Although proteins in the basic helix-loop-helix (bHLH) family are universal transcription factors in eukaryotes, the biological roles of most bHLH family members are not well understood in plants.
The Arabidopsis thaliana bHLH122 transcripts were strongly induced by drought, NaCl and osmotic stresses, but not by ABA treatment. Promoter::GUS analysis showed that bHLH122 was highly expressed in vascular tissues and guard cells. Compared with wild-type (WT) plants, transgenic plants overexpressing bHLH122 displayed greater resistance to drought, NaCl and osmotic stresses. In contrast, the bhlh122 loss-of-function mutant was more sensitive to NaCl and osmotic stresses than were WT plants.
Microarray analysis indicated that bHLH122 was important for the expression of a number of abiotic stress-responsive genes. In electrophoretic mobility shift assay and chromatin immunoprecipitation assays, bHLH122 could bind directly to the G-box/E-box cis-elements in the CYP707A3 promoter, and repress its expression. Further, up-regulation of bHLH122 substantially increased cellular ABA levels.
These results suggest that bHLH122 functions as a positive regulator of drought, NaCl and osmotic signaling.
Drought and salt stresses are the major environmental factors limiting crop productivity worldwide. To reduce the adverse effects of abiotic stresses, plants have evolved multifaceted strategies, including morphological, physiological and biochemical adaptations (Xiong et al., 2002; Zhu, 2002; Shinozaki et al., 2003; Bohnert et al., 2006). The cellular and molecular mechanisms underlying plant adaptation to environmental stresses have been researched intensively, and changes in gene expression play an important role in this process. The stress-responsive genes can be divided into two subgroups based on their putative functions (Wei et al., 2009). The first group includes enzymes required for osmoprotectants, late embryogenesis abundant (LEA) proteins, aquaporin proteins, chaperones and detoxification enzymes that protect cell membrane integrity, control ion balances and scavenge reactive oxygen species. The second group is represented by protein kinases and transcription factors involved in signal perception, signal transduction and transcriptional regulation of gene expression.
Members of the basic helix-loop-helix (bHLH) family, which are designated by the bHLH signature domain, are universal transcription factors in eukaryotes. The bHLH domain usually consists of c. 60 amino acids with two functionally distinct regions: the basic region, which contains 13–17 primarily basic amino acids for DNA binding, and the HLH region, which enables the formation of homodimers or heterodimers with one or several different partners (Toledo-Ortiz et al., 2003; Feller et al., 2011). In animals, bHLHs are involved in essential physiological and developmental processes, such as neurogenesis, myogenesis and cell differentiation (Ma et al., 1997; Tomita et al., 2000; Nieto et al., 2001). In Arabidopsis, bHLHs are the second largest kind of transcription factor. Although > 160 genes have been predicted to encode bHLHs, functions have been determined or partially determined for only c. 30% (Heim et al., 2003; Li et al., 2006).
Limited research has demonstrated that plant bHLHs are also involved in various developmental processes, including asymmetric cell division and differentiation of stomata (MacAlister et al., 2007), nodule vascular patterning (Godiard et al., 2011), floral organogenesis (Heisler et al., 2001; Reymond et al., 2012) and axillary meristem formation (Yang et al., 2012). Recently, increasing evidence has shown that bHLHs can regulate plant responses to abiotic stresses. bHLH29/FIT1 (Fer-like iron deficiency-induced transcription) interacts with bHLH38 and bHLH39 in regulating iron-uptake gene expression for iron homeostasis in Arabidopsis, and transgenic plants overexpressing FIT/bHLH38 or FIT/bHLH39 accumulate more iron in their shoots than do wild-type (WT) plants (Yuan et al., 2008). PIF4 (Phytochrome-interacting factor 4), encoding a nuclear-localized bHLH protein, interacts directly with BZR1 (Brassinazole-resistant 1) and forms a module that integrates steroid and environmental signals (Oh et al., 2012). JAM1 (ABA-inducible bHLH transcription factor/jasmonic acid (JA)-associated MYC2-like1), acting as a repressor to negatively regulate jasmonate signaling, plays a pivotal role in the fine-tuning of JA-mediated stress responses and plant growth (Nakata et al., 2013); In 35S::ICE1 (a MYC-like bHLH transcriptional activator) transgenic plants, the expression of the CBF regulon is enhanced under low temperatures, and this increases freezing tolerance in Arabidopsis (Chinnusamy et al., 2003). However, the biological roles of most bHLH family members in plants are still poorly understood.
Here, we show that the expression of bHLH122, which is an activator in the CONSTANS/FLOWERING LOCUS T (CO/FT) photoperiodic flowering pathway in Arabidopsis (Ito et al., 2012), is strongly induced by drought, NaCl and osmotic stresses, but not by ABA treatment. Our data indicate that bHLH122 functions as a positive regulator of drought, NaCl and osmotic signaling. We also demonstrate that CYP707A3 is a target gene of bHLH122. Up-regulation of bHLH122 represses CYP707A3 transcripts and substantially increases cellular ABA levels.
Materials and Methods
Plant materials and growth conditions
The Arabidopsis thaliana (L.) Heynh. ecotype Colombia (Col-0) was used as the WT and was the genetic background for transgenic plants. The T-DNA insertion mutants of bHLH122 (SALK_049022C), ABA2-1 (CS156) and ABI4-1 (CS8104) were obtained from the Arabidopsis Biological Resource Center (ABRC), Columbus, OH, USA. Seedlings in Murashige and Skoog (MS) nutrient agar medium were grown under continuous light (70 μmol m−2 s−1) at 23 ± 1°C. The MS medium was supplemented with 3% (w/v) sucrose and, unless described in the Results, with mannitol, NaCl or ABA as needed. Soil-grown Arabidopsis plants were kept in a growth chamber with a photoperiod of 16 h : 8 h light : day at 23 ± 1°C. For drought treatment, plants were grown in soil with sufficient water for 3 wk, and then the water was withheld for the durations indicated. For mannitol, NaCl and ABA treatments, 2-wk-old seedlings were pulled out of the agar medium and submerged in a solution containing 300 mM mannitol, 150 mM NaCl or 100 μM ΑΒΑ for the durations indicated.
Constructs and generation of transgenic plants
Full-length cDNA of bHLH122 was amplified with the following primers: forward, 5′-CAC CAT GGA ATC AGA ATT CCA GCA-3′; reverse, 5′-AGT TTC TAA CAA AAG AAA ATA AAC T-5′. The amplified fragments were introduced into the pENTR™/D-TOPO vector (Invitrogen) and cloned into pMDC32 or pEarleyGate202 by LR reactions (Invitrogen). For the bHLH122 promoter::β-glucuronidase (GUS) construct, a 2.0-kb fragment upstream from the initiation codon was amplified with the forward primer 5′-CAC CCT TTG TTC CAA AGT AAC GAG-3′ and reverse primer 5′-CTT AGT AGC TTT TGA GTT TCT CTC T-3′, and cloned into the pMDC164 vector following Gateway recombination.
The plasmid was electroporated into Agrobacterium tumefaciens GV3101 and transformed into Col-0 by the floral dip method (Clough & Bent, 1998). Transgenic plants were selected with the use of 35 μg ml−1 hygromycin. T3 or T4 homozygous lines were used for all experiments.
Localization of bHLH122
The full-length bHLH122 coding region was amplified with the primers containing XhoI and BamHI sites (forward primer, 5′-CCG CTCGAG ATG GAA TCA GAA TTC CAG CA-3′; reverse primer, 5′-CGC GGATCC TGC GCA CTA GAG CAT CTA CAT C-5′). The PCR products were cleaved with XhoI/BamHI (the restriction sites are shown in italic in the indicated sequences) and subcloned into pEZS-NL to generate the bHLH122 expression plasmid. The plasmid (1.5 μg) was transformed into onion epidermal cells by the Biolistic PDS-1000/He Particle Delivery System (Bio-Rad, USA), and the bHLH122-green fluorescent protein (GFP) fusion protein in transformed onion cells was microscopically detected with a Nikon Eclipse TE2000-E confocal microscope (Nikon, Tokyo, Japan). Images were analyzed using EZ-C1 software (Nikon).
Total RNA was extracted from WT, mutants and transgenic plants with Trizol reagent (Invitrogen). First-strand cDNA was synthesized using SuperScript™ III First-Strand Synthesis Supermix (Invitrogen). Quantitative real-time reverse transcription-polymerase chain reaction (RT-PCR) was carried out in an ABI 7500 system using the SYBR Premix Ex Taq™ (perfect real time) kit (TaKaRa Biomedicals, Dalian, China). PCR included a preincubation at 95°C for 3 min, followed by 40 cycles of denaturation at 95°C for 10 s, annealing at 55°C for 15 s and extension at 72°C for 45 s. The PCR products were loaded onto 1.5% agarose gels and photographed after staining with ethidium bromide. The primer pairs used for real-time RT-PCR are listed in Supporting Information Table S1. Primer efficiencies were measured and calculated (Ramakers et al., 2003). The average amplification efficiency was > 89%, and could be used to reliably detect the expression changes. The relative expression level was calculated using the comparative Ct method. Results were normalized to the expression of Tub4. Each experiment was replicated at least three times.
Electrophoretic mobility shift assay (EMSA)
Affinity-purified glutathione S-transferase (GST)-bHLH122 recombinant protein was used for EMSA. For the generation of GST-bHLH122 recombinant protein, a PCR was performed to amplify the entire coding region of bHLH122 with the following primers: forward, 5′-CG GGATCC ATG GAA TCA GAA TTC CAG CA-3′; reverse, 5′-ATA GTT TA GCGGCCGC CGC ACT AGA GCA TCT ACA TC-3′. The primers contained the restriction sites BamHI and NotI (the restriction sites are shown in italic in the indicated sequences). The amplified product was gel purified and subcloned into the BamHI/NotI-digested pGEX-4T vector. The resulting construct was verified by sequencing. GST-bHLH122 was expressed in Escherichia coli BL21 and purified by glutathione-Sepharose™ 4B (GE Healthcare, Sweden) according to the manufacturer's protocol. Purified protein was concentrated with Centricon YM-3 filter units (Millipore).
EMSA was performed with the DIG Gel Shift Kit (2nd Generation; Roche). In brief, DNA probes were amplified by PCR with the primers listed in Table S1 and labeled with digoxygenin-11-ddUTP using the recombinant terminal transferase (Roche). The binding reaction was carried out in a total volume of 20 μl containing different quantities of purified recombinant protein, 0.096 pmol probe, 20 mM Hepes (pH 7.6), 1 mM EDTA, 10 mM (NH4)2SO4, 1 mM dithiothreitol (DTT), 0.2% (w/v) Tween 20, 30 mM KCl, 1 μg of poly-(dI–dC) and 0.1 μg of poly-l-lysine. The mixture was incubated at 25°C for 30 min and separated on a 5.0% native polyacrylamide gel. Then, the DNA was transferred to Hybond-N plus membranes (Amersham). The signal was detected by chemiluminescence and recorded on X-ray film (Kodak).
Chromatin immunoprecipitation (ChIP) assays
ChIP was performed as described previously (Tsugama et al., 2012). The leaves of 4-wk-old FLAG or FLAG-bHLH122 overexpression lines were fixed in 1% formaldehyde for 15 min under a vacuum and neutralized with 125 mM glycine (Gly) under a vacuum for an additional 5 min. After homogenization, the chromatin was isolated and sonicated. Antibody against FLAG was used. The enriched DNA fragments were determined by RT-RCR using the primers listed in Table S1. Three independent biological replicates were performed for ChIP assay.
Stomatal aperture analysis
Stomatal apertures were measured as described previously (Pei et al., 1997). Leaves of similar size and age were sampled from WT and 35S::bHLH122 plants that had been subjected to drought for 10 d. Rosette leaves were floated in solutions containing 30 mM KCl, 10 mM Mes-Tris, pH 6.15, and exposed to 150 μmol m−2 s−1 light for 3 h. A light microscope (Olympus ix71, Tokyo, Japan) was used to examine the stomata on epidermal strips obtained from rosette leaves. The width and length of stomatal pores, as determined by the software Image J (http://rsbweb.nih.gov/ij), were used to calculate stomatal apertures (ratio of width to length).
Anthocyanin content measurement
Anthocyanin contents were measured as described previously (Rabino & Mancinelli, 1986). The pigments were extracted with 99 : 1 methanol : HCl (v/v) at 4°C, and the optical density at 530 nm (OD530) and OD657 for each sample were measured. OD530 – 0.25 × OD657 was used to compensate for the contribution of chlorophyll and its products to the absorption at 530 nm.
Water loss measurement
For water loss measurement, eight leaves per individual mutant and WT plant growing under normal conditions for 3 wk were excised, kept on the laboratory bench at 22°C and weighed at the designated time intervals. Four replicates were performed for each line. Water loss was represented as the percentage of initial fresh weight at each time point.
The leaves of 4-wk-old soil-grown WT and 35S::bHLH122 plants with sufficient water and drought treatment for 10 d were used for ABA measurement. Approximately 200 mg (fresh weight) of each sample were finely ground in liquid nitrogen for extraction with 1 ml methanol containing 20% water (v/v) at 4°C for 24 h. Purification was performed with an Oasis Max solid phase extraction cartridge (150 mg/6 cm3; Waters, MA, USA) after centrifugation. ABA was quantified using a high-performance liquid chromatography (HPLC)-electrospray ionization-tandem mass spectrometry method by comparing the peak areas with those of known amounts of standard ABA (Li et al., 2012). Three independent biological replicates and two technical repeats were performed for each line.
For Affymetrix GeneChip array analysis, WT and 35S::bHLH122 plants were grown in soil for 3 wk at 22°C with a photoperiod of 16 h : 8 h light : day. Total RNA was extracted with an RNeasy Plant Mini Kit (Qiagen) and used for the preparation of biotin-labeled complementary RNA targets. Microarray analysis was performed as described by Breitling et al. (2004). Two biological replicates were used for each genotype. We normalized the expression profiles with the RMA (Robust Multichip Average) method (Irizarry et al., 2003). A list of genes whose expression differed significantly between the genotypes was generated by the RankProd method, in which multiple testing was taken into account by the use of pfp (percentage of false prediction; pfp < 0.05; Gentleman et al., 2004; Hong et al., 2006). All data were deposited in the National Center for Biotechnology Information (NCBI), and are accessible through GEO Series accession number GSE46661 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE46661).
bHLH122 expression is significantly induced by abiotic stresses
To identify stress-related bHLH genes in Arabidopsis, we analyzed in silico the expression pattern of bHLH genes in the publicly available Arabidopsis eFP Browser microarray database (http://www.bar.utoronto.ca/efp/cgi-bin/efpWeb.cgi). Among the 168 Arabidopsis bHLH genes, including novel atypical bHLHs (Carretero-Paulet et al., 2010), bHLH122 (At1g51140) attracted our attention. The microarray data indicated that bHLH122 transcript levels were strongly induced in both shoots and roots by osmotic stress (300 mM mannitol) and salt stress (150 mM NaCl), but not by ‘drought’ treatment (Fig. S1). To verify the microarray data, we performed quantitative real-time RT-PCR using RNA isolated from drought-, NaCl- and mannitol-treated Arabidopsis, and the results confirmed the microarray data under osmotic and salt conditions (Fig. 1a). Unexpectedly, bHLH122 expression was also significantly up-regulated by drought stress, that is, bHLH122 expression increased c. 13-fold after water was withheld for 10 d. The inconsistency between the results of the quantitative real-time RT-PCR and microarray data might be caused by the difference in drought treatment. In microarray analysis, drought was generated by exposing 18-d-old seedlings grown on rafts floating on liquid MS medium to an air stream for 15 min; the brief air exposure was evidently insufficient to activate drought-responsive genes.
ABA accumulation is required for some drought stress-induced up-regulation of gene expression (Zhu, 2002). Thus, we tested the response of bHLH122 to ABA treatment. We could not detect any significant variations in the expression of bHLH122 with exogenous ABA treatment (100 μM; Fig. 1b). The drought-inducible expression of bHLH122 also did not change in an ABA-deficient mutant (aba2-1) and an ABA-insensitive mutant (abi4-1; Fig. 1b). This result suggests that bHLH122 expression is at least partly independent of ABA signaling.
Expression pattern of bHLH122
Quantitative real-time RT-PCR analysis revealed that bHLH122 was expressed in various tissues of Arabidopsis, including roots, leaves, stems and flowers (Fig. 2a). To further determine the expression pattern of bHLH122, we analyzed the expression of a bHLH122 promoter-reporter gene fusion, which consisted of a 2-kb fragment upstream from the initiation codon of bHLH122 fused with GUS. Analysis of > 15 independent Arabidopsis transgenic lines confirmed the presence of GUS activity at all developmental stages tested, from germination to flowering (Fig. 2b). Histochemical staining revealed that bHLH122 was expressed in cotyledons, root vascular tissues and root tips of 8-d-old seedlings (Fig. 2b, panels I, V and VI). bHLH122 expression was high in leaf tissues, including the leaf vascular systems and, importantly, guard cells (Fig. 2b, panels II, III and IV). GUS staining was also evident in floral tissues and siliques (Fig. 2b, panels VII and VIII).
The Psort program predicted a nucleus localization of bHLH122 with 95% certainty (http://psort.hgc.jp/form.html). To determine the subcellular localization of bHLH122 protein, we transformed bHLH122-GFP construct under the control of the cauliflower mosaic virus (CaMV) 35S promoter into onion epidermis cells by bombardment. A strong fluorescent signal derived from GFP alone was observed in the cytoplasm and nuclei (Fig. 2c), whereas transformed cells carrying bHLH122-GFP showed a strong green fluorescence signal in the nucleus, demonstrating the nucleus localization of bHLH122.
Overexpression of bHLH122 increases the drought resistance of Arabidopsis
The drought-inducible expression of bHLH122 and its strong expression in guard cells prompted us to analyze its potential role in drought resistance. Three independent transgenic lines (#1, #13 and #15) were chosen for further analysis based on their expression levels (Fig. 3a). WT and 35S::bHLH122 transgenic plants were grown for 3 wk in soils before water was withheld for 13 d. Most WT plants wilted and their leaves became purple, whereas the 35S::bHLH122 transgenic plants remained turgid and their leaves remained green (Fig. 3b). When data were analyzed from three different experiments, 32–50% of the 35S::bHLH122 transgenic plants survived drought stress and recovered from the stress after re-watering, which was significantly higher than for WT plants (Fig. 3c). These results suggested that WT may have depleted the soil water more rapidly than 35S::bHLH122 transgenic plants, and thus wilted more quickly. To investigate this possibility, we measured the stomatal apertures of leaves from WT and 35S::bHLH122 plants grown in soil. Under normal conditions, the stomatal aperture index of WT plants was 0.48, and those of 35S::bHLH122-13 and 35S::bHLH122-1 plants were 0.42 and 0.43, respectively (Fig. 3d). Although the difference between WT and 35S::bHLH122 transgenic plants under normal conditions was not statistically significant, the stomatal aperture indices of 35S::bHLH122-13 and 35S::bHLH122-1 plants after 10 d of drought decreased to 0.32 and 0.20, respectively (Fig. 3d), which was 40–63% smaller than that of the WT. Consistent with these results, detached leaves of 35S::bHLH122 transgenic plants lost water much more slowly than those of WT plants after dehydration for 4 h (Fig. 3e), suggesting that the more rapid appearance of wilting after water was withheld in the WT could at least partly be attributed to the inability of these plants to efficiently close their stomata and reduce transpiration. Another indicator of stress sensitivity is the accumulation of the purple flavonoid pigment, anthocyanin, in leaves. The anthocyanin levels in 35S::bHLH122-13 and 35S::bHLH122-1 plants after 10 d of drought were 9.77 and 8.64 μg g−1 fresh weight (FW), respectively, which were c. 75% less than that of the WT (Fig. 3f), again supporting the inference that 35S::bHLH122 transgenic plants are more resistant to drought stress. The results suggest that adequate expression of bHLH122 is required for drought resistance in Arabidopsis.
35S::bHLH122 transgenic plants are resistant to salt and osmotic stresses
bHLH122 is a salt- and mannitol-induced gene (Fig. 1a), indicating that bHLH122 might also be necessary for plant responses to osmotic stresses. To test this possibility, we germinated and grew WT and 35S::bHLH122 plants on MS medium containing different concentrations of NaCl. On MS medium without NaCl, no obvious differences were evident at the germination stage (Fig. 4a–c). The addition of 140 mM NaCl significantly inhibited cotyledon greening/expansion of WT, but not of 35S::bHLH122 transgenic plants (Fig. 4a). Based on radicle emergence, 35S::bHLH122 plants were more tolerant to NaCl stress than the WT at the germination stage (Fig. 4b). The number of green cotyledons 10 d after sowing also indicated that 35S::bHLH122 plants were more resistant to NaCl stress than the WT (Fig. 4c). To further evaluate the effects of NaCl stress on the growth of transgenic plants, 4-d-old WT and 35S::bHLH122 seedlings were placed on MS medium containing different concentrations of NaCl. After 10 d, the biomasses were significantly higher for 35S::bHLH122 transgenic plants than for WT plants at all NaCl concentrations tested (Fig. S2), indicating that bHLH122 is involved in responses to NaCl stress at both germination and post-germination developmental stages.
In contrast with the germination rates on MS medium containing NaCl, the germination rates on MS medium containing various concentrations of mannitol were similar for 35S::bHLH122 transgenic and WT plants (Fig. S3). In addition, root elongation of both WT and 35S::bHLH122 plants was reduced by osmotic stress (Fig. 4d,e). The sensitivity of root growth to osmotic stress, however, was significantly greater for WT than for 35S::bHLH122 plants, in that WT roots but not 35S::bHLH122 roots tended to curl abnormally when mannitol was added to the medium (Fig. 4d). On MS medium containing 250 mM mannitol, the average root length of three independent 35S::bHLH122 transgenic lines ranged from 1.9 to 2.2 cm, which was c. three-fold greater than that of WT plants (Fig. 4e). The results demonstrate that osmotic stress resistance in Arabidopsis requires bHLH122.
bhlh122 loss-of-function mutants are hypersensitive to salt and osmotic stresses
To further characterize the function of bHLH122 in abiotic stress resistance, we searched the publicly available T-DNA collections and obtained a T-DNA insertion mutant (SALK_049022C in the Columbia background) from ABRC. Plants homozygous for the T-DNA insertion were identified by PCR, and sequencing of the T-DNA flanking region confirmed the insertion site in the first exon of bHLH122 (Fig. 5a). RT-PCR analysis showed that bHLH122 transcript was absent in the T-DNA line designated as bhlh122 (Fig. 5a).
Although we did not observe obvious differences in drought resistance between WT and bhlh122 loss-of-function mutants, bhlh122 plants were more sensitive to salt and osmotic stresses than WT at germination and post-germination stages (Fig. 5b–d). These results further show that bHLH122 is required for salt and osmotic resistance in Arabidopsis.
To identify the molecular events involved in the bHLH122-mediated signaling pathway, we compared whole-transcriptome profiles (with Affymetrix Arabidopsis ATH1 Genechips) of WT and 35S::bHLH122 transgenic plants that had been grown for 3 wk in soil. We sampled only the shoots for this microarray experiment because root phenotypes and development were similar for WT and 35S::bHLH122 plants under normal conditions. The expression of 214 genes differed significantly between 35S::bHLH122 and WT plants under non-stress conditions (pfp < 0.05; Table S2). Among the 214 genes, 87 were up-regulated and 127 were down-regulated (Table S2). Some of these genes have known or presumed functions associated with abiotic stress responses, such as those that encode protease inhibitor/seed storage/lipid transfer protein family, universal stress proteins, reactive oxygen species (ROS) homeostasis protein, WRKY/MYB transcription factors and protein kinase. Of these genes affected by bHLH122 ectopic expression, 92.0% of up-regulated genes and 94.5% of down-regulated genes contain the G-box (5′-CACGTG-3′) and/or E-box (5′-CANNTG-3′) motifs in their promoter regions predicted by the PLACE program (Table S2; Higo et al., 1999); these short sequence motifs were expected because they were predicted to be DNA-binding sites of bHLHs (Toledo-Ortiz et al., 2003; Feller et al., 2011).
We then confirmed the microarray results by quantitative real-time RT-PCR. In agreement with our microarray data, the quantitative real-time RT-PCR assay showed that the expression levels of At5g15450 (Albino and pale green 6), At4g09760 (Choline synthase), At3g14205 (Phosphoinositide phosphatase family protein) and At5g64940 (Oxidative stress-related ABC1-like protein 1) were higher in 35S::bHLH122 plants under normal conditions than in WT plants (Fig. S4), suggesting that the expression of stress-responsive genes in 35S::bHLH122 plants is constitutive. As expected, At4g17490 (ERF6) and At5g57560 (TCH4) were expressed at lower levels in 35S::bHLH122 plants than in WT plants (Fig. S4), demonstrating the reliability of our microarray data.
bHLH122 can bind to the G-box cis-element
Based on maximum likelihood phylogenetic analysis, plant bHLHs have been classified into 26 subfamilies, with bHLH122 belonging to subfamily IX. In this subfamily, most members, including bHLH122, have been predicted to bind to an E-box cis-element rather than a G-box cis-element, but this prediction lacks experimental support (Pires & Dolan, 2010). To test the DNA-binding activity of bHLH122, we performed EMSAs using digoxygenin-labeled E-box or G-box elements derived from the At2g46830 promoter, and bacterially expressed bHLH122 proteins fused to GST. The retarded protein complexes were detected in the presence of GST-bHLH122 recombinant protein and probe containing the E-box cis-element, and the binding activity increased with increasing concentrations of GST-bHLH122 recombinant protein (Fig. 6a). When the unlabeled E-box probe was added to the system as a competitor, the signal was suppressed (Fig. 6a). Unexpectedly, when the E-box probes in the system were replaced by G-box elements, the retarded protein complexes were also observed, although the binding activity was weaker to G-box element than to E-box element (Fig. 6a). When the unlabeled G-box probes were added as a competitor for the E-box element, the signal did not show significant changes (Fig. 6a). In contrast, when the unlabeled E-box probes were added as a competitor for the G-box element, the signal was dismantled (Fig. 6a).
To determine whether bHLH122 interacts with G-box/E-box motifs in vivo, ChIP assays were performed. The FLAG-bHLH122 fusion protein was expressed under the control of the 35S promoter and immunoprecipitated using an anti-FLAG antibody. The DNA fragments of testing genes without E-box/G-box cis-elements were used as negative controls. By searching our microarray data for genes in which E- and G- boxes occurred separately, At3g14205 (three E-box elements), At4g17490 (two E-box elements) and At5g64940 (one G-box element) were selected for further research (Fig. S5). DNA precipitated without the anti-FLAG antibody was subjected to PCR amplification, and the testing genes were not detected (Fig. 6b). The At3g14205, At4g17490 and At4g17490 bands were amplified when FLAG-bHLH122 was precipitated, but not for corresponding DNA fragments without E-box/G-box elements (Fig. 6b), demonstrating that bHLH122 can bind to the G-box/E-box cis-elements in vivo.
bHLH122 can directly repress CYP707A3 expression
Among the 214 genes whose expression differed significantly in WT vs 35S::bHLH122 plants, CYP707A3 attracted our attention because it is a major ABA 8′-hydroxylase essential for ABA catabolism (Kushiro et al., 2004; Umezawa et al., 2006; Okamoto et al., 2009). Using the PLACE program (Higo et al., 1999), we found two G-box and six G-box cis-elements in the 1-kb region upstream from the initiation codon of the CYP707A3 gene (Table S2). A 197-bp DNA fragment containing one G-box from the CYP707A3 promoter was cloned (Fig. 7a), and labeled with digoxygenin-11-ddUTP as a probe for EMSAs. The results of the EMSAs showed that bHLH122 was tightly bound to the CYP707A3 promoter because, even at low concentrations of GST-bHLH122, the free DNA probes could not be detected (Fig. 7b). When unlabeled CYP707A3 probe was added to the system as a competitor, the signal was suppressed (Fig. 7b), further demonstrating the direct binding of the CYP707A3 promoter by bHLH122.
To test the in vivo binding ability of bHLH122 to the CYP707A3 promoter, a ChIP assay was carried out using 35S::FLAG and 35S::FLAG-bHLH122 transgenic plants. After immunoprecipitation with an antiserum against FLAG, the DNA fragments containing one G-box and one E-box motif were amplified (Fig. 7a). An obvious CYP707A3 band was amplified when FLAG-bHLH122 was precipitated (Fig. 7c). PCR products were not detected in any samples amplified with the primer set that did not span the E-box/G-box elements, demonstrating that CYP707A3 is a target gene of bHLH122.
We then analyzed bHLH122 regulation of CYP707A3. Our array data indicated that CYP707A3 was a bHLH122 down-regulated gene. To test this hypothesis, RNA was purified from two independent 35S::bHLH122 transgenic lines and the bhlh122 loss-of-function mutant, and quantitative real-time RT-PCR analysis was performed. Consistent with our array data, real-time RT-PCR results showed that the CYP707A3 expression levels were c. 60% lower in 35S::bHLH122 transgenic plants than in WT plants (Fig. 7d). Conversely, the CYP707A3 expression level in the bhlh122 mutant was about three times greater than in WT plants (Fig. 7d).
The cyp707a3 loss-of-function mutant contained higher ABA levels in turgid plants, and accumulated much greater amounts of stress-induced ABA after dehydration than did WT plants (Umezawa et al., 2006). To further verify the relationship between CYP707A3 and bHLH122, the ABA contents in 35S::bHLH122 transgenic lines with or without drought treatment were determined. We did not measure the ABA content in the bhlh122 mutant, which showed no difference from WT after drought for 13 d. The ABA contents in 35S::bHLH122-1 and 35S::bHLH122-13 plants were 8.08 and 8.18 ng g−1 FW, which were c. 33% and 35% higher than in WT, respectively (Fig. 7e). The higher ABA concentration in 35S::bHLH122 transgenic plants might lead to a similar water loss to WT plants in the early dehydration treatment. After drought for 10 d, the ABA content in WT plants was 63.6 ng g−1 FW, whereas those in 35S::bHLH122-1 and 35S::bHLH122-13 plants increased to 218.1 and 115.1 ng g−1 FW, respectively (Fig. 7e). These results further demonstrated that CYP707A3 is a bHLH122 down-regulated gene.
Gene regulation under abiotic stress is mediated by multiple transcriptional cascades (Zhu, 2002; Yamaguchi-Shinozaki & Shinozaki, 2006). In each of these cascades, a transcription factor gene is induced, which, in turn, activates or represses downstream target genes important for abiotic resistance. bHLH122 may define one of these abiotic stress-responsive transcriptional cascades by recognizing the G-box (CACGTG) and/or E-box (CANNTG) elements to regulate the downstream targets. This transcriptional cascade is critical for abiotic resistance because 35S::bHLH122 transgenic plants are more resistant to drought, salt and osmotic stresses, and bhlh122 loss-of-function mutants are hypersensitive to salt and osmotic stresses.
The tolerance of Arabidopsis to drought stress is mediated by ABA-dependent signaling and ABA-independent pathways involving other drought-responsive factors, such as DREB2A and ERD1 (Early responsive to dehydration stress 1), and crosstalk occurs between ABA-dependent signaling and ABA-independent pathways (Ishitani et al., 1997; Shinozaki et al., 2003; Shinozaki & Yamaguchi-Shinozaki, 2007). ABA-dependent and ABA-independent signaling pathways have been mainly investigated from gene marker analysis in ABA biosynthesis mutants or ABA-insensitive mutants, and the downstream signaling ceases in the dominant-positive ABA-insensitive mutants. Using the PlantCARE program (http://bioinformatics.psb.ugent.be/webtools/plantcare), we detected six ABRE (ABA-responsive element) sequences in the 2-kb region upstream from the initiation codon of the bHLH122 gene. However, we could not detect any significant variations in the expression of bHLH122 with exogenous ABA treatment (Fig. 1b). The drought-inducible expression of bHLH122 also did not change in an ABA-deficient mutant (aba2-1) and an ABA-insensitive mutant (abi4-1; Fig. 1b). Our microarray analysis showed that some stress-responsive genes insensitive to exogenous ABA application might be the targets of bHLH122, because G-box/E-box cis-elements were predicted in their promoters, such as CCA1 (Circadian clock-associated 1), SZF1 (Salt-inducible zinc finger) and ZAT10 (Salt-tolerance zinc finger). Previous research has shown that CCA1 affects the transcriptional regulation of ROS-responsive genes, ROS homeostasis and tolerance to oxidative stress (Lai et al., 2012); transgenic plants overexpressing SZF1 show reduced induction of salt stress-responsive genes and increased tolerance to salt stress (Sun et al., 2007); and relative to the WT, 35S::ZAT10 transgenic Arabidopsis exhibits elevated expression of reactive oxygen defense transcripts and enhanced tolerance to salinity, heat and osmotic stresses (Mittler et al., 2007). These results indicate that bHLH122 may directly regulate the expression of stress-responsive genes through an ABA-independent pathway (Fig. 8). However, it should be noted that up-regulation of bHLH122 produced higher levels of ABA than in WT plants, even without stress conditions (Fig. 7e), indicating that significant cross-talk between the ABA-dependent signaling and ABA-independent pathways is involved in the stress resistance in 35S::bHLH122 transgenic plants (Fig. 8).
The increased level of endogenous ABA in 35S::bHLH122 transgenic plants might be caused by the direct repression of CYP707A3 transcripts by bHLH122 (Fig. 7). CYP707A3, a major ABA 8′-hydroxylase, determines the threshold levels of ABA during dehydration and after rehydration in Arabidopsis (Umezawa et al., 2006). Moreover, the cyp707a3 loss-of-function mutant exhibits both exaggerated ABA-inducible gene expression and enhanced drought tolerance (Umezawa et al., 2006). However, the regulation of CYP707A3 was not characterized in these previous studies. VIP1 (VirE2-interacting protein 1), a bZIP transcription factor involved in Agrobacterium tumefaciens transformation of Arabidopsis, has been shown to bind to CYP707A3 promoters in EMSA and ChIP assays (Tsugama et al., 2012). In a transient reporter assay in Arabidopsis protoplasts, VIP1 enhanced CYP707A3 promoter activities (Tsugama et al., 2012). However, vip1-1 did not cause any significant difference in the responses to either mannitol or dehydration/rehydration treatments (Tsugama et al., 2012), indicating that other factors are required for the regulation of CYP707A3 response to abiotic/biotic stresses. In this research, we demonstrated that bHLH122 could bind directly to the CYP707A3 promoter by EMSA and ChIP analyses (Fig. 7b,c). The CYP707A3 mRNA levels were down-regulated in 35S::bHLH122 transgenic plants, but up-regulated in the bhlh122 mutant (Fig. 7d). These results suggest that, by repressing CYP707A3 mRNA levels and thereby enhancing ABA content, bHLH122 functions as a positive regulator of drought, salt and osmotic signaling in Arabidopsis.
We thank Dr Jun Zheng (Institute of Crop Science, Chinese Academy of Agricultural Sciences) for stomatal aperture analysis, and Dr Miaoyun Xu (Biotechnology Research Institute, Chinese Academy of Agricultural Sciences) for ChIP assays. This work was supported by the National High-Tech R&D Program (grant number 2012AA10A306), the National Science Foundation of China (grant number 31071139, 31370303) and the Core Research Budget of the Non-profit Governmental Research Institution from the China Government to the Institute of Crop Science, Chinese Academy of Agricultural Sciences (Grant number 2013001) to W-X.L.