OPEN ACCESS ARTICLE

This Article Free via Open Access: OA OA Abstract Full Text (PDF) Supplemental Data OA All Versions of this Article: 17/12/3470    most recent tpc.105.035659v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (28) Citing Articles via Google Scholar Google Scholar Articles by Fujita, Y. Articles by Yamaguchi-Shinozaki, K. Search for Related Content PubMed PubMed Citation Articles by Fujita, Y. Articles by Yamaguchi-Shinozaki, K. Agricola Articles by Fujita, Y. Articles by Yamaguchi-Shinozaki, K.

AREB1 Is a Transcription Activator of Novel ABRE-Dependent ABA Signaling That Enhances Drought Stress Tolerance in Arabidopsis^[W],[OA]

^a Biological Resources Division, Japan International Research Center for Agricultural Sciences, Tsukuba, Ibaraki 305-8686, Japan
^b Laboratory of Plant Molecular Biology, RIKEN Tsukuba Institute, Tsukuba, Ibaraki 305-0074, Japan
^c Core Research for Evolutional Science and Technology, Japan Science and Technology Agency, Kawaguchi, Saitama 332-0012, Japan
^d Plant Functional Genomics Group, RIKEN Genomic Sciences Center, Yokohama, Kanagawa 230-0045, Japan
^e Gene Function Research Center, National Institute of Advanced Industrial Science and Technology, Central 4, Tsukuba, Ibaraki 305-8562, Japan
^f Laboratory of Plant Molecular Physiology, Graduate School of Agricultural and Life Sciences, University of Tokyo, Tokyo 113-8657, Japan

¹ To whom correspondence should be addressed. E-mail kazukoys{at}jircas.affrc.go.jp; fax 81-29-838-6643.

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
ABSCISIC ACID–RESPONSIVE ELEMENT BINDING PROTEIN1 (AREB1) (i.e., ABF2) is a basic domain/leucine zipper transcription factor that binds to the abscisic acid (ABA)–responsive element (ABRE) motif in the promoter region of ABA-inducible genes. Here, we show that expression of the intact AREB1 gene on its own is insufficient to lead to expression of downstream genes under normal growth conditions. To overcome the masked transactivation activity of AREB1, we created an activated form of AREB1 (AREB1QT). AREB1QT-overexpressing plants showed ABA hypersensitivity and enhanced drought tolerance, and eight genes with two or more ABRE motifs in the promoter regions in two groups were greatly upregulated: late embryogenesis abundant class genes and ABA- and drought stress–inducible regulatory genes. By contrast, an areb1 null mutant and a dominant loss-of-function mutant of AREB1 (AREB1:RD) with a repression domain exhibited ABA insensitivity. Furthermore, AREB1:RD plants displayed reduced survival under dehydration, and three of the eight greatly upregulated genes were downregulated, including genes for linker histone H1 and AAA ATPase, which govern gene expression and multiple cellular activities through protein folding, respectively. Thus, these data suggest that AREB1 regulates novel ABRE-dependent ABA signaling that enhances drought tolerance in vegetative tissues.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Water deficit leads directly to fatal damage in all living things. Hence, the molecular machinery involved in crucial stress responses has developed redundant and complex signal transduction networks that overcome and circumvent fatal injury. However, the inherent complexity and redundancy cause difficulties in the analysis of the key signal transduction pathways. In this study, to overcome this problem, we have used two approaches: (1) an active form of a transcription factor to activate expression of target genes without an originally required posttranscriptional modification and (2) a transcription factor fused to a repression domain (RD) that consists of only 12 amino acids to overcome potential functional redundancy conferred by homologs (Hiratsu et al., 2003).

The plant hormone abscisic acid (ABA) regulates many essential processes, including inhibition of germination, maintenance of seed dormancy, control of stomatal closure, and adaptive responses to a variety of environmental stresses (for review, see Finkelstein et al., 2002). In Arabidopsis thaliana, for example, analyses of ABA-hypersensitive mutants have revealed that ABA is involved in cellular processes such as farnesylation (era1), inositol signaling (fry1), and RNA metabolism (abh1, sad1, and hyl1). In addition, genetic screens for mutations that display a reduced sensitivity to ABA have identified homologous type 2C phosphatases (ABI1 and ABI2) and three different classes of transcription factors (ABI3, ABI4, and ABI5).

ABA is produced under dehydration conditions and plays pivotal roles in response to drought stress (Shinozaki and Yamaguchi-Shinozaki, 2000; Finkelstein et al., 2002; Xiong et al., 2002). Numerous drought stress–inducible genes have been reported in vegetative tissues, and many of them are also activated by ABA (Ingram and Bartels, 1996; Seki et al., 2002). In analyses of the promoters of such ABA-regulated genes, a conserved cis-element designated ABA-responsive element (ABRE; PyACGTGGC), which controls ABA-regulated gene expression, has been identified (Bray, 1994; Giraudat et al., 1994; Busk and Pages, 1998). Various types of ABRE-like sequences have been reported, including the G-box sequence (CACGTG) and coupling element (CGCGTG), CE3, hex3, and motif III (Shen et al., 1996; Busk and Pages, 1998). So far, all isolated interactors with divergent types of ABRE sequences belong to the basic domain/leucine zipper (bZIP) class of transcription factors and can bind to them at least in vitro (Foster et al., 1994; Busk and Pages, 1998).

The RD29B promoter region carries two ABRE sequences, and the drought-inducible expression of RD29B is controlled mainly by ABA, according to analyses in the ABA-deficient and -insensitive mutants aba1 and abi1, respectively (Koornneef et al., 1984, 1992; Yamaguchi-Shinozaki and Shinozaki, 1994). Using yeast one-hybrid screening with the RD29B promoter including the ABRE sequences, we cloned three different cDNAs encoding ABRE binding proteins (AREB1, AREB2, and AREB3) of Arabidopsis (Uno et al., 2000). Expression of AREB1 and AREB2 is upregulated by ABA and by drought and high-salinity stresses. Both AREB1 and AREB2 function as trans-acting activators, as identified by transient expression analysis in protoplasts (Uno et al., 2000). In the Arabidopsis genome, to date, nine AREB homologs have been identified: AREB1/ABF2, AREB2/ABF4, AREB3/DPBF3, ABF1, ABF3/DPBF5, ABI5/DPBF1, EEL/DPBF4, DPBF2, and AT5G42910 (Choi et al., 2000; Finkelstein and Lynch, 2000; Lopez-Molina and Chua, 2000; Uno et al., 2000; Bensmihen et al., 2002; Jakoby et al., 2002; Kim et al., 2002; Suzuki et al., 2003). In this study, we show that expression of only three members, AREB1/ABF2, AREB2/ABF4, and ABF3/DPBF5, is induced by ABA, drought, and high salinity in vegetative tissues. Because AREB1 has the highest ABA and drought inducibility in RNA gel blot, histochemical, and transient analyses, we focus on AREB1/ABF2 in order to elucidate its role in ABA-dependent responses to drought in vegetative tissues.

Here, we report that AREB1 is a key positive regulator of ABA signaling in vegetative tissues under drought stress. Our results indicate that AREB1 directs expression of ABA- and dehydration-inducible regulatory genes such as linker histone H1-3 and AAA ATPase genes, as well as late embryogenesis abundant (LEA) class genes, which are thought be involved in alleviation of water stress. Also, we demonstrate that two powerful tools, a constitutive active form and a dominant loss-of-function mutant with an RD, are useful for dissecting important transcription factors that seem to be required for posttranscriptional modification or in which there is expected to be phenotypic masking due to potential functional redundancy.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Three bZIP Proteins Are Involved in the ABA-Mediated Signal Transduction Pathway under Drought and High-Salinity Conditions
The Arabidopsis AREB/ABF/ABI5/DPBF bZIP subfamily consists of nine members that contain three N-terminal (C1, C2, and C3) and one C-terminal (C4) conserved domains (Bensmihen et al., 2002; Jakoby et al., 2002; Kim et al., 2002; Suzuki et al., 2003) (Figure 1A). To identify members that are involved in an ABA signal transduction pathway under drought, we conducted RNA gel blot analysis of the nine genes. The expression of three genes, AREB1/ABF2, AREB2/ABF4, and ABF3/DPBF5, was induced by dehydration, high-salt, and exogenous ABA treatments (Figure 1B). These three members are in the same clade in the phylogenetic tree of the nine members of the subfamily (Figure 1C).

View larger version (79K): [in this window] [in a new window]   Figure 1. Expression of the AREB1 Gene and Subcellular Localization of the AREB1 Protein. (A) Structure of AREB1 family proteins. NLS, nuclear localization signal. C1 to C4 indicate conserved domains within the family. (B) Expression profiles of AREB family genes in response to dehydration, high salt, or ABA. Each lane was loaded with 20 µg of total RNA from 3-week-old Arabidopsis plants that had been dehydrated (DRY), transferred to hydroponic growth in 250 mM NaCl (NaCl), transferred to hydroponic growth in 100 µM ABA (ABA), or transferred to water (H2O). rRNAs are shown as equal loading controls. A band located in the center of each column indicates a transcript that corresponds to each gene. (C) Phylogenetic tree of AREB family proteins. Proteins were aligned using ClustalX software, and the tree was constructed using MEGA software. (D) Nuclear localization of AREB1 protein in onion epidermal cells: fluorescent images of GFP, fluorescent images stained with propidium iodide (PI), and merged images (GFP/PI). (E) Patterns of AREB1 promoter-driven GUS expression in seedlings at different ages or in different tissues: (a) 2-d-old seedling, (b) 5-d-old seedling, (c) cotyledon, (d) primary leaf, (e) 2-week-old seedling, (f) 2-week-old seedling treated with 50 µM ABA, (g) flower, (h) immature silique, (i) seeds from (h). Bars = 0.5 mm in (a) to (d) and (g) to (i) and 5.0 mm in (e) and (f). (F) Expression of AREB1 and RD29B in wild-type and 35S-AREB1 plants (line 6) induced by 50 µM ABA treatment. Representative data are shown. Each lane was loaded with 15 µg of total RNA from 2-week-old Arabidopsis plants. rRNAs are shown as equal loading controls.

AREB1 Is Localized in the Nucleus
The AREB1 nuclear localization signal is located in the basic region of the bZIP DNA binding domain (Figure 1A). To examine the subcellular localization of the AREB1 protein in plant cells, we fused the AREB1 coding region in frame to the coding region for the C-terminal side of green fluorescent protein (GFP), and the fusion gene was expressed under the control of the 35S promoter of Cauliflower mosaic virus (CaMV). Onion epidermal cells transformed with an expression plasmid for the GFP:AREB1 fusion protein exhibited GFP fluorescence in the nucleus (Figure 1D). By contrast, GFP fluorescence was observed in the entire region of the cell when intact GFP was expressed. These results show that AREB1 is localized in the nucleus.

AREB1 Is Expressed Constitutively in Roots, Leaf Vascular Tissues, and Hydathodes or in All Tissues under Stress Conditions
Previously, we detected weak basal AREB1 expression in roots and leaves but not in seeds in an RNA gel blot analysis (Uno et al., 2000). Here, to determine the temporal and spatial expression patterns of AREB1 in more detail, we analyzed transgenic Arabidopsis plants expressing an AREB1 promoter–ß-glucuronidase (GUS) transgene. GUS expression was observed in roots at all developmental stages that we assessed (Figure 1E, a, b, e, and f). In unstressed plants, leaf vascular tissues and hydathodes also exhibited AREB1 promoter activity (Figure 1E, c to e). By contrast, ABA or drought treatment of seedlings enhanced the AREB1 promoter activity in all tissues (Figure 1E, f; data not shown). In mature plants, GUS activity was observed also in anthers, stigma, and siliques (abscission zone and carpels) (Figure 1E, g and h). The false septum and carpels were stained gradually from the abscission zone to the remains of the stigma with maturity (Figure 1E, h and i).

Expression of AREB1 on Its Own Is Insufficient to Induce the Expression of the Downstream Gene RD29B
To assess the in vivo function of AREB1, we generated transgenic plants expressing the AREB1 cDNA under the control of the CaMV 35S promoter (35S-AREB1). Thirty-six T3 homozygous lines were obtained, and eight transgenic lines with higher expression levels of the transgene were selected for further analysis by RNA gel blot analysis.

No obvious difference was observed in growth phenotypes between the wild-type and 35S-AREB1 plants growing on GM agar plates containing 1 or 3% sucrose, implying the possibility that constitutive overexpression of intact AREB1 on its own is not sufficient to activate downstream genes such as RD29B. To examine this possibility, we monitored RD29B mRNA levels in the 35S-AREB1 transgenic plants at several time points with or without exogenous ABA. As shown in Figure 1F, constitutive overexpression of AREB1 did not activate expression of the downstream RD29B in the absence of exogenous ABA (time point 0). Within 2 h after the addition of ABA, RD29B was expressed in the 35S-AREB1 plants but not in the wild-type plants. These results indicate that not only the presence of AREB1 in plants but also the exogenous addition of ABA is required for the expression of downstream genes such as RD29B. These data also suggest that the preexistence of AREB1 before the addition of ABA contributed to earlier accumulation of RD29B mRNA in the 35S-AREB1 plants than in the wild-type plants. Taken together with our previous report of ABA-dependent phosphorylation of the N-terminal region of AREB1 and suppression of the activation of AREB1 by protein kinase inhibitors in a transient assay using protoplasts (Uno et al., 2000), in addition to the ABA-dependent expression of AREB1, these results indicate that the ABA-induced modification of the AREB1 protein seems to be also required for the expression of its downstream genes.

The N-Terminal Conserved Region of AREB1 Has Transactivation Activity in Protoplasts
In a previous study, we showed that the AREB1 protein transactivates the RD29B promoter–GUS fusion gene (RD29B-GUS) in Arabidopsis protoplasts (Uno et al., 2000). To identify the transcriptional activation domain of AREB1, we constructed a series of effector plasmids bearing N-terminal deletion mutants of AREB1 under the control of the constitutive CaMV 35S promoter (Figure 2). The effector plasmids were cotransfected into protoplasts prepared from Arabidopsis T87 cultured cells, with a reporter plasmid, RD29B-GUS, carrying a GUS reporter gene fused to five tandem copies of a 77-bp fragment of the RD29B promoter containing two ABRE motifs (Figure 2A). A small deletion of 60 amino acids (region P) from the N terminus of AREB1 produced a significant decrease in the transactivation of the reporter gene in protoplasts treated either with or without 100 µM ABA, suggesting the presence of a positive regulatory domain in this region (Figure 2B).

View larger version (31K): [in this window] [in a new window]   Figure 2. The N-Terminal Conserved Region of AREB1 Functions as a Transcriptional Activation Domain in Protoplasts Derived from Arabidopsis T87 Cultured Cells. All transactivation experiments were performed 3 to 10 times, and results from one representative experiment are shown. Bars indicate standard deviation; n = 3 to 5. (A) Scheme of the effector and reporter constructs used in the transactivation analysis with AREB1 bZIP DNA binding domain. The effector constructs contain the CaMV 35S promoter and TMV sequence fused to AREB1 cDNA fragments encoding different portions of AREB1. The reporter construct, RD29B-GUS, contains 77-bp fragments of the RD29B promoter connected tandemly five times. The promoters were fused to the –51 RD29B minimal TATA promoter–GUS construct. Nos-T, nopaline synthase terminator. (B) Transactivation domain analysis of AREB1 using N-terminal deletion constructs. Protoplasts were cotransfected with the RD29B-GUS reporter and the effector construct (shown on the left) carrying an N-terminal truncated form of AREB1 cDNA or pBI-35S (vector). To normalize for transfection efficiency, the pBI35S-LUC reporter was cotransfected as a control in each experiment. Bars indicate standard deviation of three replicates. "Relative activity" indicates the multiples of expression compared with the value obtained with the pBI221-35S vector control. Top numbers indicate amino acid numbers of AREB1. P, Q, R, S, T, and U indicate the partial region of the AREB1 cDNA. The region P, Q, R, or U includes a conserved domain (solid rectangle), C1, C2, C3, or C4, respectively, in Figure 1A. (C) Schematic diagram of the effector and reporter constructs used in the transactivation analysis with the GAL4 DNA binding domain. The effector plasmids encoding the GAL4 DNA binding domain are fused to AREB1 cDNA fragments encoding different portions of AREB1. The GUS reporter construct, GAL4-GUS, containing GAL4 binding sites, is fused to the minimal promoter of CaMV 35S. (D) N-terminal conserved P region of AREB1 contains sufficient domain for transcriptional activation. Protoplasts were cotransfected with the GAL4-GUS reporter and the effector construct expressing a portion of AREB1 or the vector DNA. (E) AREB1QT is a constitutive active form of AREB1. Protoplasts were cotransfected with the RD29B-GUS reporter and the effector construct expressing intact AREB1, AREB1QT, or AREB1P/RT, or vector DNA.

We further analyzed whether an AREB1 mutant protein containing the P region and its native binding domains transactivates the RD29B-GUS reporter gene without exogenous ABA treatment. AREB1QT and AREB1P/RT were constructed as effector plasmids, carrying the AREB1 internal deletion mutants containing the bZIP DNA binding domain of AREB1 and region P or Q, respectively (Figure 2E). Cotransfection of AREB1QT together with RD29B-GUS resulted in a significant activation of the GUS reporter gene even in the absence of ABA, but that of AREB1P/RT did not activate the reporter gene either with or without ABA. These results indicate that the N-terminal P region of AREB1 contains a transcriptional activation domain, and AREB1QT is a constitutive active form of AREB1 in protoplasts (Figure 2E). Also, significant reduction of activation in the AREB1P/RT deletion mutants or most of the N-terminal deletion mutants of AREB1 compared with the vector control in the presence of ABA suggests that overexpression of the AREB1P/RT or N-terminal deletion mutants dominantly inhibits ABA-induced binding of endogenous transcription factors to the ABRE motifs of the reporter plasmids (Figures 2B and 2E). Among N-terminal deletion mutants of AREB1, however, only the deletion mutant lacking the P region does not seem to repress the activation of the reporter gene in the presence of ABA, suggesting that the deletion mutant lacking the P region may not dominantly inhibit such ABA-induced binding owing to the sequence-specific effect of the deletion mutant (Figure 2B).

AREB1QT Is a Constitutive Active Form of AREB1 in Planta
Our previous study showed that the AREB1 protein binds to two ABRE sequences in the RD29B promoter and activates expression of the gene (Uno et al., 2000). As described above, overexpression of intact AREB1 does not activate the expression of the downstream RD29B gene in the absence of exogenous ABA. We then analyzed whether AREB1QT overexpression in Arabidopsis plants activates the transcription of downstream genes such as RD29B in the absence of exogenous ABA. Transgenic plants expressing the AREB1QT cDNA under the control of the CaMV 35S promoter were generated, and expression of the AREB1QT transgene and the downstream RD29B gene was analyzed in 33 independent transgenic lines under unstressed conditions. The accumulation levels of both AREB1QT and RD29B mRNA were elevated in all examined lines in the absence of exogenous ABA (data not shown); so, we selected eight transgenic lines with higher AREB1QT expression levels for phenotypic analysis. Because the eight transgenic lines behaved in a similar manner (data not shown) and because the number of seeds in each line was sometimes not enough to allow statistical analysis, in some cases, we used different lines in different experiments. Overexpression of AREB1QT in plants activated the expression of the downstream RD29B gene in the absence of exogenous ABA (Figure 3A). This result shows that AREB1QT is a constitutive active form of AREB1 even in whole plants and that the N-terminal P region of AREB1 functions as a transcriptional activation domain in plants as well as in protoplasts. At 3 weeks after stratification, the maximum rosette radius of the 35S-AREB1QT plants averaged 70% of that of the wild-type plants (Figures 3B and 3C). The 35S-AREB1QT plants were slightly smaller than the wild type throughout their life (data not shown). By contrast, the 35S-AREB1 transgenic plants showed a similar growth phenotype to that of the wild type (Figure 3B).

View larger version (81K): [in this window] [in a new window]   Figure 3. AREB1QT Is a Constitutive Active Form of AREB1 in Planta. (A) RNA gel blot analysis of AREB1 and RD29B expression in wild-type, vector control (vector), 35S-AREB1, and 35S-AREB1QT plants. Two-week-old seedlings were either not treated (–) or treated (+) with ABA for 7 h. Each lane contained 10 µg of total RNA. Two lines of the 35S-AREB1 plants (3 and 6) and three lines of the 35S-AREB1QT plants (5, 12, and 26) are shown. rRNAs on ethidium bromide–stained gel are shown as equal loading controls. (B) Growth phenotype of 35S-AREB1QT (line 5) and 35S-AREB1 (line 6) plants that were grown for 3 weeks on GM agar plates containing 1% sucrose. (C) Maximum rosette radius (i.e., length of the longest rosette leaf) of each plant on a GM agar plate containing 3% sucrose was measured 3 weeks after stratification. Three independent lines of wild-type plants and nine independent lines of 35S-AREB1QT plants were used. Bars indicate standard deviation; n = 7. (D) Expression profile of downstream genes identified by microarray analysis (Table 1) in 35S-AREB1QT plants (line 5). Two-week-old seedlings were either not treated (–) or treated (+) with ABA for 7 h. Each lane contained 7 µg of total RNA. Three to eight independent lines were used, and results from one representative experiment are shown. rRNAs on ethidium bromide–stained gel are shown as equal loading controls.

View this table: [in this window] [in a new window]   Table 1. Genes Upregulated in Plants Overexpressing AREB1QT, Identified by Microarray Analysis

Overexpression of AREB1QT Leads to Expression of LEA Proteins and Activation of ABRE-Dependent Signal Transduction Pathways

The expressions of RD29B and RAB18 in the 35S-AREB1QT plants without exogenous ABA were weak compared with the expressions in the wild-type plants in response to ABA treatment, whereas the mRNA accumulation levels of HIS1-3, AIA1, and GBF3 in the 35S-AREB1QT plants without exogenous ABA were higher than those in wild-type plants treated with exogenous ABA (Figure 3D). Expressions of AIL1, RD20, and KIN2 were similar between the 35S-AREB1QT plants without exogenous ABA and the wild-type plants with ABA (Figure 3D). The difference in these expression patterns between the 35S-AREB1QT plants without exogenous ABA and wild-type plants with exogenous ABA may reflect some difference in the composition of the transcriptional regulatory complex of the cis-element in the promoter region of the target genes.

Transgenic Plants Overexpressing AREB1QT Are Hypersensitive to ABA
To evaluate the effect of AREB1QT overexpression in transgenic plants on ABA sensitivity, we germinated 35S-AREB1QT seeds and grew the seedlings on growth medium (GM) containing various concentrations of ABA. During the germination process, no obvious difference was observed between the 35S-AREB1QT and wild-type plants (data not shown). However, the 35S-AREB1QT seedling growth, including root growth, and cotyledon greening and expansion were severely inhibited when the ABA concentration was 0.5 µM or higher (Figures 4A and 4B). By contrast, seeds of the wild-type plants germinated and the seedlings grew normally, although at a slower rate than those on ABA-free GM (Figures 4A and 4B). In addition, we found no differences in germination or seedling growth between the 35S-AREB1 and wild-type plants by 6 d after stratification, whereas at 2 weeks after stratification, the seedling growth of 35S-AREB1 plants was severely inhibited when the ABA concentration was 0.5 µM or higher compared with the wild-type plants (data not shown).

View larger version (44K): [in this window] [in a new window]   Figure 4. 35S-AREB1QT Plants Are Hypersensitive to ABA. (A) Growth of 35S-AREB1QT (line 5) and 35S-AREB1 (line 6) plants on GM agar plates containing 0, 0.5, or 1.0 µM ABA. Seeds were germinated and grown on GM agar plates for 6 d; representative plants are shown. (B) ABA dose response of root growth. Seeds were germinated and grown on GM agar plates containing various concentrations of ABA and 1% sucrose. Root elongation was measured 6 d after stratification, and relative growth compared with that on ABA-free medium is indicated. Bars indicate standard deviation; n = 26 to 38. The experiments were performed three or more times, sometimes using different transgenic lines, and the results were consistent.

Transgenic Plants Overexpressing AREB1QT Display Enhanced Drought Tolerance

AREB1QT

View larger version (74K): [in this window] [in a new window]   Figure 5. Enhanced Drought Tolerance in 35S-AREB1QT Plants. (A) Enhanced tolerance to drought in the 35S-AREB1QT plants (lines 12 and 26). Watering was withheld from 3-week-old plants for 12 d, then rewatering for 10 d, before the photograph was taken. Number codes indicate number of surviving plants out of total number. (B) Enhanced ability of 35S-AREB1QT plants (line 12) to survive the dehydration condition. Three-week-old transgenic and wild-type plants were grown on GM agar plates, transferred to Petri dishes, left unwatered for 6 h, and then rewatered. (C) Increased survival rates of the 35S-AREB1QT plants (lines 12 and 26) under dehydration. Water was withheld for 5 to 6 h from 3-week-old plants and then survival rates were counted. Surviving plants were scored on the second day. Survival rates and standard deviations (bars) were calculated from results of three independent experiments. (D) Water loss rates of 35S-AREB1QT (lines 12 and 26) plants. Each data point represents the mean of duplicate measurements (n = 7 each). Error bars represent standard deviation. (E) Standardized water content of 35S-AREB1QT (lines 12 and 26) plants. Details as in (D). Some error bars are smaller than the symbols. (F) Stomatal aperture of 35S-AREB1QT plants (line 12). Stomatal guard cells were observed in the middle of the watering cycle.

Loss-of-Function Mutants of AREB1 Display Opposite Phenotypes to That Conferred by Overexpression of AREB1QT
The constitutive expression of AREB1QT in transgenic plants resulted in significant changes in ABA-associated phenotypes, such as ABA sensitivity and drought tolerance, and altered expression of ABA/stress-responsive genes, suggesting that AREB1 functions in the ABA-mediated stress signaling pathway. The overexpression of the active form of AREB1, however, might have caused unnatural conditions; so, the overexpression phenotypes need to be interpreted with caution. For example, a high level of the active form of AREB1 may result in altered specificity of interaction with the target proteins, leading to an altered function in the cell. In addition, overexpression of the active form of AREB1 also may have other, nonspecific effects on gene expression. However, the fact that overexpression of the active form of AREB1 conferred a positive effect on ABA-mediated drought stress response suggests that AREB1 function is specific even after overexpression. Nevertheless, we cannot exclude the possibility that a gain of function occurred as a result of the overproduction of the active form of AREB1.

To further investigate the function of the AREB1 gene, we assessed two types of loss-of-function mutants: a T-DNA insertion mutant of AREB1, areb1, and transgenic plants overexpressing AREB1 fused to a fragment of the EAR motif RD, SRDX, under the control of the CaMV 35S promoter, 35S-AREB1:RD (Figures 6A to 6D). The mutant areb1 has a T-DNA insertion in the first intron of AREB1 (Figure 6A). RT-PCR analysis indicated that the knockout mutant plants did not produce AREB1 transcripts even after the application of exogenous ABA or during dehydration, while AREB1 and a larger mRNA band were observed in the wild-type plants (Figure 6C). This larger band seems to correspond to the estimated size of the unspliced form of AREB1 mRNA and was also detected using the other AREB1-specific primer pairs, suggesting that this may indeed be the unspliced form of AREB1 mRNA (Figure 6C; data not shown). By contrast, in the areb1 knockout mutant, a truncated form of AREB1 mRNA was detected using primers carrying AREB1- and T-DNA–specific binding sequences (Figure 6C). Growth of wild-type seedlings was inhibited gradually with increasing amounts of ABA, and growth of areb1 seedlings was also inhibited, but to a lower degree, especially at 3.0 µM ABA on the GM agar plates 2 weeks after stratification (Figure 6E). Furthermore, for more detailed analysis, because it is difficult to get intact roots from GM agar plates >2 weeks after stratification, we used GM plates containing 2.5% Gelrite (Merck), which is soft enough to release them even later than 2 weeks after stratification. Increased concentrations of ABA also resulted in greater growth retardation of both wild-type and areb1 primary roots, although the effect of ABA on primary root growth in the Gelrite plates seems to have been severer than in the agar plates (Figures 6E and 6F). At 1.0 µM ABA, primary root growth of wild-type plants was 7% of that in the medium without ABA, while that of the areb1 plants was 61% of that in the medium without ABA (Figure 6F). Thus, these data demonstrate insensitivity of the areb1 plants to ABA (Figures 6E and 6F) despite no obvious difference in ABA sensitivity having been observed between areb1 and wild-type plants in the germination process (data not shown). At 3 weeks after stratification, the maximum rosette radius of areb1 plants averaged 23% larger than that of the wild-type plants (Figures 6G and 6H). Thus, before the bolting stage, the areb1 plants are slightly larger than wild-type plants, but after that, the size difference gradually decreased, and then eventually the sizes were indistinguishable (data not shown), indicating that the areb1 plants grew slightly faster than the wild-type plants during the vegetative phase.

View larger version (71K): [in this window] [in a new window]   Figure 6. Analysis of AREB1 Loss-of-Function Mutants. (A) Scheme of the Arabidopsis AREB1 gene. Exons (open boxes) and introns (lines) are indicated. The position of the T-DNA insertion is shown (not to scale). (B) Schematic representation of the 35S-AREB1:RD construct used for expression of the chimeric repressor with a modified version of the EAR-motif RD (SRDX), consisting of 12 peptides. (C) Expression levels of AREB1 in the areb1 knockout mutant were determined by RT-PCR using total RNAs isolated from 2-week-old plants with or without 6-h treatment of 100 µM ABA or dehydration and grown on GM agar plates. Primers for detection of a truncated form of AREB1 mRNA, generated by T-DNA insertion into the first intron of AREB1, have T-DNA– and AREB1-specific binding sequences. The arrow and asterisks indicate expression of AREB1 and a larger band, respectively. TUB1, ß-1 tubulin transcript as a control. (D) RNA gel blot analysis of AREB1 mRNA in wild-type and 35S-AREB1:RD plants (lines 4 and 8) in the absence or presence of 50 µM ABA for 7 h. Eight micrograms of total RNA from 3-week-old seedlings was probed with AREB1 cDNA. (E) Growth of the mutant areb1 and 35S-AREB1:RD plants (line 4) on GM agar plates containing 0, 1.0, or 3.0 µM ABA, supplemented with 1% sucrose. Seeds were germinated and grown on the medium for 2 weeks. Bars = 25 mm. (F) ABA dose response of primary root growth. Seeds were geminated and grown on GM plates containing 0.25% Gelrite, 1% sucrose, and various concentrations of ABA. Primary root elongation was measured 19 d after stratification, and relative growth compared with that on ABA-free medium is indicated. Bars indicate standard deviation; n = 15 to 31. (G) Growth phenotypes of areb1 and 35S-AREB1:RD (line 4) plants that were grown for 3 weeks on GM agar plates supplemented with 1% sucrose. (H) Maximum rosette radius (i.e., length of the longest rosette leaf) of each plant on a GM agar plate containing 3% sucrose was measured 3 weeks after stratification. Three independent lines of wild-type plants, one line of the areb1 T-DNA insertion mutant, and seven independent lines of 35S-AREB1:RD plants were used. Bars indicate standard deviation; n = 7. (I) The fusion of the RD to AREB1 creates a repressor. Arabidopsis protoplasts were cotransfected with the RD29B-GUS reporter and the effector construct expressing AREB1 or AREB1:RD, or vector DNA. The RD29B-GUS reporter plasmid and the transient assay system are described in the legend of Figure 2. (J) Expression profile of downstream genes identified by microarray analysis (Table 1) in 35S-AREB1:RD plants (line 4). Two-week-old seedlings were either not treated (–) or treated (+) with ABA for 7 h. Each lane contained 7 µg of total RNA. Three to eight independent lines were used; results from one representative experiment are shown. (K) Difference in recovery after rehydration among 35S-AREB1QT (line 12), 35S-AREB1:RD (line 4), and wild-type plants. Transgenic and wild-type plants were grown on GM agar plates for 2 weeks, transferred to Petri dishes, left unwatered for 4 h, and then rewatered. The photograph was taken 2 d after rewatering. (L) Quantification of the survival rates of the wild-type and 35S-AREB1:RD plants (lines 4 and 8) after rehydration. Water was withheld for 5 h from 3-week-old plants and then survival rates were counted. Surviving plants were scored on the second day. Survival rates and standard deviations were calculated from the results of four independent experiments (n = 10 each). Bars indicate standard deviations.

To minimize the effects of phenotypic masking due to functional redundancy, we used this technology to generate 26 35S-AREB1:RD transgenic lines overexpressing AREB1 fused to the SRDX RD under the control of the CaMV 35S promoter. Expression levels of the transgene in the 26 independent transgenic lines were analyzed by RNA gel blot analysis using an AREB1-specific probe; we selected eight transgenic lines with higher AREB1:RD expression levels for phenotypic analysis. Because the eight transgenic lines behaved in a similar manner (Figures 6D to 6H; data not shown), we performed detailed analysis on representative lines 4 and 8. RNA gel blot analysis of the 35S-AREB1:RD plants showed that in the 35S-AREB1QT plants under unstressed condition, three of the eight candidate target genes, HIS1-3, AIA1, and RD29B, were downregulated significantly in the presence of ABA (Figure 6J). By contrast, expression of another four candidate target genes, AIL1, RD20, RAB18, and KIN2, was not altered, and only GBF3 expression seemed to be only slightly upregulated (Figure 6J). Thus, although ABA-dependent secondary gene expression could be induced, approximately half of the genes downstream of AREB1 were suppressed even in the presence of exogenous ABA, indicating that the 35S-AREB1:RD plants act as loss-of-function mutants. These data also suggest that the three genes that are upregulated by the overexpression of AREB1QT and downregulated in the loss-of-function mutant of AREB1 with the RD (HIS1-3, AIA1, and RD29B) are target genes regulated mainly by AREB1. Since upregulated genes such as RD20, RAB18, and KIN2 are well-known stress-inducible markers (Seki et al., 2002), we expect the expression of these genes to be also mediated by transcription factors other than AREB1.

At 3 weeks after stratification, the maximum rosette radius of 35S-AREB1:RD plants averaged 56% larger than that of wild-type plants (Figures 6G and 6H). Compared with the wild-type plants, petioles of the 35S-AREB1:RD plants were longer and thicker. This enhanced growth phenotype contrasted with the growth retardation phenotype of the 35S-AREB1QT plants during the vegetative phase. By flowering time, however, the growth phenotype of the 35S-AREB1:RD plants had become similar to that of the wild-type plants (data not shown). We also tested the ABA sensitivity of the 35S-AREB1:RD transgenic plants. The 35S-AREB1:RD plants were more resistant to ABA than the other loss-of-function mutant areb1 plants or the wild-type plants >2 weeks after stratification (Figures 6E and 6F). In the germination process, however, no obvious ABA insensitivity was observed in the 35S-AREB1:RD plants or in the areb1 plants (data not shown).

As described above, the 35S-AREB1:RD plants were insensitive to ABA, and at least three stress-inducible genes were downregulated in them (Figures 6E, 6F, and 6J). Therefore, compared with wild-type plants, the 35S-AREB1:RD plants could be expected to survive less well under dehydration. To test this, we examined the differences in recovery after dehydration using plants grown on GM agar plates. All 35S-AREB1:RD plants were dead, and many wild-type plants were partially dead, but most 35S-AREB1QT plants survived when they were rewatered afterwards (Figure 6K). To clarify this difference between the 35S-AREB1:RD and wild-type plants, we performed further tests. Compared with the wild-type plants, the survival rates of the 35S-AREB1:RD plants were reduced, confirming that the 35S-AREB1:RD plants survived less well under dehydration (Figure 6L). These phenotypes of the 35S-AREB1:RD plants also support the notion that AREB1 plays a pivotal role in the ABA-mediated stress signaling pathway in vegetative tissues. However, because such ABA-dependent phenotypes were due to overexpression of the gene, we need to interpret them with the cautions described above. Nevertheless, when we consider the two types of loss-of-function mutants, these loss-of-function phenotypes in growth, ABA sensitivity, or survival under dehydration were the opposite of the gain-of-function phenotypes conferred by the overexpression of AREB1QT, demonstrating that AREB1 is involved in the ABA-mediated tolerance to drought through regulation of the ABA-dependent expression of novel downstream genes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Here, we show that expression of the intact AREB1 gene alone is insufficient to upregulate its downstream genes under normal growth conditions. Taken together with data from our previous in-gel kinase assays and protoplast transient assays (Uno et al., 2000), the data may support the notion that not only ABA-induced transcription but also ABA-induced modification of AREB1 is required to activate expression of ABRE-dependent downstream genes. On the basis of recent reports that AREB1 and its homologs are phosphorylated in vitro or in vivo (Uno et al., 2000; Lopez-Molina et al., 2001; Kagaya et al., 2002), phosphorylation of AREB1 may be involved in the modification. We have shown that AREB1 is expressed constitutively in specific tissues (roots, hydathodes, and some vascular systems) and is induced by drought and high salt in all vegetative tissues. Therefore, activation of AREB1 by such modification without de novo protein synthesis allows ABA-dependent gene expression to respond more rapidly to stress conditions in these specific tissues. Since overexpression of intact AREB2 or ABF3, unlike AREB1, affects ABA sensitivity and stress tolerance in normal medium containing 1% sucrose (Kang et al., 2002; Kim et al., 2004; data not shown), this two-step signal perception system containing transcriptional and posttranscriptional regulation is more important in AREB1 than that in the other two homologs. Furthermore, the addition of ABA or drought stress dramatically activated AREB1 promoter activity in all tissues (Figure 1E), whereas very little induction of AREB2 or ABF3 promoter activity was observed under stress conditions in histochemical analyses (Uno et al., 2000; Kang et al., 2002). Our results show that AREB1, rather than the other bZIP homologs, plays an important role in vegetative tissues under drought stress conditions. In this scenario, the recent finding that ABF2 (AREB1) is a positive component of glucose signaling implies the possibility that AREB1-mediated stress response is involved in glucose signal transduction (Kim et al., 2004).

By differential expression analyses with microarray and RNA gel blot analyses, in combination with information on stress inducibility and cis-elements in the promoter region, we identified two groups of candidate target genes of AREB1: (1) four LEA class genes (AIL1, RD29B, KIN2, and RAB18) and (2) four regulatory genes (HIS1-3, AIA1, GBF3, and RD20). Interestingly, promoter regions of all these genes carry two or more ABRE motifs. This is consistent with recent findings that two ABRE motifs are required for activation of gene expression by AREB1 (Choi et al., 2000; Uno et al., 2000) and suggests that these genes are candidates for direct targets of AREB1. In particular, three of the eight genes (HIS1-3, AIA1, and RD29B) were upregulated in 35S-AREB1QT plants even in the absence of exogenous ABA and downregulated in 35S-AREB1:RD plants even in the presence of exogenous ABA, suggesting that these genes are directly and mainly regulated by AREB1. The remaining five upregulated genes that were not downregulated in the 35S-AREB1:RD plants even in the presence of exogenous ABA also could be regulated by cis-elements other than the ABRE sequence and dedicated transcription factors other than AREB1 because at least three genes (KIN2, RAB18, and RD20) are also induced by cold stress and are known to be multiple stress marker genes (Table 1). For example, RD20 is upregulated by overexpression of RD26, which is a transcription factor with a NAC domain, which is induced by dehydration, high salinity, or ABA (Fujita et al., 2004).

Among the four regulatory genes identified as candidate target genes of AREB1, we expect two genes, HIS1-3 and AIA1, to be direct target genes of AREB1. HIS1-3, encoding a linker histone H1-3 protein, showed the highest increase of expression in the 35S-AREB1QT plants under unstressed conditions (Table 1), and HIS1-3 expression was suppressed in the 35S-AREB1:RD plants in the presence of exogenous ABA (Figure 6J). Linker histone H1, unlike core histones (H2A, H2B, H3, and H4), is the most variable histone in eukaryotes and regulates specific gene expression, but not global transcription, throughout all tissues (Shen and Gorovsky, 1996; for review, see Jerzmanowski et al., 2000). HIS1-3 (Ascenzi and Gantt, 1997) in Arabidopsis and H1-D (Wei and O'Connell, 1996) and H1-S (Scippa et al., 2000) in tomato (Lycopersicon esculentum) have been reported as drought-inducible variants of histone H1 genes. These findings suggest that HIS1-3 plays an important role in drought-responsive gene expression mediated by chromatin remodeling. Previous results of the expression of HIS1-3 in RNA gel blot and histochemical analyses showed that the stress inducibility and location of expression are very similar to those of AREB1 (Ascenzi and Gantt, 1997, 1999). In particular, HIS1-3 expression is characteristically observed around the transition zone (the area at the junction of root and hypocotyl) and in hydathodes (Ascenzi and Gantt, 1999). AREB1 expression was also characteristically detected in the same region (Figure 1E), but expression patterns of AREB2 or ABF3, two members of the bZIP family, were distinct from that of HIS1-3 (Ascenzi and Gantt, 1999; Kang et al., 2002). Also, interestingly, expressions of AREB1, HIS1-3, RD29B, and RAB18 were not induced by ABA or dehydration treatments in the abi1 mutant (Ascenzi and Gantt, 1997; Uno et al., 2000; Y. Uno and K. Yamaguchi-Shinozaki, unpublished data), implying that the ABA-dependent expression via AREB1 is mediated by the ABI1 protein.

AIA1 encodes an AAA ATPase with chaperone-like activity (Neuwald et al., 1999). AAA ATPases form a large protein family with manifold cellular activities, including proteolysis, protein folding, and cytoskeletal regulation (Vale, 2000). In many cases, AAA domains assemble into hexameric rings that are likely to change their shape during the ATPase cycle (Vale, 2000). Although recent reports have shown that AAA ATPase is involved in multiple cellular functions via chaperone-like activity, the role of AAA ATPase in plants is still unknown. Our findings suggest that AIA1 is involved in drought response via its chaperone-like activity. Since all these candidate target genes were ABA and stress inducible, it is likely that the genes downstream of AREB1 enhance drought stress tolerance and increase fundamental cellular activities under stress conditions (Figure 7). Further dissection of such target genes will give us new insights into the ABA signal transduction network under stress conditions.

View larger version (31K): [in this window] [in a new window]   Figure 7. A Model of the Regulation of ABA Signaling by AREB1. AREB1 is postulated to mainly regulate the expression of stress-responsive genes via ABRE sequences in vegetative tissues.

All tested mutants of AREB1 showed no obvious phenotypes in the germination process compared with the wild-type plants (data not shown). Also, histochemical analysis detected very little AREB1 promoter activity in newly germinated seedlings (Figure 1E). These data suggest that AREB1 does not function during germination. This is also consistent with the finding that AREB1, unlike its bZIP homolog ABI5, has not been isolated to date in extensive genetic screening during germination (Leung and Giraudat, 1998; Finkelstein and Lynch, 2000; Lopez-Molina and Chua, 2000). By contrast, overexpression of AREB2 or ABF3 caused ABA hypersensitivity to some extent during germination, indicating that the role of AREB1 is distinct from those of AREB2 and ABF3 (Kang et al., 2002; data not shown). This finding may be related to the observation that expression of AREB1, unlike AREB2 or ABF3, was inhibited by overexpression of VP1, a key determinant of seed-specific gene expression, in the vegetative tissues (Suzuki et al., 2003). Thus, AREB1 appears to play a specific role only in the vegetative phase. Moreover, interestingly, in the AREB1- or AREB1QT-overexpressing plants, the mRNA decreased over time after ABA treatment of the plants (Figures 1F, 3D, and 6J). Although this phenomenon may be specific to the AREB1 sequence and may play some role in the regulation of AREB1, the role and mechanism are still unknown.

Our results show that the overexpression of the AREB1 mutation carrying an internal deletion between the N-terminal activation and bZIP DNA binding domains enables the constitutive activation of transcription of downstream genes in the absence of ABA. This suggests that the region of the internal deletion functions as a regulatory domain in response to ABA. Because it appears that AREB1 mutations lacking the N-terminal activation domain have a dominant negative effect on binding of endogenous transcription factors to the ABRE motifs in the promoter region of the reporter plasmid (Figures 2B and 2E), the altered function of the activation domain, rather than the change in binding activity of AREB1 to ABRE motifs, seems to help overcome the problem in some modifications. Hence, conformational change generated by the internal deletion may be associated with exposure of an activation domain masked in the normal folded complex (Figure 7). Furthermore, in the protoplast transient assay, an effector plasmid producing fusion proteins consisting of the DNA binding domain of GAL4 and the P region of AREB1 (which has a single amino acid substitution of Ser-26 to Ala in the R-x-x-S/T phosphorylation target site in the N-terminal conserved region of AREB1) did not decrease expression of the GUS reporter gene, suggesting that phosphorylation of the target site in the N-terminal conserved domain in the P region does not affect transactivation activity itself (data not shown). Taken together, these results indicate that ABA-induced modification, such as phosphorylation, might contribute to enabling the function of the N-terminal transactivation domain via conformational change of AREB1 rather than to activating the N-terminal transactivation domain itself.

Considering the potential functional redundancy of AREB family proteins and many bZIP factors interacting with ABREs, traditional loss-of-function approaches would not be appropriate for studying such bZIP factors (Foster et al., 1994; Kang et al., 2002). Although an areb1 T-DNA insertion mutant exhibited a milder phenotype, we could not determine whether this was an AREB1-specific phenotype or if it was compensated functionally by the other redundant bZIPs. To minimize the effects of phenotypic masking due to functional redundancy, we used the recently established chimeric repressor silencing technology (Hiratsu et al., 2003) in loss-of-function analysis of AREB1. Using several well-studied transcription factors, such as CUC1 and EIN3, they demonstrated that the chimeric repressors, which express transcription factors with an RD, SRDX, exhibited dominant loss-of-function phenotypes even in the presence of functionally redundant transcription factors (Hiratsu et al., 2003). In the case of AREB1, we show here that the phenotype conferred by the overexpression of the AREB1 mutant with the RD is opposite to that rendered by the overexpression of the constitutive active form of AREB1. Gain-of-function phenotypes in the AREB1 homolog are very different from each other (Kang et al., 2002), and loss-of-function phenotypes of the AREB1 mutant with the RD are also different from those of the homologs AREB2 and ABF3 (Y. Fujita and K. Yamaguchi-Shinozaki, unpublished data). These data suggest that phenotypes conferred by either type of mutant (with an internal deletion or the RD) are AREB1 specific. Thus, in the study of important transcription factors that have redundant homologs and are modulated by several means, the creation of constitutive active and chimeric repression forms of transcription factors are attractive approaches for revealing the mechanism of such transcription factors.

Here, we show that expression of the intact AREB1 gene alone is insufficient to lead to expression of the downstream genes under normal growth conditions. Furthermore, we identified the N-terminal region of AREB1 as a transcriptional activation domain and then created a constitutive active form of AREB1 carrying the N-terminal activation and bZIP DNA binding domains. Using the constitutive active form of AREB1, a T-DNA insertion knockout mutant, and a dominant loss-of-function mutant with the SRDX RD, we further demonstrate that AREB1 plays a key role in vegetative tissues under drought stress and mediates novel ABRE-dependent ABA signaling that enhances drought stress tolerance. Overall, these findings may contribute to our understanding of several important mechanisms underlying AREB1 functions in a novel ABRE-dependent ABA signal transduction pathway in vegetative tissues under drought stress.

	METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Plant Materials and Growth Conditions
Plants (Arabidopsis thaliana ecotype Columbia) were grown on GM agar plates for 2 to 3 weeks as described (Osakabe et al., 2005) under a 16-h-light/8-h-dark regime (40 ± 10 µmol photons/m²/s). The GM agar plates were supplemented with 1 or 3% sucrose and, as described in Results, with ABA as needed. T87 cultured cells, derived from Arabidopsis, were maintained, grown, and treated as described (Satoh et al., 2004). An Arabidopsis AREB1 T-DNA insertion line (SALK_002984; Col-0 ecotype) was obtained from the Arabidopsis Biological Resource Center (Columbus, OH). Insertion mutant information was obtained from the Salk