray et al., 1997; Levitt, 1980; Wanner and Junttila, 1999). To test if light is a factor regulating cold-induced gene expression mediated by the C/DRE in Arabidopsis, light-grown 4C/DRE-GUS plants were immediately treated with cold temperature at 3°C in the presence or absence of fluorescent white light, followed by histochemical GUS assay. When the transgenic plants were incubated in the dark during cold treatment, only the basal level of GUS expression in the petiole was detected (Figure 2j). A similar result was obtained with plants treated with drought stress (Figure 2l). CaMV 35S-GUS plants incubated in the cold (Figure 2h,i) or under drought stress (Figure 2m,n) did not show any difference in GUS expression between the treatments in the light and in the dark. Dark incubation of 4C/DRE-GUS plants at 23°C did not show any GUS staining over the leaf (data not shown). These results show that light is necessary for inducing gene expression through the C/DRE in response to cold and drought conditions. To examine whether photosynthesis is involved in mediating cold stress to the gene expression, DCMU (3-(3,4-dichlorophenyl)-1,1-dimethyl urea), an inhibitor of the electron transport chain, was included during cold treatment with light. RT–PCR analysis of the GUS mRNA showed that DCMU did not significantly inhibit GUS expression (Figure 3), indicating that light signalling is involved in mediating cold and drought stress to the C/DRE cis-acting element, but not the result of photosynthesis such as redox potential.

Light stimulates the expression of cor15a and its cognate transcription factor genes, CBFs, during cold treatment

To understand the role of light in cold-induced gene expression of cor15a mRNA through the C/DRE, we examined the time course of GUS mRNA accumulation in comparison to the induction of cor15a mRNA, with or without fluorescent white light during cold treatment, using RT–PCR. For this purpose, Arabidopsis seedlings were treated at 3°C for 1, 5, 24, and 48 h with or without light. A significant increase in GUS mRNA level was detected at 24 h, which decreased after further treatment of light and cold (Figure 4a). However, only a slight increase in GUS mRNA level at 24 h was found by dark treatment, consistent with histochemical GUS assay results (Figure 2). In contrast, high level of cor15a mRNA accumulation was found even at 48 h in both light and dark conditions by cold treatment. However, light stimulated the induction of the cor15a mRNA at 5 h compared to darkness in the cold, suggesting that light enhanced the induction kinetics of the cor15a mRNA during cold treatment. We next asked whether light regulates the transcription of the CBF genes encoding the cognate trans-acting factors that bind to the C/DRE cis-acting element, thereby modulating cold-induced gene expression of cor15a. Arabidopsis seedlings were treated at 3°C for 1, 5, 24, and 48 h with or without light, followed by RT–PCR analysis with specific primers for CBF1, 2 and 3 (Gilmour et al., 1998; Medina et al., 1999). Each CBF transcript had a distinct profile for cold induction as shown in Figure 4b. The accumulation of the mRNAs of CBF1, 2, and 3 was stimulated by cold in both light and dark treatments. However, light greatly enhanced the accumulation of the CBF mRNAs at 1 h compared to darkness, again showing that light stimulates the induction kinetics of expression of CBF genes and therefore the cor15a gene. All these results suggest that light signalling is linked to the C/DRE pathways at a higher step than the expression of CBF genes to activate cold-stress signalling.

Figure 4.

Light-stimulated induction of the GUS, cor15a, CBF1, 2, and 3 mRNAs in response to cold treatment.

Total RNAs from each treatment were isolated, and subjected to RT–PCR analysis.

(a) Induction kinetics of the GUS and cor15a mRNAs in response to cold. 4C/DRE-GUS plants grown for 10 days at 16 h photoperiod were incubated at 3°C with (lanes 2–5) or without light (lanes 6–9) for 1, 5, 24, and 48 h. Lane 1 indicates the sample before the treatment.

(b) Induction kinetics of the CBF1, 2, and 3 mRNAs in response to cold with or without light. 4C/DRE-GUS plants were incubated at 3°C with (lanes 2–5) or without light (lanes 6–9).

(c) Dark-suppression of cor15a mRNA induction by cold in etiolated seedlings. Ten-day-old seedlings grown in the light (lanes 1–4) or dark (lanes 5–8) were incubated at 23°C (lanes 1, 2, 5, and 6) or 3°C (lanes 3, 4, 7 and 8) and with (lanes 2, 4, 6, and 8) or without light (lanes 1, 3, 5, and 7) for 48 h. D and L indicate dark and light treatments, respectively.

We further examined whether light can stimulate the induction of cor15a mRNA in etiolated seedlings compared to light-grown seedlings during cold treatment. As shown in Figure 4c, light significantly enhanced the accumulation of the cor15a mRNA in etiolated seedlings compared to darkness even at 48 h incubation. This light-stimulated expression pattern of cor15a in the etiolated seedlings mimics the light-regulated GUS expression pattern of C/DRE-GUS during cold treatment.

Phytochrome mediates light signalling to cold-induced gene expression through the C/DRE

As a step toward understanding connection between light and cold stress signalling, we first wanted to examine whether phytochrome photoreceptor is involved in light signalling linked to the C/DRE pathway. Phytochromes can exist in two different conformers: the red light-absorbing, biologically inactive Pr form and the far-red light-absorbing, biologically active Pfr form (Neff et al. 2000). Both forms are photoconvertible by red and far-red light. Thus, in order to test whether phytochrome is a responsible photoreceptor, we examined photoreversibility of the accumulation of GUS mRNA by giving single, short pulses of red and/or far-red light for 10 min at 3°C to 4C/DRE-GUS plants. The plants treated were then further incubated in the dark at 3°C for 24 h to check the accumulation of GUS mRNA. Total RNA isolated from each plant treated was subjected to RT–PCR for the GUS mRNA. For the quantitative assay, PCR products separated by agarose gel electrophoresis were transferred onto nylon membranes, probed with 32P-labelled GUS DNA, and exposed to X-ray film (Figure 5a,b). The GUS mRNA level was increased 2.5-fold when 4C/DRE-GUS plants were treated with cold in the dark compared to 23°C in the dark (Figure 5a, lanes 1 and 2, and Figure 5b). Cold treatment under white light could enhance the accumulation of GUS mRNA up to 7-fold compared to 3°C in the dark (lane 3). Treatment of the transgenic plants with 10 min pulse of red light increased the GUS mRNA level as white light treatment for 24 h at 3°C (lane 4). Far-red light also induced the GUS mRNA, but the level was significantly lower than that of red and white light-treated seedlings, which is only 2-fold higher than dark-treated seedlings at 3°C (lane 5). When the seedlings were treated with red light, immediately followed by far-red light pulse, the induction of the GUS mRNA caused by red light was cancelled (lane 6). The GUS mRNA level went back to the level as observed with the plants treated with far-red light. These results clearly demonstrated that phytochrome mediates the light signalling to the cold-induced gene expression through the C/DRE cis-acting element.

Figure 5.

Phytochrome mediates the light signalling to activate the C/DRE promoter in response to low temperature.

(a) RT–PCR analysis of 4C/DRE-GUS plants subjected to treatments of various light sources in the cold. 4C/DRE-GUS plants were treated with 10 min pulse of red light (lane 4) or far-red light (lane 5) or red light and immediately followed by far-red light (lane 6) with simultaneous cold treatment and then incubated in the dark at 3°C for 24 h. Lanes 1–3 show samples treated at 23°C, 3°C in the dark, and 3°C in the light for 24 h, respectively. RT–PCR products were separated on the agarose gel, transferred onto nylon membranes, probed with 32P-labelled GUS DNA fragment, and exposed to X-ray film. Duplicate experiments were quantitated with a phosphoimager (b).

(b) Quantitative analysis of RT–PCR products (a) using phosphoimager. Bar indicates se.

Phytochrome B is a primary photoreceptor responsible for activating cold-induced gene expression through the C/DRE

The phytochrome (phy) family comprises five members from phyA to E, with an overlapping but distinct mode of action in Arabidopsis (Neff et al., 2000; Quail et al., 1995; Smith, 2000). Phy B-E are light-stable, whereas phyA accumulates to a high level in the dark and its Pfr form is rapidly degraded. Photoresponses exerted by the light-stable phytochromes follow the classical red/far-red photoreversibility. Analysis of mutants null for phyA, B, D, and E showed that phyB has a role at all stages of the life cycle, whereas phyA, D, and E exert their principal functions at selected stages (Robson and Smith, 1997; Whitelam and Devlin, 1997). We therefore reasoned that phyB is a likely photoreceptor for linking light signalling to cold-stress signalling. In order to further test whether phytochrome functions as a photoreceptor linking light signalling to cold-stress signalling and to determine the type of phytochrome, we took a genetic approach with phyA, phyB, and phyAphyB Arabidopsis mutants. For this approach, we crossed phyA, phyB, and phyAphyB Arabidopsis mutants with 4C/DRE-GUS, and selected homozygous lines for both phy mutant phenotypes and 4C/DRE-GUS transgene. These double mutants were then incubated at 3°C for 2 days with or without light, and GUS expression was assayed histochemically. As shown in Figure 6, while high levels of GUS expression were found in phyA mutant background as 4C/DRE-GUS, minimal levels of GUS expression were detected in phyB and phyAphyB mutant backgrounds, demonstrating that phyB is responsible for activating cold-induced gene expression through the C/DRE. This result obtained by the genetic approach is in good agreement with photoreversibility of GUS expression with red and far-red light in 4C/DRE-GUS plants (Figure 5).

Figure 6.

Histochemical analysis of GUS activity of 4C/DRE in phytochrome mutant backgrounds.

Homozygous Arabidopsis seedlings containing 4C/DRE-GUS in wild type (4C/DRE-GUS), phyA, phyB, and phyAphyB backgrounds were incubated at 3°C in the light or dark, or at 23°C in the light for 2 days, and GUS activity was assayed histochemically.

Protein phosphatase 2A might function as a positive regulator in ABA-independent pathway

Calcium was shown to act as second messenger mediating cold stress to the activation of cold-acclimation specific (cas) genes (Monroy and Dhindsa, 1995). Phospholipid/CDPK inhibitor, staurosporine, prevented the cold-induction of cas15 in alfalfa, whereas the protein phosphatase (PP) 1 and 2A inhibitor, okadaic acid, induced the accumulation of the cas15 mRNA at 25°C (Monroy et al., 1998). Calcium-dependent protein kinase (CDPK) has also been shown to activate a barley stress-inducible promoter that responds to cold, salt, and ABA treatments when it was overexpressed in protoplasts (Sheen, 1996). To test whether these components are involved in the C/DRE-cold signal transduction pathway, we added various inhibitors including staurosporine, okadaic acid, the calcium channel blocker verapamil, calcium chelator BAPTA(1,2-bis(o-aminophenoxy)ethane N, N, N′, N′-tetraacetic acid), and calcium channel ionophore A23187 to 4C/DRE-GUS plants during cold treatment and determined GUS expression histochemically (data not shown). Among the inhibitors tested, okadaic acid at 250 nm significantly inhibited cold-induced GUS expression. PP2A has 0.1–1.0 nm of IC50 (concentration for inhibiting the enzyme activity to 50%) for okadaic acid whereas PP1 has 10–100 nm of IC50 (Smith and Walker, 1996). Therefore, to identify the type of phosphatase involved in cold-induced gene expression, we incubated 4C/DRE-GUS plants at the concentrations of 1, 5, 25, and 250 nm of okadaic acid during cold treatment. RT–PCR analysis of the GUS mRNA isolated from individually treated samples showed that okadaic acid effectively inhibited GUS expression, even at the concentration of 1 nm (Figure 7), suggesting that PP2A might function as a positive regulator in cold-induced gene expression mediated by the C/DRE.

Figure 7.

The effects of okadaic acid on the expression of GUS gene fused to four copies of the C/DRE cis-acting element.

(a) 10-day-old seedlings of 4C/DRE-GUS were incubated at 3°C in the light for 2 days with varying concentrations of okadaic acid at 0, 1, 5, 25, and 250 nm (lanes 1–5). RT–PCR assays were performed as described in Figure 3.

(b) Histochemical GUS assays of 4C/DRE-GUS plants treated at 3°C for 2 days (left panel) or with okadaic acid at 1 nm (right panel).

Discussion

Previous genetic studies have suggested that extensive interaction occurs between osmotic stress, temperature stress, and ABA responses (Ishitani et al., 1997, 1998; Xiong et al., 1999a). Molecular analysis of the reporter gene expression fused to a rd29A stress-inducible promoter has further revealed that there are positive and negative interactions among these responses (Xiong et al., 1999b). While these approaches provided evidence indicative of the interactive responses between different stress signalling pathways, it was difficult to explore the precise nature of these interactions due to the synergistic, inhibitory, and additive effects of the several cis-acting elements existing in the promoter. Here we used multiple copies of the C/DRE cis-acting element as a molecular probe to study the ABA-independent pathway of the cold-stress signalling. We showed that four copies of the C/DRE could confer responsiveness on GUS reporter gene expression upon cold treatment, but not by ABA and NaCl. Therefore, this approach allowed us to eliminate the interactive impacts of different signalling converging into the various cis-acting elements.

Light is necessary for inducing gene expression through the C/DRE in response to low temperature

In barley, the accumulation of cor14b transcript and the encoded protein was stimulated by light compared to darkness when the dark-grown plants were treated with cold temperatures (Crosatti et al., 1995, 1999). It was also shown that phytochrome, and/or blue light receptor, mediated the accumulation of the cor14b protein (Crosatti et al., 1999). Expression of the Wcs9 gene in wheat was similarly enhanced by light (Chauvin et al., 1993). In our studies, using transgenic Arabidopsis containing 4C/DRE-GUS construct, we showed that light is necessary for activating gene expression through the C/DRE promoter element upon cold treatment in Arabidopsis.

Dark incubation of 4C/DRE-GUS plants substantially reduced cold-induced GUS expression up to 7-fold compared to light conditions. DCMU could not inhibit cold-induced GUS expression in the light, suggesting that light signalling, not the product of photosynthesis such as redox potential, is involved in gene expression through the C/DRE. To further understand the role of light in the expression of cor15a gene containing C/DRE within the promoter element upon cold treatment, the time course of GUS mRNA induction was determined in comparison with those of cor15a and the genes encoding cognate trans-acting factors, CBF1, 2, and 3. The results showed that light stimulates the expression of CBFs and cor15a in a consecutive manner, faster than dark conditions in the cold. This suggests that the light-stimulated induction of the cognate trans-acting factor genes resulted in faster induction kinetics of cor15a gene expression in the light compared to darkness during cold treatment. This result further suggests that the connection point between light and cold signalling is at a higher step in the pathway of the accumulation of CBF mRNAs, most likely at the transcriptional level. The saturated level of the cor15a mRNA was obtained with longer treatment of light-grown Arabidopsis seedlings with light during cold treatment, and thus no difference was found between dark and light conditions. This may result partly from the equal level of expression of CBF1 and 3 after 5 h. It is also possible that other cis-acting elements such as ABA-responsive elements and G-box elements might contribute cold-induced cor15a gene expression in the absence of light, resulting in the saturated level of gene expression at 48 h in the cold. However, in the case of etiolated seedlings, stimulated accumulation of cor15a mRNA was found in the light, even at 48 h. Significantly less of this mRNA was detected in the dark with etiolated seedlings compared to light-grown seedlings. In other words, dark conditions in the etiolated seedlings decreased the saturated level of cold-induced gene expression. This situation is similar to transgenic 4C/DRE-GUS plants that show tight regulation of GUS expression between dark and light conditions during cold treatment, indicating that other cis-acting elements in the promoter region of cor15a that are developmentally regulated play a role in cold-induced gene expression of cor15a at prolonged illumination, allowing saturated level of gene expression. The etiolated plants are morphologically and physiologically different from the light-grown plants. Thus, the photo-sensing and signalling to the C/DRE may also be affected in the etiolated plants. In any cases, all these strongly suggest that the C/DRE pathway is tightly linked to light signalling (see Figure 8 for a working model).

Figure 8.

Working model showing cross-talk between cold signalling and light signalling mediated by phytochrome.

(a) Cold signalling in the dark.

(b) Cold signalling in the light. In the absence of light, gene expression through the C/DRE is slightly induced by cold (a).

Light can convert Pr form of the phytochrome B into Pfr form, which then turns on the signalling events to activate basal level of cold-stress signalling, leading to the enhanced gene expression through the C/DRE (b). Connection point between light and cold signalling is at the higher step of CBFs gene expression. PP2A is acting as a positive factor in cold-stress signalling. Thick and thin arrows indicate activated and basal level of signalling, respectively.

Phytochrome B functions as a primary photoreceptor linking light signalling to cold stress signalling pathways

Short pulse treatment for 10 min of 4C/DRE plants with red light-induced GUS expression after cold treatment in the dark for 24 h, and this induction was cancelled by far-red light treatment. This result suggested that phytochrome is a primary photoreceptor mediating light signals to cold-induced gene expression through the C/DRE. Furthermore, GUS expression analysis in phyA, phyB, and phyAphyB mutant backgrounds obtained by crossing each phy mutant Arabidopsis with 4C/DRE-GUS, showed that only in phyA mutant background, wild type level of GUS expression was found, demonstrating that phytochrome B is a primary photoreceptor responsible for activating cold stress signalling to the C/DRE. In other phy mutants, highly curtailed GUS expression is induced in response to cold in the light. Since the C/DRE is highly conserved among the promoters of low-temperature inducible genes in Arabidopsis and among dicots, and between dicot and monocot plants (Hughes and Dunn, 1996; Thomashow, 1999), cross-talk between light and cold signalling may be a general phenomenon in plants. It is not clear why and how light signalling mediated by phytochrome is tightly linked to cold signalling pathways converging into the C/DRE through its cognate trans-acting factors. The identification and characterization of 4C/DRE-GUS mutants exhibiting GUS expression in the dark at low temperatures will reveal the components involved in this connection and the physiological functions of cross-talk between light and cold signalling.

The function of PP2A in the C/DRE cold-signal transduction pathway

Okadaic acid at the concentration of 1 nm inhibited the cold-induced GUS expression in 4C/DRE-GUS plants. This result suggested that protein phosphatase 2A, as a positive regulator, might be involved in the cold signal transduction pathway mediated by the C/DRE, as opposed to the role of protein phosphatase 2C as a negative regulator in the ABA-dependent pathway (Gosti et al., 1999). It has been shown that the transcript of a 2A phosphatase-associated protein named TAP46 is inducible by chilling treatment but not by heat or anaerobic stress in Arabidopsis (Harris et al., 1999), suggesting that PP2A may positively function in the cold-response.

In alfafa protoplasts, okadaic acid treatment induced the cas15 at 25°C (Monroy et al., 1998). On the basis of the observation of an almost complete inhibition of PP2A upon exposure of cells to cold, it has been proposed that PP2A is a target for cold-inactivation, and that this inactivation might result in induction of cas15 during cold-acclimation. Apparently, this looks contradictory to our result that PP2A might function as a positive regulator for cold-signal transduction pathway in Arabidopsis. The clear demonstration of the involvement of PP2A in the cold-signal transduction pathway and whether PP2A acts as either a positive or a negative regulator, need the isolation and characterization of the Arabidopsis mutants where the mutations of the genes encoding PP2A subunits impair cold-acclimation.

Flanking sequences of the C/DRE may determine salt-responsiveness

The rd29A mRNA is induced in response to dehydration, cold, high salt, and ABA (Liu et al., 1998; Yamaguchi-Shinozaki and Shinozaki, 1994) whereas cor15a was shown to respond to cold, drought, and ABA (Baker et al., 1994). Deletion analysis of the promoter of rd29A showed that the 20 bp direct repeat with CCGAC core sequence functioned as a dehydration-, high salt-, and low temperature-responsive element, but not as an ABA-responsive element. Mutation of -CCGA- into -TTTT- completely abolished stress-responsiveness. In our work, 4C/DRE-GUS plants did respond to cold and drought conditions, but not to high salt and ABA. The rd29A promoter contains -taCCGACat- as the C/DRE whereas the cor15a promoter contains -ggCCGACct-, suggesting that the lack of the response of 4C/DRE-GUS plants to high salt is due to the difference in the 5′- and 3′- flanking sequences of the C/DRE.

In summary, our studies showed that even a single cis-acting element, here, the C/DRE, did not constitute a simple isolated, linear signal transduction pathway in response to stress signals, but the pathway was connected to other signalling pathways such as light, providing fine tuning of coordination between stress signals for the regulation of gene expression and plant responses.

Experimental procedures

Construction of pMP-GUS, p2C/DRE-GUS, and p4C/DRE-GUS gene fusions

DNA fragment for the minimal promoter, containing BamHI at the 5′-end and XbaI at the 3′-end, was generated encompassing −90 to +4 of CaMV 35S promoter by PCR. This DNA fragment was subcloned into pBI121 (Clontech), a plant transformation vector containing GUS reporter gene, at BamHI and XbaI sites after deleting the CaMV 35S promoter to produce pMP-GUS. The synthetic C/DRE was made containing 5′-aatttcatgg CCGACctgctttttt−3′ with the compatible restriction sites, BamHI and BglII at the 5′-end on top and bottom strands, respectively. Each DNA strand was phosphorylated at the 5′-end with T4 polynucleotide kinase, and the mixture of both strands was annealed by heating the sample at 90°C for 5 min and then incubating it at 37°C for 15 min, and ligated. To remove the ligation yielding the reverse orientation, the ligation products were treated with both restriction enzymes. Dimer and tetramer of the C/DRE isolated on the agarose gel were ligated into pMP-GUS at BamHI site, producing p2C/DRE-GUS and p4C/DRE-GUS, respectively. The orientation of the insert was verified by DNA sequencing.

Arabidopsis transformation and generation of 4C/DRE-GUS/phy mutants

pCaMV35S-GUS(pBI221), pMP-GUS, p2C/DRE-GUS, and p4C/DRE-GUS were introduced into Arabidopsis thaliana (Columbia-0 ecotype) using vacuum-infiltration method through Agrobacterium-mediated transformation (Bechtold et al., 1993). Thirty independent transformants (T1) were screened on the media containing kanamycin (50 mg l−1) and selfed. The T2 transformants showing a 3 : 1 segregation ratio were selected, made for the T3 homozygous transformants, and amplified.

PhyA-211 and phyB-9 Arabidopsis mutants were from ABRC, and phyA-211phyB-9 double mutants were kindly provided by M. Soh. The 4C/DRE-GUS gene was introduced into these backgrounds by crossing them with the 4C/DRE-GUS (2514) line.

Treatment of cold, drought, ABA, high salt, DCMU, light, and the inhibitors for signalling components

The transgenic Arabidopsis thaliana was grown on germination agar plates (0.5x Murashege Skoog media with vitamins, 1.5% sucrose, 2.5 mm MES, pH 5.7, 0.8% agar) for 10 days at 23°C for a 16 h photoperiod. These seedlings were incubated at 3°C with white fluorescent light (0.4 mW cm−2) for 2 days for cold treatment. For drought stress, the seedlings were dehydrated on Whatman 3MM paper. For the treatment of ABA, high salt, and DCMU, the transgenic plants were germinated and grown on the filter paper put on top of the germination agar plate, and then wet with the solution containing 100 µm of ABA or 300 mm NaCl or 10 µm of DCMU. These plates were further incubated at 3°C or at 23°C for a given period of time with or without light. For the dark incubation, the plates containing plants were completely wrapped with aluminium foils three times. LED-controlled environment chamber (Percival Scientific, model E-30LED1) was used for the treatment of red light (34 µW cm−2) and far-red light (31 µW cm−2). Seedlings were removed from the incubation chamber for treatment 4 h after the 16 h photoperiod had started.

The transgenic plants were incubated with the following inhibitors as described above. Staurosporine, okadaic acid, verapamil, BAPTA (1,2-bis(o-aminophenoxy)ethane N, N, N′, N′-tetraacetic acid), and A23187 were added at the concentrations of 5 µm, 0.25 µm, 0.1 mm, 2 mm, and 50 µm, respectively (Monroy and Dhindsa, 1995; Monroy et al., 1998).

Histochemical GUS assay

A histochemical assay of GUS activity was performed by incubating the treated seedlings in 5-bromo-4-chloro-3-indolyl glucuronide (DUCHEFA, the Netherlands) at 37°C for 24 h and removing the chlorophyll from green tissues by incubation in 100% ethanol, as described (Jefferson and Wilson, 1991).

Rt–pcr

Total RNA was isolated from the transgenic plants treated under various conditions, using RNeasy Plant Mini Kit (Qiagen) and subjected to RT–PCR analysis with Access RT–PCR System (Promega) according to the instruction manual. 0.1 µg of the total RNA was used for RT–PCR of the GUS (25 cycles) and actin7 mRNAs (20 cycles). For Southern analysis of PCR products of the GUS mRNA, 20 cycles of PCR were performed. For RT–PCR of cor15a, CBF1, 2, and 3 mRNAs, total RNAs used were 0.4 (35 cycles), 0.8 (40 cycles), 0.2 (40 cycles), and 0.8 µg (40 cycles), respectively. Annealing temperature was 60°C except that 55 and 62°C were used for RT–PCR of cor15a and CBF3 mRNAs, respectively. All the RT–PCR products were verified by size of the PCR products and DNA sequencing. PCR primers used were 5′-ATGTTACGTCCTGTAGAAAC-3′ (upstream primer) and 5′-TCATTGTTTGCCTCCCTGCT-3′ (downstream primer) for GUS mRNA; 5′-ATGTCTTTCTCAGGAGCTGT-3′ (upstream primer), 5′-CTACTTTGTGGCATCCTTAG-3′ (downstream primer) for cor15a mRNA; 5′-ATGGCCGATGGTGAGGATAT-3′ (upstream primer), 5′-TTAGAAGCATTTCCTGTGAA-3′ (downstream primer) for actin 7 mRNA; 5′-ATGAACTCATTTTCAGCTTT-3′ (upstream primer), 5′-TTAGTAACTCCAAAGCGACA-3′ (downstream primer) for CBF1 mRNA; 5′-ATGAACTCATGTTCTGCTTT-3′ (upstream primer), 5′-TTAATAGCTCCATAAGGACA-3′ (downstream primer) for CBF2 mRNA; 5′-ATGAACTCATTTTCTGCTTT-3′ (upstream primer), 5′-TTAATAACTCCATAACGATA-3′ (downstream primer) for CBF3 mRNA.

Acknowledgements

We would like to thank Drs M.-S. Soh and S.-Y. Kim for fruitful discussion and critical reading of the manuscript. We also thank J.-H. Kang for technical assistance with plasmid construction. Y.-K. Kim and H.-J. Kim were supported by the intern research fellowship from Korea Science and Engineering Foundation. This work was supported in part by a grant (PF003105) from Plant Diversity Research Center of 21st Century Frontier Research Program, funded by Ministry of Science and Technology of Korean government and a grant from Agricultural Plant Stress Research Center funded by Korea Science and Engineering Foundation to J. Kim. This paper is KLESL publication number 53.

Ancillary