RAVL1, an upstream component of brassinosteroid signalling and biosynthesis, regulates ethylene signalling via activation of EIL1 in rice

Brassinosteroids (BRs) are steroid hormones that regulate several aspects of plant growth and development. Their signal is activated by the binding of BRs to the receptor brassinosteroid‐insensitive 1 (BRI1), which activates the transcription factors brassinazole‐resistant 1 (BZR1) and BRI1‐EMS‐SUPPRESSOR 1 (BES1) that regulate target gene expression (Kim and Wang, 2010). More recently, the crosstalk between BRs and other hormone signalling factors has been extensively investigated, and it has been suggested that BR signalling is highly integrated with other signals to regulate plant growth and development (Wang et al., 2012). We previously identified Related to ABI3/VP1RAV‐Like 1 (RAVL1), a transcriptional activator that modulates BR homeostasis by activating BRI1 and biosynthetic genes in rice (Je et al., 2010). However, RAVL1 is not modulated by BRs, suggesting a potential association between BRs and other hormones via RAVL1 in rice.

To investigate whether RAVL1 is involved in the crosstalk circuits between BRs and other hormones, the hormone‐dependent expression of RAVL1 was analysed. Seven‐day‐old wild‐type plants were subjected to various treatments using 2,4‐dichlorophenoxyacetic acid (2,4‐D), epibrassinolide (epi‐BL), abscisic acid (ABA), gibberellic acid (GA) and 1‐aminocyclopropane‐1‐carboxylic acid (ACC), and the RAVL1 level was monitored by quantitative real‐time PCR (qPCR). The results indicated that 2,4‐D (a synthetic auxin) treatment induced while ACC application suppressed the RAVL1 expression level, but the other hormones examined did not alter the RAVL1 level (Figure 1a). Analysis of the RAVL1 promoter sequences revealed an auxin‐responsive element located within 1 kb of these promoter sequences (data not shown), suggesting that auxin response factors (ARFs) might be involved in auxin induction of RAVL1. However, further analyses are required to determine whether ARFs bind to RAVL1 promoter sequences. To further investigate the effect of ethylene signalling on RAVL1, RAVL1 expression levels were examined in an ethylene‐insensitive 2 (EIN2) mutant, ethylene‐insensitive 3‐like 1 (EIL1) RNAi (eil1 Ri), ethylene response sensor 1 (ERS1) mutant and EIL1 overexpressor (EIL1 OX). The data indicate that RAVL1 was slightly higher in ein2 and eil1 Ri, while lower in ers1 and EIL1 OX compared with wild‐type plants (Figure 1b). As RAVL1 levels were suppressed by ACC, ACC‐dependent root growth was analysed in wild‐type, RAVL1 mutants (ravl1‐1 and ravl1‐2) and RAVL1 overexpression lines (RAVL1 OX and RAVL1‐GFP OX). Primary root growth was inhibited by ACC treatment in wild‐type plant, and the inhibition rate was lower in RAVL1 mutants and higher in RAVL1 overexpression lines than in wild‐type plants (Figure 1c,d), indicatingthat RAVL1 positively contribute to the ACC‐mediated root growth suppression.

Figure 1.

RAVL1 directly activates EIL1 and ERS1 regardless of brassinosteroids (BRs) signalling. (a) The expression patterns of Related to ABI3/VP1RAV‐Like 1 (RAVL1) were monitored by quantitative real‐time PCR (qPCR) performed on 7‐day‐old wild‐type (WT) seedlings treated with 1 μm of epibrassinolide (epi‐BL), 2,4‐dichlorophenoxyacetic acid (2,4‐D), abscisic acid (ABA), 1‐aminocyclopropane‐1‐carboxylic acid (ACC) and gibberellic acid (GA). RAVL1 expression patterns were evaluated after 0, 3, 6 and 9 h of treatment. Error bars represent ± standard error (SE) of three repeated experiments (n = 3). (b) RAVL1 level was examined in 7‐day‐old WT, ein2, eil1 Ri, ers1 and EIL1 OX plants. Ubiquitin was used as a control to normalize data. Error bars represent ± SE (n = 3). (c) Seven‐day‐old WT, ravl1‐1 mutant and RAVL1 overexpression (OX) plants analysed for 10 μm ACC‐dependent root growth were photographed. (d) ACC‐dependent root growth of wild‐type, ravl1‐1, ravl1‐2,RAVL1 OX and RAVL1‐green fluorescence protein (GFP) OX was calculated. Error bars indicate ± SE (n = 10). (e) Expression of EIL1,EIL2,EIN2,ERS1, ACO2 and ACO3 was examined in WT, ravl1‐1, ravl1‐2,RAVL1 OX and RAVL1‐GFP OX plants. Error bars indicate ± SE (n = 3). (f) Ethylene contents were measured from 7‐day‐old WT, ravl1‐1, ravl1‐2,RAVL1 OX and RAVL1‐GFP OX plants. Error bars indicate ± SE (n = 30). (g) Schematic diagram indicating E‐box location (red circle) within 1.5 kb EIL1 and ERS1 promoters, and probes (P) used for chromatin immunoprecipitation (ChIP) assays. Numbers (1–3) below E‐box elements indicate probes used in the electrophoretic mobility shift assay (EMSA). Relative ratios of immunoprecipitated DNA to input DNA were determined by qPCR. Input DNA was used to normalize the data. Pre‐immune serum: IgG, Ab: GFP antibody. Error bars indicate ± SE (n = 3). (h) EMSA was performed to evaluate RAVL1 affinities to each of the three E‐box elements (numbers 1–3) and mutated probes (m1–m3). (i) A yeast one‐hybrid assay was conducted to analyse the RAVL1 activation within 1.5 kb of EIL1 and ERS1 promoters. Yeast cells harbouring either AD‐RAVL1 and pEIL1‐His,AD‐RAVL1 and pERS1‐His, or E‐box‐mutated promoters mpEIL1‐His and mpERS1‐His, respectively, were grown on SD media lacking leucine (Leu) and uracil (Ura) or Leu, Ura and histone (His). (j) A transient expression assay was performed by co‐transfection with p35S:RAVL1 and each of the vectors expressing the beta‐glucuronidase gene (GUS) under the control of native and E‐box‐mutated EIL1 or ERS1 promoters in protoplast cells. The luciferase gene driven by the 35S promoter was used as an internal control to normalize GUS expression. Error bars indicate ± SE (n = 6). EIL1 and ERS1 expression levels were tested in d61‐1 (a BRI1 mutant) and RAVL1 OX genetic combinations (k) or d2 and RAVL1 OX genetic combination plants (l). Error bars indicate ± SE (n = 3). (m) BRI1 and D2 levels were monitored in eil1 and RAVL1 OX genetic combinations. Error bars indicate ± SE (n = 3). (n) RAVL1 OX, d61‐1/RAVL1 OX and eil1 Ri/RAVL1 OX 7‐day‐old plants were analysed for 1‐aminocyclopropane‐1‐carboxylic acid (ACC)‐dependent root growth (10 μm ACC‐treated and control plants were examined). (o) ACC‐dependent root growth inhibition rates of RAVL1 OX, d61‐1/RAVL1 OX and eil1 Ri/RAVL1 OX were calculated. Data are presented as means ± SE (n = 10). The ACC‐dependent expression patterns of BRI1 (p) and D2 (q) were evaluated after 0, 3, 6 and 9 h of treatment. Error bars indicate ± SE (n = 3). The relative expression rate was calculated against 0 h time point. All experiments were repeated at least three times. For all experiments, one‐way analysis of variance (ANOVA) was conducted, followed by Bonferroni's multiple comparison tests. Significant differences at P < 0.05 level are indicated by different lowercase letters.

The expression of EIL1, EIL2, EIN2, ERS1, ACC oxidase 2 (ACO2) and ACO3 was analysed in RAVL1 mutants and RAVL1 overexpression lines by qPCR. EIL1 and ERS1 levels were lower in mutants and higher in overexpression lines than in wild‐type plants. ACO2 and ACO3 levels were higher in mutants than in wild‐type and overexpression lines, while EIN2 and EIL2 levels did not change among the tested lines (Figure 1e). As the expression of ACO2 and ACO3 (two ethylene biosynthesis enzymes) was altered in ravl1‐1 mutants, ethylene content was measured. The results demonstrated that ravl1 mutants accumulated more ethylene, while RAVL1 overexpressors contained less ethylene compared with wild‐type plants (Figure 1f). As EIL1 and ERS1 are key ethylene signalling genes positively regulated by RAVL1, their promoter sequences were analysed to identify the presence of E‐box elements in the RAVL1‐binding region. Two and three putative E‐box elements were located within 1.5 kb of EIL1 and ERS1 promoters, respectively (Figure 1g). To test the binding affinity of RAVL1 to the E‐box elements, a chromatin immunoprecipitation assay was performed using 35S:GFP and 35S:RAVL1:GFP transgenic plant calli. A pre‐immune serum was used as the control for the GFP antibody to immunoprecipitate DNA. RAVL1 bound to the P2 and P3 regions of EIL1 and ERS1 promoters, respectively (Figure 1g). We further performed an electrophoretic mobility shift assay to determine which E‐box elements were responsible for the binding affinity to RAVL1, using six probes. We found that RAVL1 bound to E‐boxes in positions ‘1’, ‘2’ and ‘3’, but failed to bind their mutated probes m1, m2 and m3 (Figure 1h). The results of this binding assay were confirmed using a yeast one‐hybrid assay, which indicated that RAVL1 is only able to activate 1.5‐kb EIL1 and ERS1 promoters (pEIL1 and pERS1) if the E‐box promoters are not mutated at P2 and P3 regions of EIL1 and ERS1 promoters, respectively (mpEIL1 and mpERS1). In these mutated promoters, E‐box element sequences CANNTG were substituted by the sequence TTTTTT (Figure 1i). To verify whether these cis‐elements were responsible for the transcriptional activation of EIL1 and ERS1 promoters by RAVL1 in vivo, we performed transient expression assays using the protoplast system. Protoplast cells were co‐transformed with the 35S:RAVL1 plasmid and a vector expressing the beta‐glucuronidase gene (GUS) under the control of pEIL1 and pERS1 or mpEIL1 and mpERS1. Using 35S:Luciferase (LUC) as an internal control to normalize the transformation efficiency in each assay, protoplasts expressing RAVL1 had approximately twice the levels of activated pEIL1 and pERS1, but RAVL1 was unable to activate mpEIL1 and mpERS1 (Figure 1j). These results indicate that RAVL1 directly activates EIL1 and ERS1 via promoter binding. More interestingly, RAVL1 is able to directly activate BRI1 and ERS1, and directly regulates other key genes in BR biosynthesis (D2 and D11; Je et al., 2010) and ethylene signalling (EIL1). RAVL1 mutants are insensitive to BRs and ethylene, two important phytohormones, while RAVL1 overexpression lines are sensitive to both, which is in agreement with existing molecular data, because RAVL1 activate BR‐ and ethylene‐related genes. ERS1 is a negative regulator of ethylene signalling and function at the upstream of EIL1 (Yang et al., 2015a,b), and its mutant is sensitive to ethylene in rice (Ma et al., 2014). The positive regulation of RAVL1 on ethylene response might be caused by the activation of EIL1, which may cover the activation of ERS1 by RAVL1.

RAVL1 directly activates BRI1 and biosynthetic genes to maintain BR homeostasis (Je et al., 2010). We further examined whether RAVL1 activation of EIL1 and ERS1 depends on BR signalling using genetic combinations between RAVL1 OX and d61‐1, a mutant of BRI1, or between RAVL1 OX and d2, a BR biosynthetic gene mutant, and qPCR was performed. EIL1 and ERS1 expression levels were slightly lower in d61‐1 and higher in RAVL1 OX plants compared with wild‐type plants. However, the mutation of BRI1 in the RAVL1 OX background (d61‐1/RAVL1 OX) did not affect EIL1 and ERS1 activation by RAVL1 (Figure 1k). Similarly, EIL1 and ERS1 expression levels were lower in d2 compared with wild‐type plants; however, D2‐mutated plants showed no difference in the activation of EIL1 and ERS1 by RAVL1 (Figure 1l), suggesting that RAVL1 might activate EIL1 and ERS1 in BR signalling‐independent manner. Activation of BRI1 and D2 by RAVL1 was examined in eil1 Ri and RAVL1 OX genetic combinations. The data showed that the knock‐down of EIL1 did not affect the activation of BRI1 and D2 by RAVL1 (Figure 1m), indicating that RAVL1 might be also independent of ethylene signalling to activate BRI1 and D2. Ethylene sensitivity of RAVL1 OX, d61‐1/RAVL1 OX and eil1 Ri/RAVL1 OX was further examined, and the results indicated that RAVL1 OX and d61‐1/RAVL1 OX exhibited similar responses to exogenously applied ethylene, while eil1 Ri/RAVL1 OX reduced its sensitivity to ethylene (Figure 1n,o). These results suggest that RAVL1 is required in the ACC‐induced root growth suppression, which independents of BR signalling but requires ethylene signalling. However, EIL1 and ERS1 levels were slightly lower in d61‐1 and d2 mutants, suggesting that BR signalling might be upstream of ethylene signalling. In addition, treatment of an ethylene precursor ACC suppresses RALV1 expression, and the RAVL1 level was reduced in ers1 and EIL1 OX plants. Moreover, RAVL1 positively contributes to the ethylene signalling, implying that feedback regulation may occur between RAVL1 and ethylene signalling. Similarly, the suppression of RAVL1 by ethylene might be a result of the activation of ethylene signalling.

To analyse whether ethylene also influences BR‐related genes via RAVL1, ACC‐mediated expressions of BRI1 and D2 were further examined in 7‐day‐old wild‐type and ravl1 mutants. As the BRI1 and D2 levels were lower in ravl1 mutants than in wild‐type plants (Je et al., 2010), ACC‐mediated expressions of BRI1 and D2 were calculated as ratio to untreated control. The RT‐qPCR results indicated that ACC application suppressed BR1I and D2 expression, but the suppression was partially inhibited in ravl1 mutants (Figure 1p,q), suggesting that RAVL1 is involved in the ethylene‐mediated BR signalling gene suppression. Further molecular experiments are necessary to isolate the ethylene signalling factors that inhibit RAVL1 levels. Overall, our findings provided useful insight into the BR and ethylene crosstalk in rice. More importantly, the data revealed that RAVL1 activates both BR and ethylene signalling in rice.

Conflict of interest

The authors declare no conflict of interest.