Interaction between two cis-acting elements, ABRE and DRE, in ABA-dependent expression of Arabidopsis rd29A gene in response to dehydration and high-salinity stresses

Authors

  • Yoshihiro Narusaka,

    1. Biological Resources Division, Japan International Research Center for Agricultural Sciences (JIRCAS), 1-1 Ohwashi, Tsukuba, Ibaraki 305-8686, Japan,
    Search for more papers by this author
    • Current address: Department of Resource and Energy Science, Graduate School of Science and Technology, Kobe University, 1-1 Rokkodai-cho, Nada-ku, Kobe 657–8501, Japan.

  • Kazuo Nakashima,

    1. Biological Resources Division, Japan International Research Center for Agricultural Sciences (JIRCAS), 1-1 Ohwashi, Tsukuba, Ibaraki 305-8686, Japan,
    Search for more papers by this author
  • Zabta K. Shinwari,

    1. Biological Resources Division, Japan International Research Center for Agricultural Sciences (JIRCAS), 1-1 Ohwashi, Tsukuba, Ibaraki 305-8686, Japan,
    Search for more papers by this author
  • Yoh Sakuma,

    1. Biological Resources Division, Japan International Research Center for Agricultural Sciences (JIRCAS), 1-1 Ohwashi, Tsukuba, Ibaraki 305-8686, Japan,
    Search for more papers by this author
  • Takashi Furihata,

    1. Biological Resources Division, Japan International Research Center for Agricultural Sciences (JIRCAS), 1-1 Ohwashi, Tsukuba, Ibaraki 305-8686, Japan,
    Search for more papers by this author
  • Hiroshi Abe,

    1. Biological Resources Division, Japan International Research Center for Agricultural Sciences (JIRCAS), 1-1 Ohwashi, Tsukuba, Ibaraki 305-8686, Japan,
    Search for more papers by this author
  • Mari Narusaka,

    1. Laboratory of Plant Molecular Biology, RIKEN Tsukuba Institute, 3-1-1 Koyadai, Tsukuba, Ibaraki 305-0074, Japan, and
    2. Plant Functional Genomics Group, RIKEN Genomic Sciences Center, 3-1-1 Koyadai, Tsukuba, Ibaraki 305-0074, Japan
    Search for more papers by this author
  • Kazuo Shinozaki,

    1. Laboratory of Plant Molecular Biology, RIKEN Tsukuba Institute, 3-1-1 Koyadai, Tsukuba, Ibaraki 305-0074, Japan, and
    2. Plant Functional Genomics Group, RIKEN Genomic Sciences Center, 3-1-1 Koyadai, Tsukuba, Ibaraki 305-0074, Japan
    Search for more papers by this author
  • Kazuko Yamaguchi-Shinozaki

    Corresponding author
    1. Biological Resources Division, Japan International Research Center for Agricultural Sciences (JIRCAS), 1-1 Ohwashi, Tsukuba, Ibaraki 305-8686, Japan,
      * For correspondence (fax +81 29 838 6643; e-mail kazukoys@jircas.affrc.go.jp).
    Search for more papers by this author

* For correspondence (fax +81 29 838 6643; e-mail kazukoys@jircas.affrc.go.jp).

Summary

Many abiotic stress-inducible genes contain two cis-acting elements, namely a dehydration-responsive element (DRE; TACCGACAT) and an ABA-responsive element (ABRE; ACGTGG/TC), in their promoter regions. We precisely analyzed the 120 bp promoter region (−174 to −55) of the Arabidopsis rd29A gene whose expression is induced by dehydration, high-salinity, low-temperature, and abscisic acid (ABA) treatments and whose 120 bp promoter region contains the DRE, DRE/CRT-core motif (A/GCCGAC), and ABRE sequences. Deletion and base substitution analyses of this region showed that the DRE-core motif functions as DRE and that the DRE/DRE-core motif could be a coupling element of ABRE. Gel mobility shift assays revealed that DRE-binding proteins (DREB1s/CBFs and DREB2s) bind to both DRE and the DRE-core motif and that ABRE-binding proteins (AREBs/ABFs) bind to ABRE in the 120 bp promoter region. In addition, transactivation experiments using Arabidopsis leaf protoplasts showed that DREBs and AREBs cumulatively transactivate the expression of a GUS reporter gene fused to the 120 bp promoter region of rd29A. These results indicate that DRE and ABRE are interdependent in the ABA-responsive expression of the rd29A gene in response to ABA in Arabidopsis.

Introduction

Exposure to environmental stresses such as drought, salinity, and low temperature causes adverse effects on the growth of plants and the productivity of crops. In plants, a number of genes have been reported to be expressed in response to these stresses. Sequence analysis of these genes has revealed that their gene products might function in tolerance and response to stress in plants (Ingram and Bartels, 1996; Shinozaki and Yamaguchi-Shinozaki, 2000). Some dehydration-responsive genes are also induced by abscisic acid (ABA). It appears that dehydration triggers the production of ABA, which in turn induces various genes. Cis- and trans-acting factors involved in ABA-induced gene expression have been analyzed (reviewed in Hasegawa et al., 2000; Shinozaki and Yamaguchi-Shinozaki, 2000). Many ABA-inducible genes contain a conserved ABA-responsive cis-acting element named ABA-responsive element (ABRE; ACGTGG/TC) in their promoter regions (Bonetta and McCourt, 1998; Grill and Himmelbach, 1998; Leung and Giraudat, 1998).

Repeated copies of ABREs can confer ABA responsiveness to a minimal promoter, whereas a single copy of ABRE is not responsive to ABA (Skriver et al., 1991). In the wheat Em promoter, two ABRE motifs, Em1a and Em1b, contribute to the activity of the promoter (Guiltinan et al., 1990). These two ABREs also function as cis-acting elements in ABA-dependent expression of the rd29B gene (Uno et al., 2000). In the barley HVA1 promoter, the typical ABRE motif and the coupling element 3 (CE3) directly upstream of ABRE have been shown to be sufficient for ABA-induced gene expression (Shen and Ho, 1995). The CE3 sequence (ACGCGTGTCCTG) is similar to the typical ABRE (ACGTGG/TC) sequence, and some other coupling elements also contain an A/GCGT core sequence (Hobo et al., 1999). On the other hand, the coupling elements without A/GCGT are reported (Kao et al., 1996; Rogers and Rogers, 1992; Shen and Ho, 1995).

Three cDNAs encoding bZIP-type transcription factors have been isolated as ABRE-binding proteins in the rd29B promoter using the yeast one-hybrid system and designated as AREB1, AREB2, and AREB3 (Uno et al., 2000). Transcription of the AREB1 and AREB2 genes is upregulated by drought, high-salinity, and ABA treatments in vegetative tissues. Both the AREB1 and AREB2 proteins function as transcriptional activators and require ABA for their activation. Choi et al. (2000) also reported that four cDNAs encoding ABFs are related to or are identical to the AREB1 and AREB2 proteins. The constitutive overexpression of ABF3 or ABF4/AREB2 in Arabidopsis resulted in ABA hypersensitivity and other ABA-associated phenotypes (Kang et al., 2002).

Analyses of drought-induced genes have indicated the existence of ABA-independent signal transduction cascades between the initial signal of water deficit and the expression of specific genes (reviewed in Bray, 1997; Ingram and Bartels, 1996; Shinozaki and Yamaguchi-Shinozaki, 2000). Several stress-inducible genes such as rd29A and cor15a are induced through the ABA-independent pathway. A cis-acting element has been identified in the promoter region of rd29A and is responsible for dehydration-, high-salinity-, and low-temperature-induced expression. This sequence (TACCGACAT), namely the dehydration-responsive element (DRE), is also found in the promoter regions of many dehydration- and low-temperature stress-inducible genes (Shinozaki and Yamaguchi-Shinozaki, 2000; Thomashow, 1999; Yamaguchi-Shinozaki and Shinozaki, 1994). Similar cis-acting elements, named C-repeat (CRT) and low-temperature-responsive element (LTRE), both containing a G/ACCGAC core motif, regulate low-temperature-inducible promoters (Baker et al., 1994; Jiang et al., 1996; Thomashow, 1999).

Transcription factors (DREB/CBF) involved in DRE/CRT-responsive gene expression have been cloned (Liu et al., 1998; Stockinger et al., 1997). The homologs to DREB/CBF are also present in the Arabidopsis genome (Gilmour et al., 1998; Shinwari et al., 1998). Expression of the DREB1/CBF genes is induced by low-temperature stress, and that of the DREB2 genes is induced by dehydration and high salinity. Two independent DREB family members, DREB1/CBF and DREB2, function as transcriptional activators in two separate signal transduction pathways under low temperature and dehydration, respectively (Liu et al., 1998). Recently, Sakuma et al. (2002) reported that the DREB family includes three novel DREB1- and six DREB2-related genes and that two DREB1-related genes (DREB1D/CBF4 and DREB1F) are induced by high-salinity treatment. On the other hand, Haake et al. (2002) reported that CBF4/DREB1D functions as a transcription factor in the drought-responsive gene expression in Arabidopsis.

The Arabidopsis genes, rd29A and rd29B, are differentially induced under stress conditions and by ABA treatment. The promoter region of rd29A contains at least two cis-acting elements, two DREs and one ABRE, involved in ABA-independent and ABA-responsive gene expression (Yamaguchi-Shinozaki and Shinozaki, 1994). Recent genetic evidence indicates extensive interaction between ABA-dependent and ABA-independent pathways with a reporter gene system consisting of the firefly luciferase coding sequence driven by the rd29A promoter (Ishitani et al., 1997; Xiong et al., 2002). By contrast, the rd29B promoter contains two ABREs but no DRE and is mainly induced through the ABA-dependent signal pathway (Uno et al., 2000). Some of the genes regulated by DRE, such as cor15a, rd17/cor47, and erd10, are also highly expressed in response to the exogenous application of ABA (Seki et al., 2001; Thomashow, 1999). The DRE cis-acting element functions in the initial rapid response of rd29A to dehydration, high salinity, or low temperature, and another region that contains one ABRE is necessary for the ABA induction (Yamaguchi-Shinozaki and Shinozaki, 1994). Although the promoter region of rd29A contains only one ABRE and lacks any known coupling elements, the expression of rd29A is strongly induced by exogenous application of ABA. This suggests that a new cis-acting element may work as a coupling element in the ABA-responsive expression.

To understand the cross-talk between ABA-independent and ABA-dependent regulatory systems of gene expression, we have analyzed the regulation of rd29A gene expression and its promoter. In this study, we analyzed the 120 bp promoter region (−174 to −55) of rd29A containing DRE and ABRE and examined the interaction between DRE and ABRE in ABA-dependent gene expression. We reported here that DRE and ABRE are interdependent in the stress-responsive expression of the rd29A gene, and that DRE may function as a coupling element for ABRE in the ABA-dependent gene expression.

Results

Deletion and base-substitution analysis of the rd29A promoter for the dehydration-responsive induction of the GUS reporter gene in transgenic tobacco

To analyze the promoter region of rd29A containing DRE and ABRE, we constructed various chimeric genes using the 120 bp region (−174 to −55) of the rd29A promoter. The 120 bp region contains a DRE sequence (TACCGACAT), a DRE-core motif (GCCGAC), an ABRE motif (TACGTGTC), and an as1 sequence (GACGTC; Figure 1a). Firstly, we prepared seven kinds of 120 bp promoter fragments, with or without base substitutions, in the DRE, ABRE, and as1 sites. We also prepared four kinds of 5′-deleted DNA fragments of the 120 bp promoter region to −157, −151, −135, and −124. The monomer or tandem repeated dimer forms of these fragments were fused upstream of the −61 rd29A minimum promoter–β-glucronidase (GUS) fusion construct (Yamaguchi-Shinozaki and Shinozaki, 1994). These constructs were first introduced into tobacco plants as we could analyze GUS induction by dehydration stress more quickly and easily in tobacco than in Arabidopsis. We used leaves of primary transgenic plants containing each construct for analysis of GUS induction by dehydration stress. We analyzed more than 20 independent T1 tobacco plants for expression of each rd29A-GUS fusion gene under drought stress condition with the mutated or deleted promoter.

Figure 1.

Deletion and base-substitution analysis of the rd29A promoter for the induction of the GUS reporter gene in response to dehydration stress in transgenic tobacco.

The 5′-terminal deletion or base-substituted fragments of the rd29A promoter were fused to the GUS reporter gene and introduced into tobacco plants. Leaves of transformed tobacco plants were dehydrated for 24 h. Schematics of the 5′-terminal deletions and mutations of the promoters fused to the GUS reporter gene are shown on the left. The multiplicities of the induction of GUS activity by dehydration are shown on the right (ratio). GUS activity was measured in 20 independently obtained transgenic plants for each construct.

(a) Structure of the rd29A promoter (−174 to −55).

(b) Schematics of the chimeric constructs with a monomer of the rd29A promoter fragment and their GUS activities in transgenic tobacco.

(c) Schematics of the chimeric constructs with the tandemly repeated dimer of the DNA fragments and their GUS activities in transgenic tobacco.

The monomer form of the wild-type 120 bp fragment, which has both DRE and ABRE, showed dehydration-induced expression of GUS activity with a 26.7-fold increase (Figure 1b, WT). The dimer form of the wild-type 120 bp fragment induced a 16.4-fold increase in GUS activity after dehydration treatment, and the level of GUS activity was 10 times higher than that of the monomer form of the fragment (Figure 1c, WT). The monomer forms of the 120 bp fragments with six combinations of base substitutions in DRE, ABRE, and as1 showed considerably reduced GUS activity (Figure 1b, M1–M6). However, all six fragments with base substitutions still responded to dehydration stress. We obtained similar results with the dimer forms of the base-substituted 120 bp fragments, and each GUS induction by dehydration was higher than that of the monomer forms of the base-substituted fragments, except M3 and M6 (Figure 1c).

To analyze the role of the as1 sequence that may function in root-specific expression in GUS induction by dehydration (Yamaguchi-Shinozaki and Shinozaki, 1994), we prepared three kinds of constructs containing the base-substituted 120 bp (−174 to −55) fragments of the rd29A promoter. The as1 sequence, as1 and DRE, and as1 and ABRE were base-substituted in the three fragments, M4, M5, and M6, respectively. We analyzed both monomer and dimer forms of these fragments (Figure 1b,c). The GUS activity in the transgenic tobacco plants with these constructs was lower than that in the transformants containing the as1 sequence in most cases, but the tobacco plants still showed dehydration-induced expression of GUS activity. These results indicated that there are cis-acting elements involved in the dehydration-induced expression other than the DRE, ABRE, and as1 sequences in the 120 bp fragment of the rd29A promoter.

To examine this hypothesis, we analyzed the four 5′-deleted DNA fragments of the 120 bp promoter region to −157, −151, −135, and −124. The 5′-deletion of the 120 bp promoter region gradually reduced the level of GUS induction from −157 to −124 (Figure 1b). The dehydration-induced expression of GUS activities was very high in the −157 and −151 transformants (13.0- and 13.9-fold increase, respectively). When we deleted the region from −151 to −135, the promoter activity was reduced clearly. However, GUS induction in response to dehydration was still observed in the −135 transformant (3.0-fold increase). The −124 transformants exhibited little dehydration-induced expression of GUS activity (1.5-fold increase). We also analyzed dimer forms of these fragments and obtained similar results (Figure 1c). These results indicated that the region between −135 and −124 might contain any of the cis-acting elements that are involved in the dehydration-responsive expression of rd29A.

Analysis of the effect of various treatments on induction of the rd29A promoter-GUS fusion genes in transgenic Arabidopsis

We examined the effects of environmental stresses such as low-temperature, high-salinity, dehydration, and ABA treatments on the expression of the rd29A-GUS fusion gene with the 120 bp promoter region (−174 to −55) in transgenic Arabidopsis by RNA gel blotting. A tandem repeated dimer of the 120 bp DNA fragment fused to the −61 rd29A-GUS fusion gene responded to low-temperature, high-salinity, dehydration, and ABA treatments (Figure 2). The expression pattern of the GUS gene because of these stresses was similar to that of the endogenous rd29A gene. Base substitutions in DRE decreased the level of accumulation of the GUS mRNA by these stresses in the transformants (Figure 2, M1). The 120 bp fragment with base substitutions in ABRE did not respond to exogenous ABA (Figure 2, M2). The 120 bp fragment with base substitutions in both DRE and ABRE slightly exhibited the stress-induced accumulation of the GUS mRNA (Figure 2, M3). We also analyzed the monomer form of these DNA fragments with or without base substitutions, but the level of induction of the GUS mRNA was too weak to detect using RNA gel blotting (data not shown).

Figure 2.

RNA gel blot analysis of the effects of various treatments on the induction of the rd29A promoter-GUS fusion genes in transgenic Arabidopsis.

Schematics of the chimeric constructs are shown on the left. RNA gel blotting was carried out to measure the amount of GUS mRNA or endogenous rd29A mRNA in transgenic Arabidopsis that had either been dehydrated (dry) for 10 h, transferred from agar plates for hydroponic growth in water for 10 h (H2O), transferred from agar plates for hydroponic growth in 250 mm NaCl (NaCl) for 10 h, transferred from agar plates for hydroponic growth in 100 µm ABA (ABA) for 10 h, transferred at 4°C for 10 h (Cold), or untreated (Cont).

We analyzed the effects of base substitutions in the as1 sequence (Figure 2, M4–M6). The 120 bp fragment with base substitutions in as1 showed stress-induced expression of the GUS mRNA, similar to the wild-type 120 bp promoter (−174 to −55). The 120 bp fragment with base substitutions in both DRE and as1 showed weak induction of GUS expression by these stresses. The fragment with base substitutions in both as1 and ABRE functioned in response to low-temperature, high-salinity, and dehydration treatments, but not to ABA.

We then analyzed two deletion constructs containing promoter regions from −157 to −55 or from −135 to −55, concerning the stress-induced expression in transgenic Arabidopsis (Figure 2, D1 and D3). The tandem repeated dimer form of the fragment between −157 and −55 showed the stress-induced accumulation of the GUS mRNA as well as the wild-type 120 bp fragment (−174 to −55) of the rd29A promoter. The stress-induced GUS expression was also observed with the tandem repeated dimer form of the fragment between −135 and −55, but the level of induction by any stress treatment was very low. The region between −151 and −135 may contain an unknown cis-element, which influences the GUS activity as an enhancer.

Analysis of function of the DRE-like sequence in the rd29A promoter on the expression of the GUS reporter gene in transgenic plants in response to stresses

Base-substitution and deletion analyses indicated that the region between −135 and −124 might contain at least one cis-acting element responding to dehydration, high-salinity, and low-temperature stresses (Figures 1 and 2). The 103 bp (−157 to −55) fragment of the rd29A promoter contains a DRE-like sequence (AGCCGACAC; −130 to −122) that includes the DRE-core motif, GCCGAC (Figure 1a). Therefore, we examined the role of this DRE-core motif in the 103 bp fragment for dehydration-induced expression of the rd29A gene using transgenic tobacco plants. The monomer form of the 103 bp fragments with or without base substitutions in the ABRE-core motif showed dehydration-induced expression of GUS activity with 7.6- or 13.0-fold increase (Figure 3a). In contrast, the monomer form of the 103 bp fragments with base substitutions in the DRE-core motif or both the DRE-core motif and ABRE showed little GUS activity (1.3- or 0.9-fold, Figure 3a). These results suggested that the DRE-core motif might function as a DRE sequence. We also analyzed the dimer forms of the 103 bp fragments with or without base substitutions and obtained similar results to those obtained in case of the monomer forms of these fragments. However, the dimer form of the 103 bp fragment with base substitutions in the DRE-core motifs, which contains two ABREs, showed stress-responsive induction of the GUS activity (5.0-fold, Figure 3b). The dimer form of the 103 bp fragment with base substitutions in two DREs and one ABRE, which has only one ABRE, did not respond to dehydration stress (0.8-fold, Figure 3b).

Figure 3.

Base-substitution analysis of the 103 bp region of the rd29A promoter involved in the expression in response to dehydration in transgenic tobacco.

The 103 bp DNA fragment of rd29A promoter region (−157 to −55) that contained base substitutions in the DRE core or ABRE was fused to the GUS reporter gene and introduced into tobacco plants. Experimental conditions are the same as those described in Figure 1.

(a) The chimeric constructs with a monomer of the rd29A promoter fragment with or without base substitutions and their GUS activities in transgenic tobacco.

(b) The chimeric constructs with the tandemly repeated dimer of the rd29A promoter fragment and their GUS activities in transgenic tobacco.

Then, RNA gel blot analyses were performed using transgenic Arabidopsis plants containing the dimer forms of the 103 bp fragments with or without base substitutions (Figure 4). The dimer forms of the 103 bp fragments with base substitutions in the DRE-core motifs responded to dehydration, high-salinity, and ABA treatments, but not to low-temperature treatment (Figure 4, MD1). The dimer forms of the 103 bp fragment with base substitutions in ABREs responded to dehydration, high-salinity, and low-temperature treatments, but not to application of ABA (Figure 4, MD2). By contrast, the ABRE-mutated 103 bp fragment plus the wild-type 103 bp fragment, which had only one ABRE and two DRE-core motifs, showed expression of GUS by dehydration, high salinity, ABA, and low temperature (Figure 4, MD4). However, when we base-substituted in two DREs and one ABRE, which has only one ABRE, did not respond to these stresses at all (Figure 4, MD5). The dimer forms of the 103 bp fragment with base substitutions in two of the DRE-core motifs and two of the ABREs also did not respond to these stresses (Figure 4, MD3). These results indicated that both ABRE and DRE function as cis-acting elements interdependently for ABA-dependent expression of rd29A in response to dehydration and high salinity. These results also confirmed that DRE functions in dehydration-, high-salinity- and low-temperature-responsive gene expression in an ABA-independent manner whereas the two ABREs function in dehydration- and high-salinity-induced gene expression in an ABA-dependent manner.

Figure 4.

Northern blot analysis of the effects of various stress treatments on the induction of the GUS reporter gene fused to the 103 bp region of the rd29A promoter in transgenic Arabidopsis plants.

Schematics of the chimeric constructs are shown on the left. Northern blot analysis was used to measure the amount of GUS mRNA or endogenous rd29A mRNA in transgenic Arabidopsis plants. The plants had either been dehydrated (dry) for 10 h, transferred from agar plates for hydroponic growth in water for 10 h (H2O), transferred from agar plates for hydroponic growth in 250 mm NaCl (NaCl) for 10 h, transferred from agar plates for hydroponic growth in 100 µm ABA (ABA) for 2 h, transferred at 4°C for 10 h (Cold), or untreated (Cont).

The DREB and AREB proteins bind specifically to the cis-elements of the rd29A promoter

To identify the target sequence of the DREB and AREB proteins in the 120 bp promoter region, we expressed the DNA-binding domains of DREB and AREB as glutathione S-transferase (GST) fusion proteins in Escherichia coli. The ability of the DREB1A, DREB2A, AREB1, and AREB2 fusion proteins to bind to the 120 bp promoter fragment (WT) was examined using the gel mobility shift assay (Figure 5a,b; Liu et al., 1998). Both recombinant DREB1A and DREB2A fusion proteins bound to the 120 bp (−174 to −55) DNA fragment with base substitutions in ABRE (Figure 5a; F1). On the other hand, both fusion proteins also bound to the 120 bp fragment with base substitution in DRE or in both DRE and ABRE, which contained the DRE-core motif (F2 or F3). In addition, these proteins also bound to the 103 bp (−157 to −55) fragment with or without base substitutions in ABRE containing the DRE-core motif (F4 or F5). Furthermore, these proteins bound to the 120 bp DNA fragment containing the DRE and DRE-core motif (WT or F1) more strongly than to the 120 bp or 103 bp DNA fragments containing only the DRE-core motif (F4 or F5). However, these proteins did not bind to the 103 bp fragments with base substitutions in the DRE-core motif or in both the DRE-core motif and the ABRE (F6 or F7). These results indicated that DREB can bind to both the DRE and DRE-core motif in the 120 bp promoter region.

Figure 5.

The DNA binding affinity of the recombinant DREB1A, DREB2A, AREB1, and AREB2 proteins to the 120 bp (positions −174 to −55) and 103 bp (positions −157 to −55) fragments of the rd29A promoter.

Gel retardation assays were carried out to analyze the sequence-specific binding of the recombinant DREB1A, DREB2A, AREB1, and AREB2 proteins.

(a) The radioactive probes were incubated with the recombinant DREB1A or DREB2A.

(b) The probes were incubated with the AREB1 or AREB2 proteins.

Both the AREB1 and AREB2 fusion proteins bound to the 120 bp DNA fragment with or without a base substitution in DRE (Figure 5b; WT or F2). On the other hand, both the fusion proteins did not bind to the 120 bp fragment with base substitution in ABRE or in both DRE and ABRE (F1 or F3). In addition, these proteins also bound to the wild-type 103 bp fragment (−157 to −55) and the 103 bp fragment with base substitution in the DRE-core motif, which contained one ABRE sequence (F4 and F6), but did not bind to the 103 bp fragments with base substitutions in ABRE or in both the DRE-core motif and ABRE (F5 or F7). The AREB proteins can bind to the ABRE in the 120 bp promoter region.

The DREB and AREB proteins transactivate the rd29A promoter-GUS fusion gene

To determine whether the DREB and AREB proteins are capable of transactivating the transcription driven by the 120 bp DNA (−174 to −55) fragment of the rd29A promoter, we performed transactivation experiments using protoplasts prepared from Arabidopsis leaves. Protoplasts were transfected with a GUS reporter gene fused to the 120 bp fragments of the rd29A promoter and effector plasmid (Figure 6a). The effector plasmids consisted of the cauliflower mosaic virus 35S promoter fused to the DREB1A, DREB2A, AREB1, or AREB2 cDNAs. The tobacco mosaic virus Ω sequence was inserted upstream of these cDNAs to strengthen their translation level.

Figure 6.

Transactivation of the rd29A promoter-GUS fusion gene by the AREB1, AREB2, DREB1A, and DREB2A proteins using Arabidopsis protoplasts.

(a) Schematic diagram of the effector and reporter constructs used in co-transfection experiments. The effector constructs contained the CaMV 35S promoter and TMV Ω sequence fused to the AREB1, AREB2, DREB1, or DREB2 cDNAs. The reporter construct contained the 120 bp fragments of the rd29A promoter. The promoter was fused to the −61 rd29A minimal TATA promoter-GUS construct.

(b) Transactivation of the rd29A promoter-GUS fusion gene by the AREB1, AREB2, DREB1, or DREB2 proteins. The reporter gene was transfected with each effector plasmid or the vector as a control treatment. Transactivation experiments were performed using protoplasts prepared from Arabidopsis leaves. The transfection efficiency was normalized by co-transfecting with the CaMV 35S promoter-luciferase (LUC) plasmid in each experiment. Ratios indicate the multiples of expression compared with the value obtained with the pBI35S Ω vector.

The DREB2A gene was induced by dehydration and high-salinity treatments, whereas the DREB1A was induced by low-temperature treatment (Liu et al., 1998). The AREB1 and AREB2 genes were induced by dehydration, high-salinity, and ABA treatments (Uno et al., 2000). Co-expression of each protein in the protoplasts transactivated the expression of the GUS reporter gene (Figure 6b). Co-expression of the AREB and DREB proteins transactivated the expression of the GUS reporter gene more strongly than the expression of each of AREB1, AREB2, DREB1A, or DREB2A (Figure 6b, AREB1/DREB1A and AREB2/DREB1A). These results suggest that AREBs and DREBs function as interactive transcription activators in the expression of the rd29A gene by dehydration, high-salinity, and ABA treatments.

Discussion

We have reported that the DRE sequence in the 120 bp promoter region (−174 to −55) of rd29A functions as a cis-acting element in the expression of rd29A in response to dehydration treatment (Yamaguchi-Shinozaki and Shinozaki, 1994). Deletion and base-substitution analyses of the promoter region of rd29A in transgenic tobacco and Arabidopsis indicated that other cis-acting elements also function in its expression in response to dehydration (Figures 1 and 2), and ABRE functions at least as a cis-acting element in its ABA-dependent expression in response to dehydration (Figure 2, M2). The 103 bp (−157 to −55) region of the rd29A promoter without DRE still appeared to function in the expression in response to dehydration, high-salinity, and low-temperature stresses. We precisely analyzed the cis-acting element in the 103 bp (−157 to −55) region at the nucleotide sequence level. We found a DRE-like sequence (AGCCGACAC) containing an A/GCCGAC DRE-core motif in the 103 bp region (−157 to −55). The base-substitution analysis revealed that the DRE-core sequence also functions as a DRE (Figures 3 and 4).

The dimer forms of the 103 bp fragment with base substitutions in the DRE-core motifs responded to dehydration, high-salinity, and ABA treatments, but not to low-temperature treatment (Figure 4, MD1). This expression pattern was similar to that of the rd29B gene in which the promoter contains two ABREs but not the DRE or DRE-core sequences (Yamaguchi-Shinozaki and Shinozaki, 1994). Promoter analysis of the rd29B gene demonstrated that at least these two ABREs function as positive, cis-acting elements in the expression of rd29B in response to dehydration and ABA treatments (Uno et al., 2000). The two ABREs in the dimer forms of the 103 bp fragment also function as cis-acting elements in the expression in response to dehydration, high-salinity, and ABA treatments (Figure 4, MD1).

Although the 103 bp fragment with base substitutions in ABRE plus the wild-type 103 bp fragment contains one ABRE and two DRE-core motifs, but none of the known coupling elements, it responded to the application of ABA (Figure 4, MD4). The 103 bp fragment with base substitutions in DRE core and ABRE plus the 103 bp fragment with base substitutions in the DRE core, whose construct contains only one ABRE, did not respond to ABA (Figure 4, MD5). These results indicate that the DRE or DRE-core sequences function as a coupling element of ABRE in response to ABA.

We showed that two DREB proteins, DREB1A and DREB2A, bound to the DRE or DRE core, and that two AREB proteins, AREB1 and AREB2, bound to the ABRE in the 120 or 103 bp fragments (Figure 5). Transient expression in protoplasts showed that the level of transactivation by a combination of AREBs and DREBs was greater than that by either of them alone (Figure 6). The AREB1, AREB2, and DREB2A genes are induced by drought and high-salinity treatments, whereas DREB1A is not induced by these stresses (Liu et al., 1998; Uno et al., 2000). Therefore, the combination of DREB2A and an AREB is considered to function mainly in the expression of ABA-dependent gene expression of rd29A (Liu et al., 1998). However, the levels of transactivation using DREB1A and AREB1 or AREB2 were greater than those using DREB2A and AREB1 or AREB2 (Figure 6b). DREB1/CBF homologs, DREB1D/CBF4 and DREB1F, are induced by high-salinity stress whereas DREB1A is weakly induced by high-salinity stress (Haake et al., 2002; Sakuma et al., 2002). These DREB1-type proteins may function with AREB in the expression of rd29A under high-salinity conditions.

Most of the genes induced by low temperature are also responsive to ABA, drought, and high-salinity stresses. These genes include kin1, cor6.6/kin2, cor15A, rd17/cor47, and rd29A/cor78/lti78 (Seki et al., 2001; Thomashow, 1999). The promoter regions of these genes contain both DRE and ABRE. We have shown that DRE and DRE-core motifs are sufficient for the expression of the rd29A gene under dehydration, high-salinity, and low-temperature conditions, and that ABRE is necessary for the expression of rd29A in response to ABA (Yamaguchi-Shinozaki and Shinozaki, 1994; Figures 2 and 4). Induction of the genes by low temperature was not largely affected in both ABA-deficient and ABA-insensitive mutants in Arabidopsis (Gilmour et al., 1991; Nordin et al., 1991). The dimer forms of the 103 bp fragment with base substitutions in the DRE-core motifs responded to dehydration, high-salinity, and ABA treatments, but not to the low-temperature treatment (Figure 4, MD1). These results suggest that low-temperature stress induces gene expression through signaling pathways different from that of ABA.

Figure 7 shows a schematic model of the signal transduction pathways between initial signals of environmental stresses and the expression of rd29A. The rd29A promoter contains at least two kinds of cis-acting elements that are involved in the induction of rd29A by dehydration, high salinity, or low temperature. One of these elements is DRE which functions in response to low-temperature, dehydration, and high-salinity stresses. The rd29A promoter contains three functional DRE motifs. Expression of the DREB1 genes is induced by low temperature, and the accumulated DREB1 proteins in turn induce the DRE-dependent gene expression of rd29A (Liu et al., 1998). DRE also functions in the first rapid (within 20 min) response of rd29A to dehydration and high-salinity stresses. (Yamaguchi-Shinozaki and Shinozaki, 1993, 1994). The DREB2 proteins are activated under dehydration and high-salinity conditions and then induce the DRE-dependent expression of rd29A. ABA is not involved in this rapid process, as ABA biosynthesis is not observed within 1 h of dehydration (Kiyosue et al., 1994).

Figure 7.

A model for the induction of rd29A gene expression in response to dehydration, high-salinity, and low-temperature stresses.

There are several independent signal transduction pathways between the perception of environmental stress and expression of the rd29A gene. DRE functions in the ABA-independent pathway, and ABRE is a cis-acting element in the ABA-dependent induction of rd29A in response to ABA. Two independent families of DRE binding proteins, DREB1/CBF and DREB2, function as trans-acting factors in two separate signal transduction pathways; DREB1 in response to low-temperature stress and DREB2 in response to drought and high-salinity stresses. DRE also functions as a coupling element of ABRE in ABA-dependent gene expression. The ABRE binding proteins, AREB1 and AREB2, bind to ABRE in the rd29A promoter. The DREB proteins may co-operate with the AREB proteins in ABA-dependent gene expression of rd29A. Black arrows indicate signal transduction pathways and white arrow represents protein–protein interaction.

The other cis-acting element is ABRE. This element functions in the second slow induction of rd29 after the accumulation of ABA under dehydration and high-salinity stresses. ABA biosynthesis is induced within 2 h of dehydration and high-salinity stresses (Kiyosue et al., 1994), and the accumulated ABA induces the expression of rd29A. Expression of the AREB genes is induced by dehydration and high-salinity stresses, and then the accumulated AREB proteins function as transcriptional activators in the ABA-dependent expression of rd29A. However, interaction between the AREB/ABRE and DREB/DRE regulatory systems is necessary for the slow expression of rd29A in response to endogenous ABA under drought and high-salinity conditions. Genetic studies using the mutants also suggested that the pathways do share certain components, although separate signaling pathways exist for osmotic stress and ABA (Ishitani et al., 1997; Xiong et al., 1999). These common components may mediate the synergistic interaction between osmotic stress and ABA response. This study suggests that DREB interacts with AREB in ABA-induced slow gene expression in vegetative tissues under dehydration stress conditions.

Experimental procedures

Plant growth and stress treatment

Arabidopsis thaliana (Columbia ecotype) was grown on germination medium (GM) agar plates (Valvekens et al., 1988) containing 0.8% of Bacto-agar (Difco, Detroit, MI, USA) at 22°C for 3 weeks under a 16 h light (3000 lux)/8 h dark cycle and was used in stress treatment experiments prior to bolting. Arabidopsis rosette plants were harvested from GM agar plates and were then dehydrated in Petri dishes at 22°C and 60–70% humidity under dim light. Plants subjected to treatment with ABA and to high-salinity stress were grown hydroponically in a solution containing 100 µm ABA and 250 mm NaCl, respectively, under dim light. Low-temperature treatment was conducted by exposure of plants grown at 22°C to a temperature of 4°C under dim light. Plants were subjected to each stress treatment for 2 or 10 h and were frozen in liquid nitrogen.

Tobacco plants were grown on Murashige and Skoog medium containing 0.8% of Bacto-agar (Difco, Detroit, MI, USA) at 25°C for 3 weeks under a 16 h light (3000 lux)/8 h dark cycle and were used in stress treatment experiments. Some leaves from transgenic tobacco plants were dehydrated for 24 h and used for assay of GUS activity.

Construction of deleted or base-substituted promoter regions to a β-glucuronidase reporter gene

The construction of various chimeric genes with the rd29A promoter fused to a GUS reporter gene has been previously described (Yamaguchi-Shinozaki and Shinozaki, 1993, 1994). The fusion gene contains a 120 bp (−174 to −55) region of the rd29A promoter, 134 bp of the minimal TATA sequence containing the TATA box and untranslated leader sequence (−54 to +81), and a 15 bp fragment of the coding region of rd29A.

A 120 bp (−174 to −55) fragment of rd29A (wild type), its base-substituted fragments, and deletion fragments were prepared by polymerase chain reaction (PCR) and were subcloned into the pBluescript II SK (pSK-GX1) vector (Stratagene, La Jolla, CA, USA) containing a 134 bp fragment of the minimal TATA sequence containing the TATA-box and untranslated leader sequence (−54 to +81), and a 15 bp fragment of the coding region of rd29A. These constructs were ligated into the BamHI–SalI site of the promoterless GUS expression vector pBI101 (CLONTECH, Palo Alto, CA, USA). The structures of the fusion constructs were confirmed by sequencing the boundary sites of the fused gene from both the multicloning site in pBI101 (primer: 5′-CTCGTATGTTGTGTGGAATTGT-3′) and a sequence from the GUS coding region (primer: 5′-GATTTCACGGGTTGGGGTTTCT-3′) as primers (Yamaguchi-Shinozaki and Shinozaki, 1994).

Transgenic plants

The vectors containing the promoter-GUS fusion constructs were then transferred from E. coli DH5α into Agrobacterium tumefaciens via triparental mating with an E. coli strain that contained a mobilization plasmid pRK2013. The pBI101 vectors containing the desired rd29A promoter-GUS fusion constructs were transferred into A. tumefaciens C58 for transformation of Arabidopsis and into Agrobacterium strain LBA4404 for transformation of tobacco as described previously (Yamaguchi-Shinozaki and Shinozaki, 1994). Arabidopsis plants (Columbia ecotype) were transformed by the vacuum infiltration method as described by Bechtold et al. (1993). Transformation of Nicotiana tabacum cv SR1 was performed as described previously (Yamaguchi-Shinozaki and Shinozaki, 1994). Transgenic tobacco and Arabidopsis plants were grown at 25 and 22°C, respectively, under a 16 h light/8 h dark cycle.

Assays of GUS activity

GUS activity was assayed as described previously (Yamaguchi-Shinozaki and Shinozaki, 1994).

RNA gel blot analysis of the accumulation of the GUS mRNA in transgenic Arabidopsis

Transgenic Arabidopsis rosette plants grown in GM agar plates were subjected to dehydration, low temperature, or high salinity for 10 h, or ABA treatment for 2 h (Yamaguchi-Shinozaki and Shinozaki, 1994) and were frozen in liquid nitrogen. The rd29A gene is strongly expressed 10 h after exposure to dehydration, high salinity, and low temperature. On the other hand, the maximum level of the rd29A mRNA is detected 2 h after exposure to ABA, and subsequently the mRNA level decreases. Therefore, Arabidopsis plants were subjected to stress treatments for 2 or 10 h, depending on the experiments. To detect GUS mRNA by ABA treatment, plants were subjected to ABA treatment for 2 h. RNA was extracted from the frozen plants as described previously (Yamaguchi-Shinozaki and Shinozaki, 1994). RNA gel blot hybridization was performed as described above, using the DNA fragments containing the coding region of the GUS reporter gene or a fragment containing the full length of rd29A gene.

Gel mobility shift assays

The GST fusion proteins, DREB1A, DREB2A, AREB1, and AREB2, were produced and purified as described previously (Liu et al., 1998). The 120 and 103 bp fragments containing DRE and ABRE of the rd29A promoter with or without base substitutions were labeled with [α-32P]dCTP, as described previously (Urao et al., 1993).

Transactivation experiments with protoplasts

Effector plasmids used in the transient transactivation experiment were constructed with DNA fragments containing the DREB1A, DREB2A, AREB1, or AREB2 coding regions, which were cloned into NotI sites of the plant expression vector pBI35SΩ. The pBI35SΩ vector was constructed as described previously (Abe et al., 1997). To construct a reporter plasmid, we replaced the 35S promoter of pBI221 (Yamaguchi-Shinozaki and Shinozaki, 1994) with the rd29A minimal TATA promoter, and then ligated the 120 bp fragment of the rd29A promoter into the HindIII site located upstream of the rd29A minimal TATA promoter as one copy. A transactivation experiment with Arabidopsis mesophyll protoplasts was performed as described previously (Mizoguchi et al., 1993).

Acknowledgements

The authors thank Ms Ekuko Ohgawara, Ms Mie Yamamoto, Ms Fumie Saito, Ms Kyoko Murai, and Fumie Saitoh for their excellent technical assistance. This work was supported in part by the Program for Promotion of Basic Research Activities for Innovative Biosciences. This work was also supported in part by a project grant from the Ministry of Agriculture, Forestry, and Fisheries, Japan.

Ancillary