Abstract
Proline (Pro) metabolism plays important roles in protein synthesis, redox balance, and abiotic stress response. However, it is not known if cross-talk occurs between proline and brassinosteroid (BR) signaling pathways. Here, an Arabidopsis intergenic enhancer double mutant, namely proline content alterative 41 (pca41), was generated by inserting a T-DNA tag in the Arabidopsis thaliana ring zinc finger 1 (atrzf1 ) mutant background. pca41 had a T-DNA inserted at the site of the gene encoding BES1/BZR1 Homolog 3 (BEH3). pca41 has a drought-insensitive phenotype that is stronger than atrzf1 under osmotic stress, including high Pro accumulation and decreased amounts of reactive oxygen species. Analysis of physiological, genetic, and molecular networks revealed that negative regulation of BEH3 during abiotic stress was linked to the BR signaling pathway. Our data also suggest that AtRZF1, an E3 ubiquitin ligase, might control osmotic stress, abscisic acid, and BR responses in a BEH3-dependent manner. Under darkness, pca41 displays a long hypocotyl phenotype, which is similar to atrzf1 and beh3, suggesting that BEH3 acts in the same pathway as AtRZF1. Overexpression of BEH3 results in an osmotic stress-sensitive phenotype, which is reversed by exogenous BR application. Taken together, our results indicate that AtRZF1 and BEH3 may play important roles in the osmotic stress response via ubiquitination and BR signaling.
Introduction
Proline (Pro) is a compatible solute that regulates plant responses to various abiotic stresses such as salinity, drought, cold, oxidative, and biotic stress (Verbruggen and Hermans, 2008; Szabados and Savouré, 2010; Singh et al., 2014). Under abiotic stress conditions, Pro plays an important role in attenuating dehydration and balancing turgor pressure, stabilizing protein structure, and maintaining cellular redox potential (Szabados and Savouré, 2010).
Arabidopsis thaliana Ring Zinc Finger 1 (AtRZF1) encodes a deduced C3H2C3-type E3 ubiquitin ligase. It was first reported as a suppressor of Pro accumulation (Ju et al., 2013). An atrzf1 mutant showed a phenotype insensitive to various abiotic stresses (Ju et al., 2013; Kim et al., 2017; Nguyen et al., 2019; Shin et al., 2019). Moreover, AtRZF1 regulates abiotic stress responses by modeling the abscisic acid (ABA) signaling pathway, suggesting that it is an ABA-related component. To further study Pro metabolism function in plant responses to drought stress, we generated a T-DNA mutant pool using the atrzf1 background. The mutants exhibiting Pro levels distinct from the parent atrzf1 were collected. They were termed pca (proline content alterative) mutants. Among these pca mutants, some mutants showed suppressed atrzf1 drought-insensitivity traits while others showed enhanced atrzf1 drought-insensitivity traits under abiotic stress (Kim et al., 2017; Park et al., 2017; Nguyen et al., 2019; Shin et al., 2019). It was found that pca mutants showing enhanced atrzf1 drought-insensitivity traits under abiotic stress had higher Pro content than the atrzf1 mutants under osmotic stress conditions at early seedling growth stages. One of these pca mutants, pca41, was characterized in this study. A T-DNA tagged At4g18890 gene of pca41 encodes BEH3 (BES1/BZR1 Homolog 3), known to regulate brassinosteroid (BR) signaling (Wang et al., 2002).
BRs are essential steroid hormones that regulate plant growth, development, and responses to environmental stresses (Krishna, 2003; Clouse, 2011). In recent decades, BR signaling and its related components, including BES1/BZR1 (BRI1 EMS SUPPRESSOR 1/BRASSINAZOLE RESISTANT 1) and its downstream transcription factors that regulate BR responses and BR biosynthesis, have been identified and characterized (Wang et al., 2002; Yin et al., 2005; Yu et al., 2011). To date, the cross-linking between BR signaling and abiotic stress has been discovered. Several studies have implied that exogenous BRs could enhance tolerance of plants to drought stress (Sairam, 1994; Krishna, 2003; Kagale et al., 2007). However, other groups have reported that some BR-deficient mutants exhibit a drought-tolerance phenotype (Noguchi et al., 1999; Beste et al., 2011; Feng et al., 2015). BES1 and BZR1 can interact with some drought-inducible transcription factors such as Responsive to Desiccation 26 (RD26) and its homologs (Sun et al., 2010; Yu et al., 2011), to regulate drought stress responses. Furthermore, upstream regulators of BES1/BZR1, such as BRASSINOSTEROID INSENSITIVE 2. (BIN2 ) and its homologs, can control ABA signaling output responses in adaptation to abiotic stress (Wang and Wang, 2018). Nevertheless, the mechanisms by which BEH3 is involved in BR signaling to regulate plant growth, development, and environmental responses is not known. Indeed, BRs can regulate hypocotyl elongation by interplaying with other plant hormones (Peres et al., 2019). For example, transcriptional regulation of Auxin Response Factors (ARFs) genes by BR occurs during hypocotyl elongation (Peres et al., 2019), implying that there is an interaction between BR and auxin signaling. Recently, several groups have suggested that exogenous BRs can enhance proline accumulation under temperature stress conditions (Wu et al., 2014; Zhang et al., 2014). However, the detailed molecular network linking these interactions in plant growth remains elusive. Thus, further investigation into the cross-talk between BR and other hormones or signaling molecules (such as proline) will be necessary to clarify plant growth processes under abiotic stress.
In this study, we provide evidence that loss of BEH3 in pca41 further increases Pro accumulation and water content. It also reduces malondialdehyde (MDA) and reactive oxygen species (ROS) accumulation, resulting in an enhanced drought-insensitive trait of atrzf1 mutant. Analysis of physiological, genetic, and molecular networks revealed that BEH3 was a negative regulator of osmotic stress, linked to BR signaling responses via an AtRZF1-BEH3 coordinated pathway. Moreover, BEH3 negatively regulates hypocotyl elongation during dark growth by suppressing the expression of BR biosynthesis genes. These data imply that AtRZF1 and BEH3 can co-regulate osmotic stress and BR signaling response cross-talk.
Materials and methods
Plant growth, development, and stress induction
Arabidopsis thaliana ecotype Columbia-0 (Col-0), atrzf1, pca41, beh3 and BEH3 transgenic plants were grown in standard chambers (22 °C, 60% relative humidity, and 16 h of day light, 120 μmol m-2 s-1). For ABA and osmotic stress, two-week-old plants were submerged and gently shaken in solutions containing 100 μM ABA or 400 mM mannitol. Samples were harvested at 0, 3, 6, 12, and 24 h after treatment with ABA or mannitol. In each case, obtained seedlings were frozen in liquid nitrogen and then stored at -80 °C. For drought stress, plants were grown in the same pots for two weeks. These plants were then divided into two groups. One group was well-watered every four days under normal conditions, while the other group was treated with drought stress without watering for 14 days. Surviving plants were counted after seven days of re-watering.
Determination of proline, malondialdehyde content, and water loss rate
Pro accumulation was estimated according to Bates et al. (1973). Briefly, leaves (0.5 g) were ground in 5 ml of 3% sulfosalicylic acid. Then 250 μl of the extract was used to react with 150 μl of ninhydrin reagent buffer solution (80% glacial acetic acid, 6.8% phosphoric acid, 70.17 mM ninhydrin) for 1 h at 90 °C, cooled rapidly, and centrifuged at 2,570 × g for 20 min at 4 °C. The supernatant was collected and 400 μl toluene was added, followed by light swirling. The absorbance value of the toluene layer was measured at 520 nm with a UV/VIS spectrophotometer (JASCO, Tokyo, Japan). Pro concentration was determined based on a standard curve.
Malondialdehyde (MDA) content was determined by the thiobarbituric acid (TBA) reaction, as described previously (Shin et al., 2019). Briefly, leaf samples (2-5 g) were ground with 2-5 ml of 0.1% trichloroacetic acid (TCA) and centrifuged at 10 280 × g for 15 min. Following this, 2-5 ml of the supernatant was mixed with 0.6% TBA, incubated in a hot water (90 °C) bath for 15 min, cooled immediately on ice, and centrifuged at 2,570 × g for 15 min. The absorbance value of the colored supernatant was measured at wavelengths of 532 nm, 450 nm, and 600 nm. MDA concentration was estimated by subtracting non-specific absorbance at 450 nm and 600 nm, compared with absorbance at 532 nm.
Water loss rate was determined after detached leaves were placed in open-lid plastic dishes at room temperature with 60% moisture under dim light (43 μmol m-2 s-1). The weight of each sample was measured at different times. Water content in the leaf was calculated based on the initial fresh weight of the leaf.
Analysis of physiological phenotype under abscisic acid, mannitol, and brassinosteroid treatments
For phenotype analysis and abiotic stress experiments, seeds of all genotypes were grown on tissue culture plates with one-half-strength Murashige and Skoog (MS) medium containing 1.5% (w/v) sucrose, without or with 400 mM mannitol or 0.7 μM ABA. Germination rate was evaluated up to 6 d after seeds were sown. Cotyledon greening rate was recorded at 7-12 d after seeds were sown. For hypocotyl length analysis, seeds of all genotypes were grown in the same MS plate without or with 0.5 nM eBL (24-epibrassinolide) or 1 μM BRZ (brassinazole). These plates were vertically oriented in darkness. Each experiment consisted of 50 seeds of each genotype. The experiment was conducted three times.
Total RNA extraction, qPCR, and reverse transcription (RT)–PCR analysis
Total RNA was extracted from frozen samples (wild-type, atrzf1, pca41, beh3, BEH3 RNAi and BEH3-overexpressing plants) using a Plant RNeasy Extraction Kit (Qiagen, Valencia, CA, USA). During total RNA extraction, genomic DNA was removed by treating with RNase-free DNase I (Qiagen) according to the manufacturer’s instructions. The concentration of RNA was accurately quantified by spectrophotometric measurements. Total RNA (5 µg) was then loaded on a 1.2% formaldehyde gel to determine its concentration and integrity. Quantitative real-time PCR (qPCR) was performed with a Rotor-Gene 6000 equipment (Corbett Research, Mortlake, NSW, Australia). Results were analyzed using RG6000 1.7 software (Corbett Research). qPCR was implemented using a SensiMix One-Step (Quantance, London, UK). Arabidopsis Actin 1 (ACT1) gene was used as an internal control for normalization. Quantitative analysis was performed using the delta-delta CT method (Livak and Schmittgen, 2001). Each sample was repeated in triplicate. Primers used in qPCR are listed in Supplementary Table S1 at JXB online. RT–PCR was used to determine expression of various genes in seedlings. Total RNA (300 ng) was used as template for RT–PCR using gene specific primers listed in Table S1.
Identification of T-DNA location site in genomic DNA by Thermal Asymmetric Interlaced (TAIL)-PCR
TAIL-PCR is an efficient technique for recovery of genomic DNA sequences flanking T-DNA insertions in Arabidopsis (Liu and Chen, 2007). Genomic DNA was isolated from pca41. TAIL-PCR was performed using arbitrary degenerate 2 (AD2) and T-DNA left border-specific (LB) primers (Supplementary Table S1). PCR fragments obtained from tertiary TAIL-PCR were inserted into pGEM T-easy vector (Promega, Madison, WI, USA) for DNA sequence analysis. The obtained sequence was subjected to NCBI BLAST search. Gene-specific primers were designed (Table S1) and used in combination with LB3 primer to amplify DNA fragments. DNA fragments were then sequenced to determine the T-DNA location site.
Construction and sub-cellular localization of BEH3-GFP fusion proteins in transgenic plants
To determine the sub-cellular localization of BEH3, full-length BEH3 cDNA from WT rosette leaves was amplified by RT–PCR. The PCR primers are shown in Supplementary Table S1. BEH3 cDNA fragments were cloned into pDONR/ZEO vector (Invitrogen, Carlsbad, CA, USA) and confirmed by DNA sequencing. This construct was sub-cloned into pEarlyGate103, a plant expression vector (Earley et al., 2006) fused with a constitutive 35S promoter. To determine the sub-cellular localization of BEH3-GFP protein in transgenic plant root cells, 3- or 4-d-old seedlings were mounted onto microscope slides and monitored with an Olympus FluoView1000 confocal microscopy (Olympus, Tokyo, Japan). The reagent 4’,6-diamidino-2-phenylindole (DAPI; Sigma, St. Louis, MO, USA) was used for staining nuclei. The images generated from the constitutive CaMV35S promoter-GFP line (Min et al., 2015) were used as a negative control for GFP expression. Each GFP image was collected and processed with FV10-ASW 1.7A (Olympus) software.
Construction of BEH3 transgenic lines and analysis of GUS expression
To generate BEH3-overexpressing transgenic lines, full-length BEH3 cDNA was amplified using specific primers (Supplementary Table S1). The PCR product was cloned into a donor vector, pDONR/ZEO (Invitrogen) and confirmed by DNA sequencing. This construct was sub-cloned into a plant expression vector pGWB514 under the control of constitutive 35S promoter. The BEH3-overexpression construct was then transformed into Arabidopsis WT or pca41 mutant through in planta vacuum infiltration (Bechtold and Pelletier, 1998). Hygromycin B (A.G. Science, San Diego, CA, USA) resistance of T2 transgenic plants then segregated as a single locus. T3 or T4 homozygous transgenic lines were used to evaluate various stress responses.
To generate BEH3 RNAi lines, the BEH3 cDNA fragment was amplified using specific primers (Supplementary Table S1). PCR products were cloned into the donor vector pDONR/ZEO and confirmed by DNA sequencing. The RNAi construct was sub-cloned into a plant expression vector pB7GWIWG2ii (Karimi et al., 2002) under the control of constitutive 35S promoter. The BEH3 RNAi construct was transformed into Arabidopsis WT or atrzf1 mutant. Phosphinothricin (Duchefa Biochemie BV, Haarlem, Netherlands) resistance of T2 transgenic plants then segregated as a single locus. T3 or T4 homozygous transgenic lines were used to evaluate various stress responses.
To generate a BEH3 promoter (PRO)-GUS construct, a 1939 bp genomic DNA fragment upstream from the BEH3 translation start codon was amplified by PCR using specific primers (Supplementary Table S1). This product was cloned into the pDONR/ZEO vector for DNA sequence analysis. This construct was then sub-cloned into a plant expression vector pBGWFs7.0. The BEH3 PRO-GUS construct was transformed into Arabidopsis WT. Phosphinothricin resistance of T2 transgenic plants then segregated as a single locus. T3 or T4 homozygous transgenic lines were used to evaluate GUS expression.
Histochemical staining in BEH3 PRO-GUS transgenic plants for GUS activity was carried out as described previously (Jefferson et al., 1987). Seedlings were soaked in a solution containing 1 mM 5-bromo-4-chloro-3-indolyl-glucuronic acid (X-Gluc), 100 mM sodium phosphate, pH 7.0, 0.5 mM potassium ferrocyanide, 0.5 mM potassium ferricyanide, 10 mM EDTA, and 0.1% Triton X-100 buffer solution, followed by incubation at 37 °C for 4 h. Chlorophyll was removed from sample tissues by soaking in 75% ethanol. Whole seedlings or plant tissues were stained for GUS expression.
Isolation of beh3 single mutation line from pca41
To isolate beh3 single mutation lines from pca41 (atrzf1/beh3) mutant, pca41 was crossed with Arabidopsis WT. Responses of F1 and F2 progenies to Basta resistance during germination were then analyzed. All F1 progenies were resistant to Basta. F2 progeny displayed a segregation ratio of 3:1 for Basta resistance to Basta-sensitive germinating seeds (χ 23:1=0.022, P≤0.05; Table S2). From all Basta-resistant plants in the F3 progeny, the beh3 single homozygous mutant was isolated by genomic DNA PCR using specific region or T-DNA insertion site primers (SupplementaryTables S1, S3) of beh3 or atrzf1 mutant, respectively.
Hydrogen peroxide determination
Hydrogen peroxide (H2O2) accumulation was detected using 3,3′-diaminobenzidine (DAB) staining method, as described previously (Daudi and O’Brien, 2012). Briefly, DAB was dissolved in distilled water and adjusted to pH 3.8 with KOH. Then 400 mM mannitol-treated 14-day-old seedlings were incubated with DAB staining solution for 12 h. Stained seedling samples were bleached with wash buffer (ethanol, acetic acid, and glycerol, 3:1:1, v:v:v). H2O2 content of each sample was then determined by a colorimetric method using potassium iodide, as described by Junglee et al. (2014). Briefly, 300 mg frozen sample in 2 ml of buffer solution containing 0.5 ml of 0.1% (w/v) trichloroacetic acid (TCA), 1 ml of 1 M potassium iodide (KI), and 0.5 ml of 10 mM potassium phosphate buffer (pH 8.0) was homogenized at 4 °C. The homogenate was centrifuged at 10 000 × g for 20 min at 4 °C. The supernatant was then measured spectrophotometrically at 350 nm. A calibration curve obtained with H2O2 solution prepared in 0.1% TCA was used for quantification.
Measuring activities of antioxidant enzymes
Activities of antioxidant enzymes were analyzed as described by Venisse et al. (2001). Briefly, 300 mg of 400 mM mannitol-treated 14 day-old seedlings were homogenized in 3 ml extraction buffer (50 mM sodium phosphate buffer, pH 7.5, 1 mM polyethyleneglycol, 1 mM phenylmethylsulfonyl fluoride, 8% polyvinylpolypyrolydone, 0.01% Triton X-100). Homogenates were centrifuged at 14 810 × g for 30 min at 4 °C. The supernatant was then collected.
Ascorbate peroxidase (APX) activity was measured spectrophotometrically at 290 nm as described by Venisse et al. (2001). The reaction mixture contained 1 ml buffer (0.2 M Tris/HCl, pH 7.8, 0.25 mM ascorbic acid, 0.5 mM H2O2). Peroxidase (POX) activity was measured spectrophotometrically at 470 nm as described by Venisse et al. (2001). The reaction mixture contained 2 ml buffer (50 mM sodium acetate, pH 7, 25 mM guaiacol, 25 mM H2O2). Catalase (CAT) activity was measured spectrophotometrically at 240 nm as described by Mishra et al. (1993). The reaction mixture contained 3 ml buffer (50 mM potassium phosphate, pH 7, 11 mM H2O2).
Statistical analysis
All statistical analyses including one-way analysis of variance (ANOVA) and Duncan’s multi-range test were carried out using SPSS 23.0 software (IBM Co, Armonk, NY, USA). Different letters on graphs indicate that means are statistically significant at P<0.05.
Results
pca41 shows increase in tolerance to drought
Previously, it has been found that Pro and water content in the Arabidopsis atrzf1 mutant were highly increased, consistent with increased dehydration stress tolerance (Ju et al., 2013). To further analyze the role of AtRZF1 in Pro metabolism and cross-linked molecules in response to abiotic stress, pca mutants were generated. To find candidate molecules for the enhancer of the atrzf1 drought-insensitive phenotype, Pro accumulation was assessed in excised leaves from five-week-old T1 transgenic plants and atrzf1, subjected to drought stress. Candidate enhancer mutants were isolated in which Pro accumulation was more increased than in atrzf1. One of these enhancer mutants, proline content alterative 41 (pca41), is presented in Fig. 1A. Pro content was up-regulated in wild-type (WT), atrzf1, and pca41 after drought stress, compared with controls. Among these, pca41 showed significantly higher (P<0.05, ANOVA) Pro content than both WT and atrzf1 (Fig. 1A).
Comparative analysis of drought-induced physiological parameters in WT, atrzf1, and pca41 plants. (A) Proline accumulation in each plant genotype under normal or drought stress condition. Data represent mean and error bars from three different replicates; each replicate contained at least 10 individual seedlings per sample. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05). (B) Drought insensitive phenotype of pca41 mutant. Two-week-old plants of each genotype were grown on soil for 14 d without watering, and then re-watered for 7 d. Scale bar=5 cm. (C) Survival rate of plants in each genotype. Two-week-old seedlings were grown on soil for 14 d without watering, then re-watered for 7 d. Surviving seedlings were scored. Data represent mean and error bars from three different replicates; each replicate contained at least 35 individual plants per sample. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05). (D) Water loss rate in four-week-old plants of each genotype. Rosette leaves of the same stage were excised after drought treatment, and weighed at indicated time points. (E) MDA accumulation in plants of each genotype. Four-week-old plants were grown for 14 d without watering. Drought-treated leaf tissues were excised and used to measure MDA content. For (D) and (E), data represent mean and error bars from three different replicates; each replicate contained at least seven individual leaves per sample. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05).
To further observe the response of pca41 to drought stress, survival analysis of WT, atrzf1, and pca41 plants following drought stress treatment was performed. As shown in Fig. 1B, there was no phenotypic difference among two-week-old WT, atrzf1 and pca41 plants under well-watered soil conditions. However, when watering was stopped for 14 d, WT and atrzf1 showed more wilting than pca41. After seven days of re-watering, atrzf1 and pca41 displayed greener and healthier phenotypes than WT, whereas pca41 showed less signs of stress than atrzf1 mutants. We also estimated survival rates of WT, atrzf1, and pca41 plants after re-watering for seven days. Survival rates were approximately 82.3% for pca41, 63% for atrzf1, and 44.2% for WT (Fig. 1C).
We further investigated drought tolerance, water loss rate, and MDA content of WT, atrzf1, and pca41 plants. Water loss rate and MDA content of pca41 were lower than those of WT and atrzf1 plants under drought conditions (Fig. 1D, E). These results indicated that pca41 had enhanced drought tolerance phenotype, compared with atrzf1.
pca41 shows enhanced insensitivity to osmotic stress and abscisic acid in early seedling growth
To evaluate whether pca41 was involved in dehydration and ABA responses, we measured the percentage of seed germination, cotyledon greening rate, and fresh weight of WT, atrzf1, and pca41, in response to mannitol or ABA.
Under normal condition (MS medium), there were no differences in early seedling growth among WT, atrzf1, and pca41 plants (Fig. 2A). The seed germination rate was similar among WT, atrzf1, and pca41 at indicated days, although WT showed lower seed germination than atrzf1 and pca41 two days after seeds were sown on MS medium containing 400 mM mannitol (Fig. 2B). However, the seed germination rate of pca41 was higher than WT and atrzf1 plants at 2 d and 3 d after sowing seeds on MS medium containing 0.7 μM ABA (Fig. 2C). These results indicated that pca41 mutant showed higher germination rates than atrzf1 under ABA treatment, but not during osmotic stress conditions.
Germination rates, greening rates and fresh weight of WT, atrzf1, and pca41 treated with osmotic stress or ABA. (A) Seeds of each genotype were sown on MS plates and permitted to grow for 7 d. (B and C) Seed germination rate. The percentage of seeds that germinated on 400 mM mannitol (B) or 0.7 μM ABA (C) is shown for each genotype at indicated days. (D) Seeds were sown on MS plates containing 400 mM mannitol and permitted to grow for 9 d or 11 d. (E and F) Cotyledon greening rate. Seeds were sown on MS plates supplemented with 400 mM mannitol (E) or 0.7 μM ABA (F) and allowed to grow for indicated days. Seedlings with green cotyledons were counted. For (C), (D), (E) and (F), data represent mean and error bars from three different replicate; each replicate contained at least 50 seeds per sample. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05). (G) Seeds were sown on MS plates containing 0.7 μM ABA and permitted to grow for 8 d or 10 d. (H and I) Inhibition of fresh weight by exposure to osmotic stress and ABA. Fresh weight was measured for seedlings from each genotype grown on MS plates supplemented with 400 mM mannitol (H) for 12 d or 0.7 μM ABA (I) for 10 d. Data represent mean and error bars from three different replicates; each replicate contained at least 20 individual seedlings per sample. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05). Scale bars=1 cm.
Comparison of pca41 with atrzf1 showed notably higher percentage of green cotyledons at indicated days after sowing seeds on MS medium containing 400 mM mannitol (Fig. 2D, E) or 0.7 μM ABA (Fig. 2F, G). Furthermore, the fresh weight of pca41 was significantly higher (P<0.05, ANOVA) than WT and atrzf1 under mannitol or ABA stress conditions (Fig. 2H, I). These physiological findings demonstrated that pca41 showed more insensitivity in early seedling growth than WT and atrzf1, under osmotic stress or ABA.
PCA41 acts negatively on osmotic stress-related genes
It has been reported that expression of Responsive to Desiccation 29A (RD29A), RD29B, Arabidopsis thaliana MYB 2 (AtMYB2), Arabidopsis thaliana oxidation-related zinc finger 2 (AtOZF2), Delta 1-pyrroline-5-carboxylate synthase 1 (P5CS1), and Delta 1-pyrroline-5-carboxylate reductase (P5CR) genes that are up-regulated under osmotic stress are induced more in atrzf1 than in WT (Min et al., 2015; Kim et al., 2017; Shin et al., 2019). To analyze the expression of these stress-inducible genes in pca41 in comparison with WT and atrzf1, quantitative real-time PCR (qPCR) was performed. As shown in Fig. S1A-F, expression of RD29A, RD29B, AtMYB2, AtOZF2, P5CS1, and P5CR was highly induced after mannitol treatment. The expression of all these genes in atrzf1 and pca41 were higher than those in WT. These observations indicate that the expression of these stress-inducible marker genes in pca41 and atrzf1 was regulated similarly under osmotic stress conditions. It is likely that both AtRZF1 and PCA41 act negatively on osmotic stress-related genes. Thus, PCA41 may share a similar function with AtRZF1in the osmotic stress response pathway .
pca41 has a single T-DNA fragment inserted near the 3’-untranslated region of At4g18890
To determine the location of the T-DNA insertion site on pca41 genomic DNA, TAIL-PCR was conducted using T-DNA specific LB3 and arbitrary AD2 primers (Supplementary Table S1). The T-DNA insertion in pca41 was identified 2169 bp downstream from the translation start codon of At4g18890 (Fig. 3A). To ensure the location of the T-DNA insertion site from pca41 genomic DNA, we prepared genome-specific primers (F1, forward 1; R1, reverse 1; Fig. 3A). As shown in Fig. 3B, using these primers, DNA fragments (567 bp) were obtained in WT and atrzf1. However, no PCR product was detected in the pca41 mutant. When LB3 and R1 primers were used, a DNA fragment (349 bp) appeared only in the pca41 mutant. In addition, AtRZF1 expression was detected in WT by reverse transcription (RT)–PCR, although it was completely abolished in atrzf1 and pca41 mutants (Fig. 3C). These PCR-based genotyping results indicated that the T-DNA fragment insertion was associated with the pca41 phenotype.
Identification of T-DNA location site in pca41 and expression pattern of BEH3 in Arabidopsis. (A) Location of T-DNA insertion site in pca41 is shown by a triangle. pca41 has a T-DNA insertion at the 3’-UTR of At4g18890. The T-DNA sequences of the left border side from pSKI015 plasmid are given in lower case letters and the genomic flanking sequences in capital letters. Small arrows indicate locations of At4g18890-specific forward I (F1) and reverse I (R1) primers and a left border-specific primer (LB3) used for PCR amplification. (B) Genomic DNA-PCR based genotyping of WT, atrzf1, and pca41 plants. Upper and lower photographs show amplicons produced using At4g18890-specific primers (F1 and R1) and the combination of R1 and LB3 primers in genomic DNA-PCR experiments, respectively. RT–PCR analysis of (C) AtRZF1 expression and (D) At4g18880, At4g18890, and At4g18900 expression in WT, atrzf1, and pca41. ACT1 was used as a loading control. (E) Expression analysis of BEH3. Expression of BEH3 was determined by qPCR using total RNA extracted from roots (Rt), stems (St), leaves (Lv), and flowers (Fl). (F) Nuclear localization of BEH3. Sub-cellular localization analysis of BEH3 fused to GFP in three-day-old transgenic Arabidopsis root cells. Confocal images show GFP fluorescence and 4’,6-diamidino-2-phenylindole (DAPI) signal in root cells from GFP vector and BEH3-GFP transgenic seedlings. DAPI staining indicates the positions of the nuclei in cells. Images were merged to show signal overlap. Scale bar=30 μm. (G and H) Expression of BEH3 in Arabidopsis seedlings following osmotic stress and ABA treatment. Expression of BEH3 was analyzed by qPCR using total RNA obtained from 14 d-old seedlings supplemented with 400 mM mannitol (G) or 100 µM ABA (H) at indicated time points. Data represent mean and error bars from three different replicates; each qPCR experiment was conducted with total RNA of each sample obtained from ten individual seedlings pool. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05). Actin 1 was used as an internal control. RAB18 served as a control for osmotic stress or ABA treatment. (I) Expression pattern of BEH3 in two-week-old BEH3 PRO-GUS transgenic lines in control, osmotic stress, and ABA treatments. These GUS transgenic seedlings were analyzed histochemically after treatment with water, 400 mM mannitol, or 100 µM ABA for 12 h. This experiment was repeated twice. Each GUS staining assay was analyzed from five individual seedlings per experiment, and similar results were obtained.
To determine whether four copies of CaMV35S enhancer elements in the activation-tagging vector might affect the neighboring locus close to the T-DNA insertion site, the expression of neighboring genes was determined by RT–PCR. Expression of At4g18880 and At4g18900 in pca41 did not differ markedly from those of WT and atrzf1, whereas the expression of At4g18890 in pca41 was abolished completely (Fig. 3D). Interestingly, the expression of At4g18890 was lower in atrzf1 than that in the WT (Fig. 3D). However, the expression of At4g18890 was notably higher in an AtRZF1-overexpressing transgenic line (Ju et al., 2013) compared with that in WT or atrzf1 (Supplementary Fig. S2A), implying that AtRZF1 could regulate the expression of At4g18890. Thus, it is quite likely that the loss of expression of At4g18890 is closely related to the pca41 phenotype under ABA and osmotic stress conditions.
On the basis of amino acid sequence alignments, At4g18890 belongs to the BES1/BZR1 homolog gene family in Arabidopsis (Supplementary Fig. S2B). At4g18890, also known as BES1/BZR1 homolog 3 (BEH3), was predicted to be a transcription factor that could regulate BR signaling responses (Wang et al., 2002). The full-length cDNA of BEH3 encodes a protein of 284 amino acids (aa) containing a BES1 N-terminal active Domain (BES1-ND, 31 – 104 aa; Fig. S2C). BEH3 shares more than 84% identity with BES1/BZR1 homologs in Arabidopsis. A phylogenetic tree showing the distance among the sequence groups was constructed using cluster algorithms (Fig. S2D).
To investigate the function of BEH3, its expression pattern was initially assessed by qPCR. As shown in Fig. 3E, BEH3 was expressed strongly in roots, while it showed moderate expression in flowers and leaves, and low expression in stems. Moreover, we determined BEH3 expression in tissues through histochemical β-glucuronidase (GUS) staining of Arabidopsis transgenic lines harboring a 1939-bp BEH3 promoter (PRO)-GUS fusion construct. Analysis of BEH3 PRO-GUS transgenic lines revealed GUS activity in roots, cotyledons, developing rosette leaves, and flowers, particularly in anthers and veins of petals (Supplementary Fig. S3A–C). In siliques, GUS staining was not detected in seeds. However, it was detected in the lower part of the silique (Fig. S3D).
To assess sub-cellular localization of BEH3, its accumulation was determined by examining transgenic plants that carried a transgene encoding a translational fusion between BEH3 and green fluorescent protein (GFP). The fluorescence of BEH3-GFP was mainly detected in the nuclei of transgenic root cells (Fig. 3F). In control transgenic seedlings expressing GFP driven by the 35S promoter (GFP vector), the GFP fluorescence was detected in the cytosol (Fig. 3F). These results showed that BEH3 is a nuclear-localized protein.
To further analyze BEH3 expression in Arabidopsis in response to osmotic stress and ABA, qPCR and GUS staining assays were performed. As shown in Figs. 3G and 3H, BEH3 expression reached its maximum within 12 h or 6 h after mannitol or ABA treatment, respectively. Its expression then declined for the remaining duration of these treatments. The abiotic stress-inducible Responsive to ABA 18 (RAB18) gene served as a reference for mannitol and ABA treatments. GUS activity was also strongly induced in roots and rosette leaves of BEH3 PRO-GUS transgenic lines during mannitol or ABA treatment (Fig. 3I). Taken together, these results indicate that BEH3 expression is regulated by osmotic stress and ABA.
Loss of BEH3 improves the osmotic stress- and ABA-insensitive trait of atrzf1
To confirm whether a loss of BEH3 was responsible for the pca41 phenotype during dehydration and ABA treatment, we performed BEH3 RNA interference (ri), or overexpressed BEH3 using a BEH3-overexpressing (OX) construct under the control of a CaMV 35S promoter into atrzf1 or pca41 mutant, respectively. For each construct, we obtained 15 homozygous lines (T3 generation). Of these, three individual atrzf1/beh3-ri atrzf1/BEH3 RNAi construct complementary lines (atrzf1/ri2-4, atrzf1/ri4-7, and atrzf1/ri6-8, the BEH3 RNAi lines in atrzf1 background) exhibited reduced expression of BEH3, while three individual pca41/BEH3-OX construct complementary lines (pca41/OX1-3, pca41/OX5-2, and pca41/OX7-4) exhibited high expression of BEH3. They were selected for phenotype characterization (Supplementary Fig. S4A). Under normal conditions, there was no difference in seed germination or early seedling growth among WT, atrzf1, pca41, three atrzf1/beh3-ri, and three pca41/BEH3-OX complementary plants (Fig. S4B). Comparison of pca41/BEH3-OX lines with pca41 mutant showed markedly fewer cotyledons expanding and greening at 12 d or 10 d after seed germination, when grown on MS plates containing 400 mM mannitol or 0.7 μM ABA, respectively (Fig. 4A–C). On the contrary, the rate of cotyledon greening was similar between atrzf1/beh3-ri line and pca41 mutant, whereas atrzf1/BEH3 RNAi lines were slightly more insensitive to cotyledon greening than atrzf1 mutant (Fig. 4A–C). These data indicate that loss of BEH3 in pca41 mutant is able to improve the osmotic stress- and ABA-insensitive trait of atrzf1 mutant. Thus, atrzf1/BEH3 RNAi complementary lines are similar to the pca41 phenotype under osmotic stress and ABA conditions.
BEH3 negatively regulates dehydration and ABA responses. (A-F) Seeds of WT, atrzf1, pca41, atrzf1/BEH3 RNAi (atrzf1/ri2-4, atrzf1/ri4-7, atrzf1/ri6-8), pca41/BEH3-overexpressing (pca41/OX1-3, pca41/OX5-2, pca41/OX7-4), BEH3 RNAi (ri8-7, ri9-1), beh3, and BEH3-overexpressing (OE3-8, OE5-3) plants were sown in MS medium supplemented with 400 mM mannitol or 0.7 μM ABA and allowed to grow for 12 d or 10 d, respectively. (A and B) Photographs showing that pca41 and atrzf1/BEH3 RNAi lines have better growth than WT, atrzf1, and pca41/BEH3-overexpressing plants under osmotic stress (A) and ABA (B) conditions. (C) Effects of osmotic stress and ABA on cotyledon greening in complementary lines. Seeds were sown on MS plates supplemented with or without mannitol and ABA. Seedlings with green cotyledons were counted. (D and E) Photographs showing that beh3 and BEH3 RNAi lines have better growth than WT and BEH3-overexpressing plants under osmotic stress (D) and ABA (E) conditions. (F) Effects of osmotic stress and ABA on cotyledon greening in BEH3 transgenic lines. Seeds were sown on MS plates supplemented with or without mannitol and ABA. Seedlings with green cotyledons were counted. For (C) and (F), data represent mean and error bars from three different replicates; each replicate contained at least 50 seeds per sample. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05).
Overexpression of BEH3 confers sensitivity to osmotic stress and ABA
To further determine the physiological role of BEH3, we chose two independent homozygous transgenic BEH3-overexpressing (OE3-8 and OE5-3) and BEH3 RNAi (ri8-7 and ri9-1) lines and tested their phenotypes in response to 400 mM mannitol or 0.7 μM ABA. BEH3 expression was evaluated by qPCR. OE3-8 and OE5-3 lines displayed high amounts of BEH3 transgene expression compared with WT plants, while ri8-7 and ri9-1 lines exhibited low levels of BEH3 expression (Supplementary Fig. S5A). To determine whether the BEH3 RNAi construct influences the expression of the BEH3 homologs (Fig. S2D) BEH1, BEH2, BEH4, BES1, and BZR1, in ri8-7 and ri9-1 lines, we performed qPCR analysis. The analysis revealed that expression of the BEH3 RNAi construct does not inhibit the expression of BEH1, BEH2, BEH4, BES1, and BZR1, indicating that the BEH3 RNAi construct is specific to its own target transcript (Fig. S5B).
Additionally, to isolate beh3 single homozygous mutant from pca41 (atrzf1/beh3) mutant, pca41 was crossed with WT. Responses of F1 and F2 progenies to Basta during seed germination were analyzed. The F1 progeny exhibited resistance to Basta. Seed germination in F2 progeny displayed a segregation ratio of 3:1 of Basta resistance to Basta-sensitive germinating seeds (χ 23:1=0.022, P<0.05; Table S2). We obtained beh3 homozygous mutant lines in F2 progeny. A detailed description of genotypes of plants obtained in the cross is presented in Table S3. Finally, a beh3 single homozygous mutant was isolated based on genomic DNA PCR using beh3-specific and atrzf1-specific T-DNA insertion site primers (Supplementary Tables S1, S3). The absence of BEH3 transcript was verified by qPCR (Fig. S5A). To analyze the expression of AtRZF1 in line with the expression of BEH3, we extracted total RNA from WT, beh3, BEH3 RNAi (ri8-7), and BEH3-overexpresssing (OE3-8) plants. As shown in Fig. S5C, the expression of AtRZF1 was significantly lower (P<0.05, ANOVA) in beh3 and ri8-7 lines compared with the WT. However, the expression of AtRZF1 was markedly higher in OE3-8 transgenic line compared with WT, beh3, and ri8-7 plants (Fig. S5C), implying that BEH3 could regulate AtRZF1 gene expression. Thus, transcriptional regulation of AtRZF1 and BEH3 were interactively controlled by each other.
To investigate the effects of BEH3 expression on cotyledon greening under osmotic and ABA stress, seeds of WT, beh3, and BEH3 transgenic plants were germinated on MS plates supplemented with 400 mM mannitol or 0.7 μM ABA. They were then permitted to grow for 12 d or 10 d, respectively. Without treatment, no difference appeared between WT, beh3, BEH3 RNAi (ri8-7 and ri9-1), and BEH3-overexpressing (OE3-8 and OE5-3) plants (Supplementary Fig. S5D). Comparison of beh3, WT and BEH3-overexpressing transgenic plants revealed less greening of cotyledons when plants were grown on MS plates containing 400 mM mannitol or 0.7 μM ABA (Fig. 4D–F). On the contrary, the rate of cotyledon greening was similar between beh3 and beh3-ri lines (Fig. 4D–F). These results indicate that BEH3 is necessary to control early seedling growth under osmotic stress and ABA treatment.
We further investigated the dehydration insensitive response of beh3 and beh3-ri lines by analyzing Pro and MDA content in WT, beh3, and BEH3-OE plants. As shown in Fig. S5E, F, Pro content was higher in beh3 and beh3-ri lines than in WT and BEH3-OE plants under osmotic stress conditions. However, MDA content was lower in beh3 and beh3-ri lines than in WT and BEH3-overexpressing plants. These results indicated that beh3 mutant displayed increased insensitivity to osmotic stress. Therefore, BEH3 might play a role in activating lipid peroxidation-driven MDA formation, under osmotic stress.
BEH3 regulates hypersensitivity to osmotic stress by stimulating the oxidative signaling pathway
Previous reports have suggested that ROS can serve as vital messengers in various biological processes involved in abiotic stress defense (Mittler et al., 2011; Schippers et al., 2012). Besides, BR signaling can regulate ROS production to confer abiotic stress tolerance (Choudhary et al., 2012). It has been reported that ROS lead to lipid peroxidation through increased levels of MDA (Zoelle et al., 2012). Since BEH3 regulates lipid peroxidation in response to osmotic stress (Supplementary Fig. S5F), detection and quantification of H2O2 in WT, atrzf1, pca41, beh3, and BEH3-OE seedlings were performed after treatment with 400 mM mannitol using 3,3’-diaminobenzidine (DAB) staining. As shown in Fig. 5A, B, control-treated seedlings all exhibited much weaker DAB staining than seedlings after mannitol treatment. Under osmotic stress conditions, although there was an increase in H2O2 accumulation compared with controls in atrzf1, pca41 and beh3 whole seedlings, there was no significant variation (P<0.05, ANOVA) between them; however, H2O2 content was increased further in WT and BEH3-OE whole seedlings, compared with controls (Fig. 5A, B). These results demonstrate that BEH3 and AtRZF1 are required for regulating H2O2 production in early seedling stages under abiotic stress.
Analysis of osmotic stress-induced H2O2 content and antioxidant enzyme activity in WT, atrzf1, pca41, beh3, and BEH3-overexpressing seedlings. (A) DAB staining for H2O2 accumulation in each 14 day-old whole seedling untreated or treated with 400 mM mannitol for 6 h. Scale bars=20 mm. (B) Analysis of H2O2 accumulation with 14 day-old whole seedlings in the absence or presence of 400 mM mannitol. Data represent mean and error bars from three different replicates; each replicate contained at least 15 individual seedlings per sample. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05). (C-E) Activities of antioxidant enzymes in each seedling. APX (C), CAT (D), and guaiacol POX (E) activities were measured in detached rosette leaves of each four week-old line after treatment without or with 400 mM mannitol. Data represent mean and error bars from three different replicates; each replicate contained at least 10 individual leaves per sample. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05).
Subsequently, to further analyze the regulation of antioxidant enzymes by osmotic stress, ascorbate peroxidase (APX), catalase (CAT), and guaiacol peroxidase (POX) activities in WT, atrzf1, pca41, beh3, and BEH3-OE plants treated with or without mannitol were investigated. As shown in Fig. 5C–E, there was no significant difference (P<0.05, ANOVA) in APX, CAT, or POX activities between WT, atrzf1, pca41, beh3, and BEH3-OE plants after control treatment. However, APX, CAT, and POX activities in all plant samples were induced in response to mannitol treatment. Overall, the amount of osmotic stress-induced activities of these antioxidant enzymes were much higher in atrzf1, pca41, and beh3 mutants than in WT and BEH3-OE plants (Fig. 5C–E). In addition, APX and CAT activities were higher in pca41 than in atrzf1 and beh3 mutants under osmotic stress conditions (Fig. 5C–E). These data suggest strongly that BEH3 and AtRZF1 can repress activities of these antioxidant enzymes under osmotic stress. Therefore, transcriptional co-regulation between BEH3 and AtRZF1 might play an important role in drought, ABA, and osmotic stress responses through a ROS-mediated signaling pathway.
BEH3 and AtRZF1 negatively regulate hypocotyl elongation by modulating BR concentrations
BEH3 has been identified as a regulatory factor in BR signaling (Wang et al., 2002). However, the mechanism of regulation of BEH3 in BR signaling responses is still unclear. To determine the effects of BEH3 expression in signaling mediated via BR that is involved in the photomorphogenic processes, seeds of WT, atrzf1, pca41, beh3 and BEH3-OE plants were germinated on MS plates under light or dark conditions. The hypocotyl length of each seedling was measured 5 d or 3 d after germinating in light- or dark-grown conditions, respectively. Under light-grown conditions, the hypocotyl lengths were not significantly different (P<0.05, ANOVA) between WT, atrzf1, pca41, beh3, and BEH3-OE seedlings grown on MS plates (Supplementary Fig. S6A, B). Furthermore, the shoot phenotypes, including leaf shape, petiole length, leaf size, and bolting time were not markedly different between WT, beh3, and BEH3-OE plants grown in soil (Fig. S7). These observations suggest that BEH3 may not fulfill specific functions in BR signaling in response to light conditions. However, hypocotyl lengths of atrzf1, pca41, and beh3 mutants in dark-grown seedlings were much longer than WT and BEH3-OE seedlings (Fig. 6A, B). In addition, hypocotyl lengths of BEH3-OE seedlings were much shorter than that of WT, whereas pca41 mutant had the longest hypocotyl length in dark-grown conditions, implying that BEH3 could negatively regulate hypocotyl elongation in the dark (Fig. 6A, B). To further confirm whether a lack of BEH3 would cause the long hypocotyl phenotype during darkness, a second set of experiments was conducted to assay hypocotyl length in BEH3 RNAi (ri8-7 and ri9-1) lines (Supplementary Fig. S8A, B). Our data revealed that hypocotyl lengths of ri8-7 and ri9-1 lines in dark-grown seedlings were longer than WT and BEH3-OE seedlings, confirming that BEH3 and AtRZF1 (Fig. 6A, B) negatively regulate hypocotyl elongation in the dark.
beh3 and atrzf1 regulate hypocotyl elongation of dark-grown seedling by modulating BR amounts. Seeds of WT, atrzf1, pca41, beh3, and BEH3-overexpressing plants were sown in MS medium supplemented with and without eBL or BRZ by dark-grown vertical-oriented assay. (A, B) Hypocotyl elongation assay with three day-old dark-grown seedlings. (C, D) Hypocotyl elongation assay with three day-old dark-grown seedlings in the presence of 0.5 nM eBL. (E, F) Hypocotyl elongation assay with three day-old dark-grown seedlings in the presence of 1 μM BRZ. Data represent mean and error bars from three different replicates; each replicate contained at least 50 individual seedlings per sample. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05). Scale bars=0.5 cm. (G and H) Expression of hypocotyl elongation-regulated genes in three day-old dark-grown seedlings. ARF4 (G) and ARF8 (H) expression was analyzed by qPCR. Actin 1 was used as internal control. Data represent mean and error bars from three different replicates; each qPCR experiment was conducted with total RNA of each sample obtained from ten individual seedlings pooled. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05).
Based on these results, we hypothesized that decreased endogenous BR concentrations in the BEH3-OE transgenic line could lead to short hypocotyls in darkness. To test this hypothesis we exogenously applied 24-epibrassinolide (eBL) to BEH3-OE seedlings, and observed whether the short hypocotyl phenotype in darkness could be rescued. We analyzed hypocotyl lengths of WT, atrzf1, pca41, beh3, and BEH3-OE seedlings in the presence of 0.5 nM eBL. Under darkness, hypocotyl lengths of BEH3-OE and WT seedlings were very similar to each other under eBL, although they were shorter than those of atrzf1, pca41, and beh3 mutants in response to eBL (Fig. 6C, D). Furthermore, hypocotyl lengths of atrzf1, pca41 and beh3 mutants were not significantly different (P<0.05, ANOVA) (Fig. 6C, D). Therefore, eBL was able to slightly reverse BR-induced hypocotyl elongation defects of the BEH3-OE transgenic line in darkness.
We then observed hypocotyl elongation of dark-grown seedlings in the presence of brassinazole (BRZ), a BR-biosynthesis inhibitor. BEH3-OE seedlings treated with 1 μM BRZ showed slightly shorter hypocotyl lengths than WT seedlings (Figs. 6E, F), suggesting that BEH3-OE transgenic lines displayed relatively weak BR-deficient phenotype. In contrast, atrzf1, pca41, and beh3 mutants exhibited enhanced hypocotyl elongation compared with WT when seedlings were treated with 1 μM BRZ (Fig. 6E, F). These genetic experiments suggest that both BEH3 and AtRZF1 can negatively regulate hypocotyl elongation of dark-grown seedlings by modulating BR amounts.
To further define whether BEH3 and AtRZF1 could be related to the regulation of hypocotyl elongation in darkness, the expression of Auxin Response Factor 4 (ARF4) or ARF8 genes, related to hypocotyl elongation (Peres et al., 2019), was assessed in dark-grown seedlings using qPCR. The results showed that accumulation of ARF4 and ARF8 mRNA was significantly increased 1.4- to 2.5-fold, and 1.4- to 1.9-fold (P<0.05, ANOVA), respectively, in the BEH3-OE line, compared with that in WT, atrzf1, pca41, or beh3 plants, whereas expression of ARF4 and ARF8 was lower in atrzf1, pca41, and beh3 mutants than in WT (Fig. 6G, H). Expression of these two genes were lower in pca41 and beh3 than in atrzf1 mutant (Fig. 6G, H). These results indicate that both BEH3 and AtRZF1 can regulate expression of these hypocotyl elongation-related genes under darkness conditions. Collectively, the data shown in Fig. 6 support the hypothesis that coordination between BEH3 and AtRZF1 plays a crucial role in regulating hypocotyl elongation through BR signaling under darkness.
BEH3 and AtRZF1 regulate the expression of BR-related genes under osmotic stress
It has been relatively well reported that BR-metabolic and -responsive genes are effectors of abiotic stress signals that regulate plant growth and development (Sharma et al., 2017; Anwar et al., 2018). To check whether there are different sensitivities of atrzf1, pca41, beh3 and BEH3-OE plants in cotyledon greening experiments in response to osmotic stress, brassinosteroid-6-oxidase 2 (BR6OX2), constitutive photomorphogenic dwarf (CPD), dwarf 4 (DWF4) and phyb-4 activation-tagged suppressor 1 (BAS1) genes (Szekeres et al., 1996; Divi and Krishna, 2010; Wang et al., 2012; Divi et al., 2016) were chosen to monitor endogenous BR concentrations in WT, atrzf1, pca41, beh3 and BEH3-OE seedlings. Expression of BR biosynthetic BR6OX2, CPD, and DWF4, or BR-catabolic BAS1 gene in WT, atrzf1, pca41, beh3, and BEH3-OE seedlings under normal conditions were very similar (Fig. 7A–D). Under osmotic stress, expression of BR6OX2, CPD, and DWF4 was higher in atrzf1, pca41, and beh3 mutants than in WT and BEH3-OE plants, whereas expression of these three genes were not much different between WT and BEH3-OE plants (Fig. 7A–D). In contrast, the expression of BAS1, a BR catabolic gene, was lower in atrzf1, pca41, and beh3 mutants than that in WT or BEH3-OE plants (Fig. 7D). In addition, the accumulation of BAS1 transcript was higher in BEH3-OE than in WT under osmotic stress (Fig. 7D). These results demonstrate that BEH3 and AtRZF1 are required for BR regulation of osmotic stress-induced sensitivity in Arabidopsis.
Responses of cotyledon greening to BR and expression of BR-related genes in WT, atrzf1, pca41, beh3, and BEH3-overexpressing plants under osmotic stress conditions. (A-F) qPCR analysis of expression of BR6OX2 (A), CPD (B), DWF4 (C), BAS1 (D), SAUR9 (E), and TCH4 (F) involved in osmotic stress response. Total RNA samples obtained from two week-old seedlings were treated without or with 400 mM mannitol for 6 h. Data represent mean and error bars from three different replicates; each qPCR experiment was conducted with total RNA of each sample obtained from ten individual seedlings pool. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05). (G and H) Response of cotyledon greening to eBL or BRZ treatment under osmotic stress conditions. Seeds of WT, atrzf1, pca41, beh3, and BEH3-overexpressing (OE) plants were grown in MS medium with 0.1 μM eBL or 2 μM BRZ in the presence of 400 mM mannitol and permitted to grow for 12 d. (G) Photographs showing that atrzf1, pca41, and beh3 mutants have better growth than WT and BEH3-OE plants in presence of 0.1 μM eBL or 2 μM BRZ under osmotic stress conditions. (H) Effects of BR or BRZ on cotyledon greening in seedlings grown under osmotic stress conditions. Seedlings with green cotyledons were counted. Data represent mean and error bars from three different replicates; each replicate contained at least 50 individual seeds per sample. Different letters above bars indicate statistically significant difference (ANOVA, P<0.05).
Additionally, to investigate the BR-responsive genes associated with osmotic stress, we assessed two genes, small auxin up RNA 9 (SAUR9) and touch 4 (TCH4), both of which are regulated by various abiotic stresses (Jain and Khurana, 2009; Ren and Gray, 2015; Xie et al., 2019). The expression of SAUR9 and TCH4 was significantly higher (1.3-1.7 fold and 1.4-1.9-fold higher, respectively, P<0.05, ANOVA) in atrzf1, pca41, and beh3 than in WT and BEH3-OE plants under osmotic stress, whereas expression of TCH4 was significantly lower (0.8-fold lower, P<0.05, ANOVA) in BEH3-OE than in WT plants (Fig. 7E, F). These observations indicate that AtRZF1 and BEH3 regulate the expression of these BR-inducible marker genes under osmotic stress conditions.
Response of cotyledon greening to BR under osmotic stress
Since the expression of BR biosynthesis genes was lower in BEH3-OE plants compared with those in beh3 mutant under osmotic stress conditions (Fig. 7A–C), we further examined if the mannitol-induced sensitive phenotype of BEH3-OE could be rescued by exogenous BR application. In the presence of exogenous eBL (0.1 μM), the rate of cotyledon greening increased in WT and BEH3-OE plants under osmotic stress conditions (Fig. 7G, H), and the relative induction in the cotyledon greening rate of BEH3-OE line was lower in WT plants (Fig. 7H). When eBL was applied to mutants, cotyledon greening rates of atrzf1, pca41 and beh3 were not significantly different (P<0.05, ANOVA) under osmotic stress conditions (Fig. 7H). Fundamentally, this analysis suggests that mannitol-induced sensitive phenotype of BEH3-OE line is reversed by exogenous BR application.
We next investigated the role of BEH3 in regulating cotyledon greening using a BR biosynthesis inhibitor, BRZ (2 μM), under osmotic stress conditions. Under osmotic stress, the atrzf1, pca41, and beh3 mutants were less sensitive to BRZ treatment than WT and BEH3-OE seedlings, based on the extent of cotyledon greening (Fig. 7G, H). Besides, BEH3-OE lines treated with BRZ under osmotic stress showed less cotyledon greening than WT seedlings (Fig. 7H). From this experiment, we concluded that BEH3 is involved in BR-regulated cotyledon greening during osmotic stress.
Discussion
Many plants under abiotic stress accumulate compatible solutes, including proline (Pro), fructans, glycine betaine, inositol, and pinitol, in the cytoplasm of their cells (Slama et al., 2015). Pro is an important signal molecule in plant responses to water deficit stress. It has many protective functions, including scavenging ROS, regulating antioxidant production, and protecting cellular membrane structure (Szabados and Savouré, 2010). In addition, ABA and BR also regulate the accumulation of Pro to confer abiotic stress tolerance (Sharma et al., 2017; Sharma et al., 2019).
Previously, we have reported that a mutation in AtRZF1, which encodes an ubiquitin E3 ligase, results in a drought tolerance phenotype by increasing Pro content (Ju et al., 2013). To isolate novel components involved in Pro regulatory metabolism in Arabidopsis, we used the atrzf1 mutant as a parental line for T-DNA mutagenesis (Kim et al., 2017). In this study, we isolated an atrzf1 mutant, pca41. It displayed an enhanced insensitive phenotype to drought, as well as to ABA and osmotic stress, compared with the atrzf1 mutant (Figs. 1, 2). Moreover, analysis of physiological parameters of pca41 under drought, ABA, or osmotic stress revealed higher Pro accumulation with higher relative water content and fresh weight, but lower MDA content than those in WT and atrzf1 (Figs. 1, 2). As shown in Fig. S1, expression of RD29A, RD29B, AtMYB2, P5CS1, and P5CR, known to regulate abiotic stress responses, was higher in pca41 than in WT and atrzf1. These data indicated that pca41 could enhance the insensitivity of atrzf1 to drought, ABA, and osmotic stress during early seedling growth.
A single T-DNA inserted in At4g18890 of pca41 genomic DNA, was identified (Fig. 3A–D). The expression of At4g18890 was completely abolished in pca41 mutant. At4g18890 encodes BEH3 (Supplementary Fig. S2B, C), which is a BR signaling regulatory transcription factor that can regulate various environmental responses, plant growth and development (Wang et al., 2002; Yin et al., 2005; Wang and Wang, 2018). Furthermore, osmotic stress or ABA-induced phenotypes of atrzf1/beh3-ri and pca41/BEH3-OX complementary lines were similar to those of pca41 and WT, respectively (Fig. 4A–C, Fig. S4). Thus, the BEH3 mutation was tightly connected to the insensitive phenotype of pca41 during abiotic stress. Interestingly, the expression of BEH3 was much lower in atrzf1, but higher in AtRZF1-overexpressing (AtRZF1-OE) transgenic plants (Fig. S2A), indicating that BEH3 might be regulated by a ubiquitination-mediated pathway. Recently, several ubiquitin E3 ligase-substrate interactions in the BR pathway have been reported. One study has demonstrated that BES1 can be ubiquitinated by E3 ligase SINAT (for SINA of Arabidopsis thaliana) to degrade BES1 under light and starvation stress conditions (Yang et al., 2017).
As shown in Fig. 3E and Fig. S3, BEH3 was highly expressed in flowers and roots, but not in stems or leaves. It was strongly induced by osmotic stress or ABA (Fig. 3G–I). These results indicate that BEH3 is regulated by various developmental and abiotic stress signals. As shown in Fig. 3F, BEH3 mainly localizes in the nucleus, which coincides with its predicted function as a transcription factor that regulates the BR signaling response (Wang et al., 2002).
Based on physiological experiments such as ABA- or osmotic stress-induced sensitivity, Pro accumulation and MDA content assays showed that beh3 mutant and BEH3 RNAi lines were more insensitive than BEH3-OE transgenic lines to ABA or osmotic stress, implying that BEH3 could negatively regulate ABA and osmotic stress responses during early seedling growth (Fig. S5).
Yin et al. (2005) have demonstrated that BEH3 and its homologs can bind to the E-box (5’-CANNTG-3’) element of the promoter region of BR-related genes. Previous studies have reported that the E-box cis-element is required to regulate ABA-, drought-, osmotic stress-, salt stress-, and cold-responsive genes (Liu et al., 2014; Zang et al., 2019). We found that five copies of E-box elements existed in the upstream region (between positions -489 and -1260 nt) of AtRZF1, suggesting that AtRZF1 might regulate ABA, osmotic stress, and BR responses in a BEH3-dependent manner. To test this possibility, accumulation of AtRZF1 mRNA was monitored in beh3 and BEH3 transgenic plants. As shown in Fig. S5C, expression of AtRZF1 was decreased in beh3 and BEH3 RNAi lines compared with those in WT and BEH3-OE plants, whereas AtRZF1 expression was induced more in BEH3-OE line than in WT plants. This result indicates that BEH3 binds to the E-box element binding transcriptional activator which regulates the expression of abiotic stress- or BR-responsive AtRZF1. Based on results shown in Figs. 3D, S2A, and S5C, coordinated regulation between BEH3 and AtRZF1 genes might play an important role in abiotic stress- and BR-mediated signaling pathways.
It is well known that ROS serve as secondary messengers in different biological processes involved in abiotic stress responses (Schippers et al., 2012). In addition, Pro and BR signaling can regulate ROS accumulation that confers abiotic stress tolerance (Szabados and Savouré, 2010; Choudhary et al., 2012). Herein, H2O2 content in BEH3-OE lines were higher than those in atrzf1, pca41, and beh3 mutants, whereas antioxidant enzyme activities were reduced by increasing BEH3 expression, in response to osmotic stress (Fig. 5). These observations suggest that BEH3 might play a role in ROS homeostasis during the osmotic stress response.
Recently, Lachowiec et al. (2018) have shown that dark-grown, 7-d-old beh3-1 mutant produces markedly shorter hypocotyls than WT; however, hypocotyl length was not affected in light-grown seedlings. Thus, it was of interest to determine hypocotyl elongation of dark-grown beh3 seedlings isolated from atrzf1/beh3 double (pca41) mutant. Under light-grown conditions, hypocotyl length and shoot phenotypes were not different among WT, atrzf1, pca41, beh3, and BEH3-OE seedlings (Supplementary Figs. S6, S7), indicating that BEH3 and AtRZF1 were not required for the regulation of hypocotyl elongation in light conditions. For dark-grown seedlings, hypocotyl lengths of atrzf1, pca41, and beh3 were much longer than either WT or BEH3-OE seedlings, whereas BEH3-OE lines displayed much shorter hypocotyl lengths than WT (Fig. 6A, B), implying that BEH3 and AtRZF1 negatively regulated hypocotyl elongation in dark growth. These observations are different from those of Lachowiec et al. (2018). It is possible that these differences are because the beh3-1 mutant used by Lachowiec et al. (2018) has a different mutation site of BEH3 compared with our beh3 mutant. Moreover, BEH3-OE lines exhibited relatively higher expression of hypocotyl elongation-regulatory genes (ARF4 and ARF8) compared with other samples. However, hypocotyl lengths of BR-treated BEH3-OE lines recovered to the phenotype of WT in darkness (Fig. 6C, D, G, H). These physiological and genetic data show that both BEH3 and AtRZF1 are negative components in regulating hypocotyl elongation of dark-grown seedlings by modulating BR concentrations.
Even though BES1/BZR1 family members function redundantly and indispensably in the BR response (Chen et al., 2019), their biological functions may not be fully redundant. Previously, bes1-D, a gain-of-function mutant, was found to be more sensitive to drought stress than WT (Ye et al., 2017). In the present study, BEH3-overexpressing seedlings displayed enhanced sensitivity to ABA and osmotic stress responses in comparison with WT seedlings (Fig. 4D–F). Recently, Sun et al. (2020) reported that heterologous expression of maize ZmBES1/BZR1-5 in transgenic Arabidopsis resulted in decreased salt-, drought-, and ABA-sensitivity, indicating that it positively regulates the abiotic stress and ABA responses. In addition, BZR1 positively regulates plant stress tolerance (Li et al., 2017; Yin et al., 2018). These data suggest that each BES1/BZR1 family member may fulfill specific functions in response to abiotic stresses.
Although BRs have been functionally implicated in diverse abiotic stress responses, their exact abiotic stress mechanism has not been clarified yet (Sharma et al., 2017). As shown in Fig. 7A–F, all samples tested in this study treated with osmotic stress showed reduced expression of BR biosynthesis (BR6OX2, CPD, DWF4) and BR-responsive (SAUR9, TCH4) genes, than untreated samples. Nevertheless, expression of these BR biosynthesis and -responsive genes, with the exception of BR-catabolic BAS1, were increased in osmotic stress-tolerant atrzf1, pca41, and beh3 mutants, than in osmotic stress-sensitive BEH3-OE lines, after osmotic stress treatment. Furthermore, osmotic stress-induced sensitive phenotype of BEH3-OE line was reversed by exogenous BR application (Fig. 7G, H). These observations demonstrate that BEH3 and AtRZF1 alone or together could regulate the activity or inactivity of BR signaling-related molecules by mediating endogenous BR concentrations in an osmotic stress-triggered defective plant growth response.
In conclusion, the present study characterized an enhanced drought tolerance phenotype of the pca41 mutant. Based on Figs. S2A and S5C, coordinated regulation between AtRZF1 and BEH3 expression is important for Pro and ROS metabolism in response to osmotic stress. In darkness, BEH3 negatively regulated hypocotyl elongation by increasing the expression of ARF4 and ARF8 genes, or suppressing the expression of BR biosynthesis genes. BEH3 repressed physiological parameters including Pro content, survival rate, ROS accumulation, and activity of antioxidant enzymes under abiotic stress conditions. Thus, co-regulation between BEH3 and AtRZF1 expression negatively mediates water deficit stress response through their BR signaling and ubiquitination action. Further functional studies investigating BEH3 and AtRZF1, especially their relationship in the BR signaling pathway, are needed to elucidate the drought response networks in plants.
Supplementary data
The following supplementary data are available at JXB online.
Table S1. Primers used for TAIL-PCR, qPCR, RT–PCR, and gene cloning.
Table S2. Genetic analysis of pca41 performed in the presence of Basta.
Table S3. Detailed description of genotypes and phenotypes in the cross of WT × pca41.
Fig. S1. Expression of osmotic stress-induced genes in WT, atrzf1, and pca41 plants.
Fig. S2. Alignment of full-length deduced amino acid sequences of BEH3 and Arabidopsis homologues.
Fig. S3. BEH3 promoter-GUS expression in transgenic Arabidopsis lines.
Fig. S4. Phenotype of complementary transgenic lines in light-grown normal conditions.
Fig. S5. Comparative analysis of osmotic stress-induced Pro and MDA accumulation in WT, beh3, BEH3 RNAi, and BEH3-overexpressing plants.
Fig. S6. Phenotypes of WT, atrzf1, pca41, beh3, and BEH3-overexpressing plants in light-grown conditions.
Fig. S7. Phenotypes of WT, beh3, and BEH3-overexpressing plants during different growth stages under light-grown conditions.
Fig. S8. Hypocotyl elongation response of WT, BEH3 RNAi, and BEH3-overexpressing plants in dark-grown conditions.
Abbreviations
- AtRZF1
Arabidopsis thaliana Ring Zinc Finger 1
- BEH3
BES1/BZR1 Homolog 3
- GFP
green fluorescent protein
- PCA41
proline content alterative 41
- ROS
reactive oxygen species
- WT
wild-type
Acknowledgements
This work was supported by grants to C.S.K. from the Next-Generation BioGreen21 (SSAC, PJ013171) and New Breeding Technology program (PJ01477701) funded by the Rural Development Administration, Republic of Korea.
Author contributions
TV N and CSK designed the study and interpreted study results; KHL provided technical assistance with TAIL-PCR analysis; TVN, CRP, KHL and SL performed interpretation and analysis of experimental results; TVN and CSK wrote the manuscript. All authors approved the final manuscript.
Data availability
All data generated in this study are available within the article and its Supplementary data.
Comments