Research ArticlePlant biology

An Interaction Between BZR1 and DELLAs Mediates Direct Signaling Crosstalk Between Brassinosteroids and Gibberellins in Arabidopsis

See allHide authors and affiliations

Science Signaling  02 Oct 2012:
Vol. 5, Issue 244, pp. ra72
DOI: 10.1126/scisignal.2002908


Plant growth and development are coordinated by several groups of small-molecule hormones, including brassinosteroids (BRs) and gibberellins (GAs). Physiological and molecular studies have suggested the existence of crosstalk between BR and GA signaling. We report that BZR1, a key transcription factor activated by BR signaling, interacts in vitro and in vivo with REPRESSOR OF ga1-3 (RGA), a member of the DELLA family of transcriptional regulators that inhibits the GA signaling pathway in Arabidopsis thaliana. Genetic analyses of plants with mutations in the genes encoding RGA and BZR1 revealed that RGA suppressed root and hypocotyl elongation of the gain-of-function mutant bzr1-1D. Ectopic expression of proteins of the DELLA family reduced the abundance and transcriptional activity of BZR1. Reporter gene analyses further indicated that BZR1 and RGA antagonize each other’s transcriptional activity. Our data indicated that BZR1 and RGA served as positive and negative regulators, respectively, of both the BR and the GA signaling pathways and establish DELLAs as mediators of signaling crosstalk between BRs and GAs in controlling cell elongation and regulation of plant growth.


Plant hormones play central roles in the regulation of plant growth and development; they not only orchestrate intrinsic developmental programs but also convey environmental inputs (1). Several classes of plant hormones have been characterized, and these control a wide range of physiological and developmental processes. Although each of these phytohormones regulates many different cellular processes, the same biological processes are controlled by different hormones (2), suggesting cooperation and crosstalk between their signaling pathways. For example, gibberellins [or gibberellic acids (GAs)] and brassinosteroids (BRs) are two classes of phytohormones that share many overlapping functions, including promoting stem elongation, seed germination, and flowering (3). Arabidopsis mutants disrupted in either GA or BR biosynthesis or perception show a typical dwarf phenotype, suggesting their essential roles in regulating plant growth (1). How GAs and BRs coordinate to control plant growth and other processes that they both regulate is poorly understood.

BRs are the polyhydroxylated steroidal hormones that are ubiquitously present in the plant kingdom. Extensive molecular genetic studies in the model plant Arabidopsis have established the essential roles of BRs in plant growth and development and have revealed a BR signal transduction pathway from specific cell surface receptor kinases to specific nuclear transcription factors. Briefly, BR signal is perceived by the BR receptor BRI1, a transmembrane leucine-rich repeat containing receptor-like kinase (LRR-RLK) (4). BRs directly bind the extracellular domain of BRI1, resulting in fast activation of the receptor’s intracellular kinase domain by means of phosphorylation and homodimerization (5). Activated BRI1 phosphorylates and activates its co-receptor BAK1 (also a LRR-RLK), which in turn phosphorylates and enhances the kinase activity of BRI1 (68). BR-induced BRI1 activation involves the release of BRI1 from another inhibitory membrane protein, BKI1 (BRI1 kinase inhibitor 1), which inhibits the kinase activity of BRI1 when the amount of BR is low or BR is absent (9). Activated BRI1 sequentially phosphorylates and activates two downstream BR signaling kinases, BSK1 and CDG1 (10, 11), and a Ser/Thr phosphatase BSU1 that inactivates the inhibitory kinase BIN2 (12). BIN2 is a cytoplasmic glycogen synthase kinase 3 (GSK3)–like protein kinase that inhibits BR signaling by phosphorylating and inactivating BZR1 and BES1, two key transcription factors that mediate BR responses (1315). In the presence of BR, BZR1 and BES1 become dephosphorylated and activated, presumably through the action of the phosphatase PP2A (protein phosphatase 2A) (16). Further protein-DNA binding experiments have demonstrated that BZR1 and BES1 can both bind the 5′-CGTG(T/C)G-3′ elements (BRRE) and 5′-CANNTG-3′ (E-box) of their target genes, with a preference for BRRE (1720).

The GA signaling pathway involves a different paradigm. GA is perceived by a soluble receptor, GIBBERELLIN INSENSITIVE DWARF 1 (GID1), which was originally identified from rice (Oryza sativa) and has three redundantly acting homologs, GID1A, B, and C, in Arabidopsis (21, 22). The current model suggests that GA binding to GID receptor family leads to conformation changes and nuclear localization of the receptors, which promotes their interaction with the DELLA proteins (23, 24). DELLAs, including GIBBERELLIC ACID INSENSITIVE (GAI), REPRESSOR OF ga1-3 (RGA), RGA-like 1 (RGL1), RGL2, and RGL3, are transcriptional regulators that restrict plant growth presumably by causing transcriptional reprogramming (25). The GA-GID-DELLA complex stabilizes the interaction of DELLA proteins with the SLEEPY1 (SLY1) F-box protein, a SCF-type E3 ubiquitin ligase component (26, 27) that targets DELLA proteins for degradation by the ubiquitin-proteasome pathway. Therefore, it is believed that GA promotes plant growth by means of removing the inhibitory DELLA proteins.

DELLA proteins are not only core components for GA signaling, but they also play a central role in integrating plant responses to various hormonal and environmental signals. For example, the effects of three phytohormones—auxin, ethylene, and jasmonate (JA)—on root growth all involve modulation of DELLA stability (28), and stress-triggered growth restriction by abscisic acid (ABA) and ethylene is likely also mediated through stabilization of DELLA proteins (29, 30). Moreover, DELLAs mediate the interaction between GA and light in controlling cell elongation (31, 32) and are involved in changes in root architecture and the accumulation of anthocyanins in response to phosphorus starvation (33). DELLA proteins promote JA signaling through competitive binding to JA ZIM-domain (JAZ) proteins, key repressors of JA signaling (34). DELLA proteins also promote the formation of functional chloroplasts during deetiolation, a process stimulated by light (35).

Because of the pivotal roles of DELLAs in mediating crosstalk among these various signaling pathways, we hypothesized that these proteins may also mediate the signaling crosstalk between BRs and GAs. Possible crosstalk at the transcriptional level has been reported: BZR1 binds the promoters of genes encoding DELLA family members and, thus, may regulate their expression (19). Here, we demonstrated that BZR1 interacts with DELLAs at the protein level to modulate both BR- and GA-induced cell elongation processes.


BZR1 and DELLAs interact in vitro and in vivo

To determine whether BZR1 and DELLAs mediate BR and GA crosstalk, we first examined the physical interaction between BZR1 and members of the DELLA family with in vitro and in vivo studies. Yeast two-hybrid analyses with full-length BZR1 and full-length DELLA proteins indicated that BZR1 interacted with all the members of the DELLA family (RGA, GAI, RGL1, RGL2, and RGL3) (Fig. 1A), albeit with different affinities. BES1, the close homolog of BZR1, also interacted with RGA in yeast two-hybrid assays (fig. S1). To determine which domains of the BZR1 and RGA proteins were responsible for their interaction, we created a set of deletion constructs for both proteins. The DELLA proteins contain the conserved Asp-Glu-Leu-Leu-Ala (DELLA) and Val-His-Tyr-Asn-Pro (VHYNP) domains at the N terminus and a conserved GRAS domain at the C terminus. The GRAS domain consists of five distinct sequence motifs: the leucine heptad repeat I (LHR I), the LHR II, the Val-His-Ile-Ile-Asp (VHIID), the Pro-Phe-Tyr-Arg-Glu (PFYRE), and the Ser-Ala-Trp (SAW) motif (36). The BZR1 protein contains a putative nuclear localization sequence, followed by a DNA binding domain (DBD), a BIN2 phosphorylation domain, a Pro-Glu-Ser-Thr (PEST) motif (involved in protein degradation), and a C-terminal domain (37). The yeast two-hybrid results revealed that the BIN2 phosphorylation domain of BZR1 and the LHR1 domain of RGA were required for their interaction (fig. S2, A and B). The BIN2 phosphorylation domain of BZR1 is not only responsible for BIN2-regulated BZR1 stability but also mediates BZR1 interaction with the PP2A phosphatase (16, 37). The LHR1 domain of RGA is essential for its activity in vivo (38, 39) and mediates the interaction of RGA with other transcription factors, such as phytochrome-interacting factors (PIFs) (31) and JAZ1 (34). We confirmed the direct interaction of BZR1 and RGA by in vitro pull-down assay, which showed that BZR1 fused to maltose-binding protein (MBP) specifically interacted with RGA fused to glutathione S-transferase (GST) (Fig. 1B).

Fig. 1

BZR1 interacts with DELLA in vitro and in vivo. (A) Yeast two-hybrid assay for interaction between BZR1 and DELLA proteins. Yeast cells transformed with bait (pGBKT7-BZR1) and prey (the indicated DELLA member cloned into pGADT7) pairs were selected on SD-Trp/-Leu/-His medium supplemented with 50 mM 3-aminotriazol (3-AT). pGADT7 is the empty vector control. (B) In vitro pull-down assay for BZR1 and RGA interaction. MBP-BZR1 was pulled down by GST-RGA, and the BZR1 and RGA interaction was detected with antibodies recognizing MBP (anti-MBP). (C) Colocalization of BZR1-GFP and RFP-RGA in tobacco epidermal cells. The scale bars in the upper and lower panels are 50 and 5 μm, respectively. (D) BiFC analysis of BZR1 and RGA interaction in planta. Vectors containing the indicated constructs were cotransformed into Nicotiana benthamiana leaves. Both YFP fluorescence images (upper panel) and fluorescence images merged with light view images (lower panel) are shown. (E) Coimmunoprecipitation assay of BZR1 with RGA. Total protein extracted from transgenic plants coexpressing BZR1-GFP and RGA-myc or BZR1-GFP and the empty myc vector was immunoprecipitated (IP) by c-Myc antibody–conjugated agarose beads. The precipitated proteins were detected with either an antibody recognizing GFP (anti-GFP) or an antibody recognizing myc (anti-myc). The input samples were detected with anti-GFP.

We used three different methods to demonstrate the interaction between BZR1 and RGA in plants. First, coexpression of BZR1 tagged with green fluorescent protein (GFP) and RGA tagged with red fluorescent protein (RFP) in the tobacco leaf epidermal cells indicated that both proteins colocalized in the nucleus (Fig. 1C). Second, bimolecular fluorescence complementation (BiFC) assay showed that strong fluorescence signals were only observed from the nuclei of cells coexpressing BZR-cYFP (C-terminal fragment of yellow fluorescent protein) and RGA-nYFP (N-terminal fragment of yellow fluorescent protein) in tobacco leaf epidermis, indicating that BZR1 and RGA interacted in the nucleus (Fig. 1D). BiFC analysis also showed that BES1 interacted with RGA in the nucleus (fig. S3). To further confirm the interaction between BZR1 and RGA in Arabidopsis by coimmunoprecipitation, we introduced the pRGA::RGA-myc plasmid into the pBZR1::BZR1-CFP plants [expressing BZR1 tagged with cyan fluorescent protein (CFP) driven by the BZR1 promoter] to make the double-transformed transgenic plants. However, we could not detect any expression of the RGA-myc protein in the resulting transgenic plants. Therefore, we used transgenic tobacco plants coexpressing the proteins from the 35S::BZR1-GFP and 35S::RGA-myc plasmids for the coimmunoprecipitation experiments and detected an interaction between BZR1 and RGA (Fig. 1E).

DELLAs and BZR1 modulate both BR and GA signaling

We used genetic interaction assays to gain insight into how the interaction between BZR1 and DELLA affected plant growth. We crossed the GA-insensitive mutant gai, which expresses a constitutively active DELLA (GAI lacking the DELLA domain) (40), with bzr1-1D, a dominant mutant of BZR1 that exhibits increased BZR1 stability and thus constitutive, ligand-independent BR pathway activity (41). We also crossed the gai mutant with bri1-5. The bri1-5 mutant is a weak loss-of-function mutant for the BR receptor BRI1, which produces plants with a dwarf phenotype (42); the bzr1-1D plants are similar in size to wild type with thick and dark green leaves (41); and the gai plants exhibit a severe dwarf phenotype (40).

Although the single mutant plants gai and bzr1-1D had different hypocotyl lengths, the gai bzr1-1D double mutant plants were dwarfed, similar in size to the gai single mutant, albeit with different leaf morphologies (Fig. 2A). Similarly, the double mutant gai bzr1-1D plants also had a short root phenotype similar to that of the gai single mutants (fig. S4). These data suggested that gai suppressed the constitutive activity of the bzr1-1D mutant. To assess whether other DELLA proteins could suppress the growth of the bzr1-1D plants, we created a mutant RGA lacking the DELLA domain (RGAΔ17). Overexpression of RGAΔ17 in the bzr1-1D background also resulted in dwarf phenotypes (Fig. 2B). We also assessed the responses of the gai bzr1-1D double mutant and RGAΔ17/bzr1-1D transgenic plants to brassinazole (BRZ), a specific BR biosynthetic inhibitor (43), and compared these responses to those of bzr1-1D single mutants. Like bzr1-1D, both the gai bzr1-1D and RGAΔ17/bzr1-1D plants were not sensitive to BRZ treatment as quantified by the ratio of hypocotyl lengths under BRZ treatment and without treatment (Fig. 2C and fig. S5, A to C), suggesting that DELLAs did not alter BRZ insensitivity phenotype of the bzr1-1D mutant. The bri1-5 gai double mutant had a more severe dwarf phenotype than either of the parental mutants (Fig. 2D and fig. S6).

Fig. 2

DELLAs negatively regulate the BR pathway, and BZR1 positively regulates the GA pathway. (A) Growth phenotypes of 4-week-old bzr1-1D and its wild-type Col, gai bzr1-1D, and gai and its wild-type Ler plants. (B) Phenotypes of 4-week-old Col, bzr1-1D, and 35S::RGAΔ17/bzr1-1D plants. The lower panel shows the RT-PCR results for RGAΔ17 or RGA expression in the indicated plants. UBC served as an internal control. Data are representative of three experiments. (C) Comparison of the relative BRZ response between the gai bzr1-1D double mutant, the indicated single mutants, and their respective wild-type plants. The data of each genotype represent the ratios of hypocotyl length of plants under BRZ treatment (1 μM for 6 days) relative to that of the mock (DMSO)–treated plants. (D) Phenotypes of 4-week-old bri1-5 and its wild-type WS, and gai bri1-5 and gai, and their wild-type Ler. (E) Comparison of BRZ sensitivity between bzr1-1D and Col and between rga-24 gai-t6 and its wild-type Ler. Six-day-old dark-grown seedlings were treated with indicated concentrations of BRZ for 6 days, and hypocotyl length was measured. (F) Comparison of PAC sensitivity between bzr1-1D and Col, bri1-5 and WS, and rga-24 gai-t6 and Ler. Six-day-old dark-grown seedlings were treated with different concentrations (0, 0.025, 0.05, 0.1, and 0.2 μM) of PAC. In (C), (E), and (F), the relative BRZ or PAC sensitivity represents the ratio of hypocotyl length of each genotype under each concentration of BRZ or PAC treatment relative to that of the mock-treated matching wild-type control. Data were analyzed by one-way analysis of variance (ANOVA) followed by Bonferroni’s correction. Error bars represent SD (n > 20 seedlings). *P < 0.05; **P < 0.01; NS, not significant. For (C), significance between the indicated genotypes is indicated. For (E) and (F), significant differences from the wild-type control under the same treatment conditions are indicated.

We assessed the sensitivity of rga-24 gai-t6, an RGA and GAI loss-of-function double mutant that has long hypocotyl and low sensitivity to the GA biosynthetic inhibitor paclobutrazol (PAC) (44), to BR and BRZ treatments. Compared to the wild-type control (Ler), the rga-24 gai-t6 mutant showed a higher sensitivity to brassinolide (BL, the most active BR) in a primary root elongation assay (fig. S7, A to C). In contrast, root length (fig. S7, D to F) and hypocotyl length (Fig. 2E and fig. S8) were less reduced by BRZ in rga-24 gai-t6 compared to wild-type plants exposed to the same concentrations of BRZ. These results indicate that DELLAs can modulate BR signaling.

To determine whether plants with altered BR signaling exhibited altered responses to activation or inhibition of the GA pathway, we tested the BR mutants bri1-5 and bzr1-1D for their responses to GA3 (an active GA) and PAC. The bri1-5 and bzr1-1D seedlings showed similar responses to GA3 treatment to the responses of their wild types [Wassilewskija-2 (WS) and Columbia (Col-0), respectively] (fig. S9, A to C): The GA-induced increase of hypocotyl length in each mutant was similar to the increase of hypocotyl length in their respective wild types exposed to the same concentrations of GA3. In contrast, bzr1-1D reduced and bri1-5 enhanced plant sensitivity to PAC in both root elongation assay (fig. S9, D to I) and hypocotyl elongation assay (Fig. 2F and fig. S10), indicating that the BR pathway can modulate the GA pathway.

The abundance of BZR1 is regulated by GA, but that of DELLAs is not regulated by BR

To determine whether BR or GA can affect each other’s biosynthesis, we analyzed the expression of CPD, DWF4, GA20ox2, and GA3ox1, which are rate-limiting genes for BR (CPD, DWF4) and GA (GA20ox2, GA3ox1) biosynthesis. The expression of GA20ox2 and GA3ox1 was significantly increased by both BL treatment and bzr1-1D mutation (Fig. 3, A and B). The expression of CPD and DWF4 was not affected by GA treatment, but transcript abundance was increased in the gai mutant (Fig. 3, C and D). These results suggest that BR and GA can modulate each other’s biosynthesis through pathways involving BZR1 and DELLA.

Fig. 3

BR does not affect DELLA abundance, but GA stabilizes BZR1 and promotes its dephosphorylation. (A and B) Transcript abundance of the indicated genes in plants exposed to BL (10−6 M, 2 hours) (A) or in the bzr1-1D mutant (B) was determined by qRT-PCR analyses. GA20ox2 and GA3ox1 are GA biosynthetic genes. The BR biosynthetic genes CPD and DWF4 served as positive controls for BR treatment. Data were normalized to the abundance of UBC. (C and D) Transcript abundance of the indicated genes in plants exposed to GA3 (50 μM, 2 hours) (C) or in the gai mutant (D). The GA biosynthetic genes GA20ox2 and GA3ox1 in (C) served as positive controls for GA treatment. Data were normalized to the abundance of EF-1a. In (A) to (D), error bars represent SD (n = 3 experiments). *P < 0.05; **P < 0.01 (Student’s t test). (E) Effects of GA3 (50 μM), PAC (1 μM), BL (1 μM), or BRZ (2 μM) treatment on the abundance of GFP-RGA in root tips of pRGA::GFP-RGA transgenic plants. Plants were treated for 12 days, and GFP signals were observed with a confocal microscope. The red color represents cells stained with propidium iodide, a dye to stain DNA. (F) Effect of BL (10−6 M) treatment on the abundance of RGA-GFP (detected with an antibody recognizing GFP). (G) Effect of BL (10−6 M, 2 hours) on DELLA protein abundance in transgenic plants overexpressing the indicated TAP-tagged DELLA proteins (32). Asterisk denotes a nonspecific protein band detected with the antibody recognizing the myc tag. (H) Effect of GA3 (50 μM) on BZR1 accumulation and phosphorylation status in pBZR1::BZR1-CFP plants compared to mock-treated plants exposed to EtOH. (I) Effect of PAC (1 μM) on BZR1 accumulation and phosphorylation status in pBZR1::BZR1-CFP and pBZR1::mBZR1-CFP transgenic plants compared to mock-treated plants exposed to methanol (MetOH). (J and K) Effects of BIN2 inhibition with LiCl, with and without, and GA (J) or PP2A inhibition with OA, with and without, and GA (K) on BZR1 abundance and phosphorylation status in pBZR1::BZR1-CFP seedlings. (L) Effects of combinational treatments with LiCl, OA, and GA on BZR1 abundance and phosphorylation status. In (H) to (L), BZR1 was detected with the antibody against GFP that recognizes the CFP tag. The numbers under the blots are normalized relative protein abundance of phosphorylated (pBZR1) or unphosphorylated BZR1 (BZR1) compared with that of pBZR1 in the first sample, which was set to 1. Rubisco was used as a loading control. Every Western blot is the representative of at least three experiments. See fig. S12 for quantification of the data represented in (H) to (L).

We also investigated whether the abundance of BZR1 was regulated by GA and that of RGA by BR at both the transcriptional and translational levels. At the transcriptional level, exposure of plant seedlings to GA did not affect BZR1 or BES1 expression, and exposure of plants to BR did not affect RGA and GAI expression (fig. S11, A and B). BZR1 or BES1 expression was also not altered in gai mutant plants, and RGA and GAI expressions was not affected in the bzr1-1D plants (fig. S11, C and D). At the translational level, both confocal microscopy (Fig. 3E) and Western blot analysis (Fig. 3F) indicated that neither activation nor inhibition of BR signaling with BL or BRZ, respectively, altered the abundance of RGA in the pRGA::GFP-RGA transgenic plants. Western analyses of other DELLA proteins using the overexpression transgenic lines also indicated that activation of the BR pathway with BL did not alter their abundance (Fig. 3G).

In contrast, in pBZR1::BZR1-CFP transgenic plants, GA treatment did not affect the abundance of phosphorylated BZR1 (pBZR1). However, there was an apparent, although not statistically significant, increase in the abundance of the dephosphorylated form (BZR1) compared with the mock-treated samples (Fig. 3H and fig. S12A). Inhibition of GA signaling with PAC caused an apparent decrease in the abundance of dephosphorylated BZR1, but not of phosphorylated BZR1, in pBZR1::BZR1-CFP transgenic plants; in pBZR1::mBZR1-CFP plants (expressing the BZR1 with the bzr1-1D mutation and tagged with CFP driven by the BZR1 promoter), PAC treatment decreased both pBZR1 and BZR1 abundance, indicating an overall reduction in the amount of the mBZR1 protein (Fig. 3I and fig. S12B).

To understand the underlying mechanisms of the effect of GA on BZR1 abundance and phosphorylation state, we examined how GA affected BZR1 in the pBZR1::BZR1-CFP plants in the presence of inhibitors of the kinase BIN2 or the phosphatase PP2A, two proteins that mediate BZR1 phosphorylation and dephosphorylation, respectively. Although there was substantial variability across the experiments, it appeared that 4-hour treatment with 2 mM LiCl (an inhibitor of BIN2) (45) induced dephosphorylation of BZR1 compared with the mock [KCl and ethanol (EtOH)]–treated sample (Fig. 3J, compare lane 2 with lane 4, and fig. S12C). Compared to plants exposed to LiCl or GA3 alone, even less phosphorylated BZR1 and more dephosphorylated BZR1 were detected in samples from plants exposed to both GA3 and LiCl (Fig. 3J, compare lane 1 with lanes 2 and 3, and fig. S12C). When the plants were treated with the protein phosphatase inhibitor okadaic acid (OA) (16), the GA3-induced enhancement of BZR1 dephosphorylation was abolished (Fig. 3, K and L, and fig. S12, D and E). Thus, GA appeared to promote BZR1 dephosphorylation, and this process may be mediated by PP2A.

Ectopic expression of DELLAs causes destabilization and inactivation of the BZR1 protein

Because the GA pathway, which controls DELLA stability, affected BZR1 protein abundance and phosphorylation status, we examined the effects of altering the abundance of DELLA proteins on BZR1 in plants. We introduced 35S::RGAΔ17 and pGAI::GAIΔ17 fragments into the pBZR1::BZR1-CFP and pBZR1::mBZR1-CFP plants, respectively, and assessed their impact on plant growth and on BZR1-CFP or mBZR1-CFP accumulation. Overexpression of RGAΔ17 or GAIΔ17 resulted in extremely dwarfed plants (fig. S13, A and B) and a reduction in BZR1 protein abundance (Fig. 4, A and B), suggesting that DELLAs restrain plant growth or cell elongation, at least in part, by affecting BZR1 protein abundance.

Fig. 4

Ectopic expression of DELLAs reduces BZR1 abundance, and DELLAs preferably bind to the dephosphorylated form of BZR1. (A) pBZR1 and BZR1 abundance in 35S::RGAΔ17/pBZR1::BZR1-CFP transgenic plants. (B) Abundance of pBZR1 and BZR1 in gai pBZR1::mBZR1-CFP plants. (C) PP2AB′α and PP2AB’β protein accumulation in 35S::PP2AB′α-YFP and 35S::PP2AB′β-YFP transgenic plants treated with PAC (1 μM). (D) Cell-free degradation of PP2AB′α-YFP and PP2AB′β-YFP proteins in the presence or absence of the proteome inhibitor MG132. In (A) to (D), both BZR1-CFP and PP2A-YFP were detected with an antibody against GFP that recognizes the CFP and YFP tag. Rubisco served as the loading control of Western blots. (E) In vitro overlay assay of GST-RGA binding with pBZR1 and BZR1. The overlay results and inputs were detected with anti-GST or anti-MBP, respectively. (F) Semi–in vivo pull-down assay of RGA interaction with pBZR1 and BZR1 in 35S::mBZR1-myc transgenic seedlings treated with or without BL (10−6 M, 4 hours). The BZR1-myc proteins pulled down by GST-RGA were detected by Western blotting using an anti-myc. Each Western blot is the representative of at least three experiments.

We explored various potential mechanisms for the effect of DELLAs on BZR1 protein stability and phosphorylation status. We used a yeast two-hybrid assay to assess whether RGA interacted with BIN2, which could enhance the interaction between BIN2 and BZR1 by recruiting more BIN2 to BZR1. However, we did not detect an interaction between RGA and BIN2 (fig. S14). We used an in vitro competition pull-down assay to show that RGA did not affect the interaction of PP2A with pBZR1 (fig. S15).

We found that the abundance of two PP2AB′ subunits (PP2AB′α or PP2AB′β) in plants overexpressing YFP-tagged forms of either of these subunits was reduced by treatment of the plants with PAC to inhibit GA signaling and increase the abundance of DELLAs (Fig. 4C). The PAC-induced loss of PP2A abundance was not due to reduced transcription because transcripts for these two proteins actually increased in response to PAC treatment or in plants with the gai mutation (fig. S16, A and B). A cell-free degradation assay performed with proteins extracted from the 35S::PP2AB′α and 35S::PP2AB′β transgenic plants revealed that the degradation of PP2AB′ proteins was inhibited by treatment with the proteasome inhibitor MG132 (Fig. 4D), suggesting that PP2A degradation is mediated by the 26S proteasome pathway.

We used an in vitro overlay assay to show that more RGA bound to the dephosphorylated form of BZR1 than to the phosphorylated form (Fig. 4E). Similar results were obtained from a “semi–in vivo” pull-down assay in which BZR1 was pulled down from the protein extracts from 35S::mBZR1-myc transgenic plants by GST-RGA and was detected with an antibody that recognizes the myc tag (Fig. 4F). These results suggest that the interaction of DELLAs with BZR1 is ligand-dependent and that the accumulation of DELLAs could impair BZR1 activity by binding to its dephosphorylated (and thus activated) form.

BZR1 and RGA attenuate each other’s transcriptional activity and target gene expression

Because both BZR1 and RGA are transcription regulators, we used a transient assay system (46) to test whether their interaction affects each other’s transcriptional activity (Fig. 5A). For this purpose, RGA was expressed as a fusion protein with the GAL4 DBD under the control of a 35S promoter (Fig. 5A). The reporter construct consisted of a minimal 35S promoter with the GAL4 DNA binding site (DBS) driving a Luciferase (LUC) reporter gene. The Renilla Luciferase (RNL LUC) gene under the control of the 35S promoter served as an internal control. These constructs were expressed transiently in protoplasts from 4-week-old Col-0 or bzr1-1D plants, and the activated LUC reporter activity was assayed. Expression of RGA in the system stimulated LUC expression compared to that of the vector control in Col-0 protoplasts, and this stimulation was significantly suppressed by BL treatment (Fig. 5B). Compared to expression of the constructs in the wild-type protoplasts, in the bzr1-1D protoplasts, RGA transcriptional activity was less effective at stimulating the reporter, and the activity of RGA was not further reduced by BL (Fig. 5B). These data suggest that BZR1 attenuated the transcriptional activity of RGA.

Fig. 5

RGA and BZR1 antagonistically regulate the transcription of growth-related genes. (A) Constructs used for RGA transcriptional activity assay in the pMN6 system (46). LUC, firefly luciferase; REN LUC, Renilla luciferase. (B) Effects of BL treatment or BZR1 overexpression on RGA transcriptional activity performed in Col or bzr1-1D protoplasts. pMN6 represents vector-transfected cells; all others represent RGA-expressing cells. Error bars represent SD (n = 4 experiments). (C) A schematic map of the transient expression vector pGreenII-0800-LUC (49). (D) Effects of BZR1 and BZR1(ΔBIN2) on RGA transcriptional activity at the SCL3 promoter reporter in Col protoplasts. (E) Effect of RGA on BZR1 transcriptional activity at the SAUR-AC promoter reporter in Col protoplasts. (F) Effects of RGA and RGA(ΔLHR1) on mBZR1 transcriptional activity at the SAUR-AC promoter reporter in Col protoplasts. In (D) to (F), error bars represent SD (n = 5 experiments). (G and H) Transcript abundance (detected by qRT-PCR analyses) of genes controlling cell elongation in bes1-D and gai mutants. UBC and EF-1a were used as an internal control for data normalization in (G) and (H), respectively. Error bars represent SD (n = 3 experiments). *P < 0.05; **P < 0.01; NS, not significant (ANOVA with Bonferroni’s correction).

We examined whether RGA and BZR1 affected the expression of each other’s target genes in an antagonistic way. We chose to use the promoters of SCARECROW-LIKE 3 (SCL3), a direct target gene of RGA (47, 48), which was suppressed by GA and enhanced by PAC treatment or gai mutation (fig. S17), and SAUR-AC, a BZR1 target gene (19), to assess the effects of RGA and BZR1 on transcription with a different transient assay system (49) designed to quantify the interactions between transcription factors and target promoter interactions (Fig. 5C). We found that the LUC expression from the SCL3 promoter was enhanced by RGA but not by BZR1 (Fig. 5D). However, when both BZR1 and RGA were present, the RGA-mediated stimulation of SCL3-driven LUC transcription was reduced (Fig. 5D), suggesting that BZR1 has an inhibitory effect on the transcriptional activity of RGA. Analysis of a mutant form of BZR1 lacking the BIN2 phosphorylation domain, which was required for the interaction between BZR1 and RGA, showed that this interaction was necessary for the inhibition of RGA activity by BZR1 (Fig. 5D), suggesting that the suppression effect of BZR1 on RGA transcriptional activity depended on the BZR1-RGA interaction.

To evaluate BZR1 transcriptional activity, we modified the assay in two ways. In addition to the BZR1 construct and the RGA construct, we included a construct encoding the BES1-interactive Myc-like 1 (BIM1) in the assay because previous reports indicated that binding of BES1, the homolog of BZR1, to target gene promoters was enhanced by the presence of BIM1 (18). We also used the 35S::mBZR1 construct because it produced more stable and active mBZR1 than the wild-type BZR1. Cotransformation of 35S::BIM1 with 35S::mBZR1 into Col-0 protoplasts significantly activated the SAUR-AC reporter; however, this activation was eliminated by the introduction of 35S::RGA (Fig. 5F), indicating that RGA suppressed the transcriptional activity of BZR1. To determine whether the RGA-induced suppression of BZR1 transcriptional activity is caused by RGA-BZR1 interaction, we introduced a construct encoding RGA(ΔLHR1), which lacks the domain required for the interaction, with BZR1 in the transient assay system. Without the LHR1 domain, RGA failed to suppress BZR1-mediated activation of the SAUR-AC reporter (Fig. 5F).

Because DELLA proteins affected both the activity and the stability of BZR1, we hypothesized that overexpression of DELLA proteins would reduce the expression of BR-responsive genes, including those controlling cell elongation, which would produce the phenotypic effects observed in the functional genetic studies. We performed quantitative expression analysis of BR-responsive genes involved in regulation of cell elongation in the bes1-D and gai dominant mutants. We used bes1-D because BZR1 can activate a negative feedback regulatory pathway that may compromise BZR1’s transcription activating activity for some of its target genes, whereas BES1 constitutively activates BR response and is not sensitive to this feedback inhibition (15, 41). We analyzed the following genes encoding proteins that induce cell wall modification to mediate cell elongation (50, 51): xyloglucan endotransglycosylase/hydrolase (XTH), expansins (EXP), and cellulose synthase genes (CESAs). Quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis of transcript abundance showed that all the tested genes exhibited an increase in expression in bes1-D and a decrease in expression in the gai mutant (Fig. 5, G and H).


Questions related to the coordination of hormone signaling in plants

One of the outstanding questions that have been challenging plant biologists is how a single plant hormone can fulfill so many different physiological functions and, at the same time, how a subset of different hormones can redundantly control the same response. For example, whereas each of the eight plant hormones (auxins, GAs, ABA, cytokinins, ethylene, BRs, JAs, and strigolactones) has their distinct functions, they are all connected to growth regulation, sometimes in a tissue-specific manner (52, 53). Little is known regarding how these different hormones communicate with each other and coordinate plant growth at the cell, organ, and tissue levels. Pair-wise interaction studies have suggested that different hormone signals may integrate at the transcriptional level, and there are points of crosstalk among their signaling pathways. However, only a few components that mediate direct crosstalk have been reported (53). Direct signaling crosstalk between BR and auxin signaling was reported: The BR-regulated kinase BIN2 directly phosphorylated and inactivated ARF2, a repressor of auxin signaling, thus resulting in increased transcription of auxin-responsive genes (54).

The interaction of BZR1 and RGA as the point of crosstalk between BR and GA signaling pathways

Surprisingly, previous comparative transcriptome studies have revealed little overlap between BR- and GA-regulated genes, which led to the conclusion that these two hormones may regulate plant growth independently (55). However, ample evidence of crosstalk between BR and GA signaling includes (i) physiological studies indicating that BR and GA co-regulate hypocotyl elongation (56), seed germination (57, 58), and flowering time (59); (ii) studies showing that these hormones affect each other’s biosynthesis (3); and (iii) studies showing that they can affect the expression of each other’s responsive genes (60). Here, we identified a physical connection between the BR and GA signaling pathways, which is mediated by the direct interaction between BZR1 and RGA, two central transcriptional regulators in the BR and GA pathways, respectively.

Although previous microarray studies have indicated little overlap between BR- and GA-regulated genes (55), comparison of RGA-regulated genes from published microarray data (47) and published BZR1 target genes (19) indicated that up to 30% of RGA-responsive genes were also direct targets of BZR1 (fig. S18), suggesting that BZR1 and RGA may control a common transcriptional module or function as two hubs in the BR- and GA-regulated transcription network. Because DELLAs do not contain a canonical DBD (25), it is likely that they regulate cell elongation by modulating the expression of their target genes by interacting with other transcription factors. For example, DELLAs interact directly with PIFs (31, 32) and JAZ family of proteins (34) to mediate crosstalk between GA signaling and light and JA signaling, respectively. Here, we demonstrate that DELLAs directly interact with BZR1 and BES1 from the BR pathway, further highlighting the central roles of DELLAs in mediating crosstalk among plant signaling pathways.

In a chromatin immunoprecipitation (ChIP)–chip study of BZR1 target genes, four of the five DELLA-encoding genes (RGA, GAI, RGL1, and RGL3) were identified as direct targets of BZR1 (19), which may represent another level of BR and GA interaction. However, our qRT-PCR analysis indicated that BZR1 did not affect the transcription of RGA and GAI, suggesting that BZR1 and DELLA coordinate the interaction between BR and GA in controlling cell elongation primarily through protein-protein interaction rather than through BZR1 regulation of DELLA gene expression.

RGA and GAI are the main DELLA members that repress cell growth (47). Genetic and pharmacological analyses showed that DELLAs inhibited the BR signaling pathway (Fig. 2). The rga-24 gai-t6 double knockout mutant was partially resistant to the BR biosynthetic inhibitor BRZ, suggesting that the enhancement of GA signaling caused by loss of function of DELLAs can compromise the response to BR. Likewise, the bzr1-1D mutant was partially resistant to the GA inhibitor PAC, indicating that enhanced BR signaling due to gain of function of BZR1 could also compromise the response to GA. Furthermore, the gain-of-function mutants of RGA and GAI suppressed cell elongation in the bzr1-1D plants.

Mechanisms of BR and GA coordination in the regulation of cell elongation and plant growth

Genetic, biochemical, and structural studies have elucidated a pivotal regulatory module for GA in plant growth and development by triggering the degradation of the growth-inhibiting DELLA proteins (25). Our results suggested that GA may contribute to the stabilization of BZR1 and promote its dephosphorylation and, thus, accumulation of the active form of BZR1 (Fig. 3). In contrast, accumulation of the nondegradable RGAΔ17 mutant protein reduced the abundance of BZR1 in vivo (Fig. 4). Although previous studies have suggested that the 26S proteasome pathway is involved in BZR1 and BES1 degradation (14, 15), the ubiquitin E3 ligase(s) responsible has not been identified. To affect BZR1 stability, binding of DELLAs to BZR1 could enhance the phosphorylation of BZR1 by recruiting the kinase BIN2 or by preventing the dephosphorylation of BZR1 by the phosphatase PP2A, or both. We consider the first possibility less likely because we did not detect an interaction between BIN2 and RGA by the yeast two-hybrid assay. However, in plants in which DELLAs were stabilized due to inhibition of GA biosynthesis, the abundance of PP2A was diminished, suggesting that this could be a mechanism by which DELLAs could affect BZR1 stability. Consistent with this model, we found that PP2A abundance was regulated by the 26S proteasome and that pharmacological inhibition of PP2A blocked GA-induced BZR1 dephosphorylation. Whether the interaction of RGA and BZR1 can also enhance the binding of BZR1 with a BZR1-specific E3 ubiquitin ligase, thereby facilitating its degradation through the 26S-ubiquitin proteasome pathway, was not tested.

Pull-down assays showed that DELLAs interacted preferentially with the dephosphorylated BZR1, and transcriptional activity assays using both wild-type and mutated RGA proteins in Arabidopsis protoplasts showed that the attenuating effect of RGA on the transcriptional activity of BZR1 depended on their interaction. In vivo, the inhibitory effect of DELLAs on the expression of several BR-induced genes required for cell elongation was confirmed using gai plants. Thus, our data indicated that the effect of DELLAs on cell elongation involves their ability to reduce BZR1 or BES1 stability and transcriptional activity.

Whereas DELLAs appeared to affect both BZR1 or BES1 stability and activity, overexpression of BZR1 or BL treatment appeared to only interfere with the transcriptional regulatory activity of DELLAs without affecting the abundance of DELLA proteins. In protoplast transient assays, a mutant BZR1 with deletion of the RGA-interacting domain failed to antagonize RGA’s transcriptional activity, suggesting that BZR1 interferes with DELLA activity by a direct interaction. This effect of BR on DELLAs is different from the effect of GA, which reduces the abundance of DELLAs through a proteasome-dependent degradation mechanism (25). However, GA-mediated plant growth does not always require the degradation of DELLAs (61, 62). Therefore, we propose that BR signaling enhances GA signaling by promoting the interaction between DELLAs and BZR1 or BES1 to alleviate DELLA’s restraint on GA-mediated growth.

On the basis of our data, we proposed a model to illustrate how BR and GA coordinate to regulate cell elongation and plant growth (Fig. 6). Briefly, BR dephosphorylates and stabilizes BZR1 and BES1 to trigger their nuclear translocation and activation of their target genes, including activation of those genes controlling cell elongation. Meanwhile, accumulated BZR1 and BES1 interact with DELLAs in the nucleus to suppress the latter’s transcriptional activity. Unlike BR, GA promotes cell elongation by stabilizing the GA-GID1-DELLA complex and by initiating the rapid degradation of DELLAs, therefore releasing the suppressive effects of DELLAs on growth-promoting factors such as BZR1 and BES1. When BR or GA signal is absent, BZR1 and BES1 are degraded, and more DELLAs accumulate in the nucleus, resulting in reduced cell elongation and plant growth.

Fig. 6

A model for BR and GA signaling crosstalk mediated by BZR1 and DELLA interaction. (A) In the absence of BR signal, BZR1 and BES1 are mostly present in their phosphorylated forms and can be degraded by the 26S proteasome. For simplicity, BZR1 and BES1 are indicated with a single shape. These proteins are not a complex. In the presence of BR, BZR1 and BES1 are dephosphorylated and activated by the phosphatase PP2A and can stimulate their downstream target genes to promote cell elongation. The dephosphorylated active BZR1 and BES1 could also sequester the DELLA proteins to enhance GA-regulated cell elongation. (B) When GA is absent, DELLA proteins accumulate and bind to the dephosphorylated form of BZR1 or BES1 and trigger the degradation of PP2A, thus inhibiting the activity of BZR1 and BES1 and cell elongation. When GA is present, DELLAs, GA, and the GA receptor GID1 form a complex, leading to degradation of DELLAs through the 26S proteasome. GA also induces PP2A-mediated dephosphorylation of BZR1 and BES1, and expression of growth-promoting genes occurs.

Materials and Methods

Plant materials and growth conditions

The BR mutants used in this study included bri1-5, bzr1-1D, and bes1-D, and their respective wild types WS, Col-0, and Ertkheim-2 (EN2). The GA mutants included gai and the double mutant rga-24 gai-t6, which are in the Landsberg erecta (Ler) background. The double mutants gai bzr1-1D and gai bri1-5 were generated by crossing bzr1-1D and bri1-5 to the gai mutant, respectively. Transgenic Arabidopsis lines were generated by the Agrobacterium-mediated floral dip method (63).

Sterilized Arabidopsis seeds were grown on Murashige and Skoog (MS) medium under a 16-hour/8-hour light/dark cycle at 22°C. For hypocotyl elongation and protein subcellular localization experiments, seeds were transferred to new MS media with or without BL or BRZ, GA3, or PAC and grown for 6 days. For root elongation assays, the germinated seedlings were transferred to MS media in square petri dishes with BL or BRZ, GA3, or PAC and grown vertically for 7 days. At least 20 seedlings were measured for each genotype in each set of experiments.

Yeast two-hybrid assays

The full-length BZR1 or RGA cDNA (complementary DNA) was cloned into the pGBKT7 bait vector and transformed into the yeast strain AH109. Yeast cells carrying the bait vector were then transformed with the prey plasmids containing the full-length DELLA fragments or different RGA and BZR1 deletions. Transformants were selected on medium lacking histidine (His) but containing 50 mM (for BZR1) or 80 mM (for RGA) 3-AT, respectively. Primers used for plasmid construction in yeast two-hybrid assays were listed in table S1.

Gene expression analysis by qRT-PCR

Total RNA was extracted from the entire Arabidopsis seedlings with the RNeasy Plant Mini Kit (Qiagen), treated with DNase I (Qiagen), and reverse-transcribed with the SuperScript first-strand synthesis system (Invitrogen). Real-time PCR was performed with the MyiQ real-time system (Bio-Rad) and the iQ SYBR Green Supermix (Bio-Rad). UBC and EF-1a were used as internal controls for BR- and GA-related genes, respectively. Primer sequences of genes tested in qRT-PCR were listed in table S2.

In vitro pull-down, overlay, and the semi–in vivo pull-down assays

GST-RGA and MBP-BZR1 were expressed in the Escherichia coli strain BL21 by induction with 0.3 mM IPTG (isopropyl-β-d-thiogalactopyranoside). The GST and MBP fusion proteins were purified with glutathione Sepharose beads (GE Healthcare) or amylose agarose beads (New England Biolabs), respectively. The in vitro pull-down assay of MBP-BZR1 and GST-RGA interaction follows the previously described procedure (32).

The in vitro overlay assay follows the protocol of Tang et al. (16). The tag-free BIN2 was produced from the BIN2-pPAL7 construct with the Profinity eXact protein purification system (Bio-Rad).

For semi–in vivo pull-down assay, protein extracts from 12-day-old 35S::mBZR1-myc transgenic seedlings treated with or without BL (1 μM BL, 4 hours) were incubated with glutathione Sepharose beads containing GST-RGA or GST for 1 hour at 4°C. The mBZR1-myc proteins pulled down by GST-RGA were detected with an antibody recognizing the myc epitope (Sigma). The primer sequences used for plasmid construction for the in vitro pull-down, overlay, and the semi–in vivo pull-down assays were listed in table S3.

BiFC assay

Full-length RGA and BZR1 cDNAs were cloned into the pBiFC vectors, which contained either N-terminal or C-terminal half of YFP. The resulting constructs were transformed into the Agrobacterium strain GV3101. The detailed protocol follows Gampala et al. (64). The primers used for plasmid construction for the BiFC assay were listed in table S3.

Coimmunoprecipitation assay

Agrobacteria (GV3101) containing the 35S::BZR1-GFP expression vector was infiltrated alone or together with Agrobacteria harboring 35S::RGA-myc into the leaves of tobacco plants. About 40 hours after infiltration, the transformed tobacco leaves were collected for the coimmunoprecipitation assay (64). The BZR1-GFP pulled down by RGA-myc was detected with an antibody recognizing GFP (Clontech) and RGA-myc detected with an antibody recognizing myc (Sigma). The primers used for plasmid construction for the coimmunoprecipitation experiment were listed in table S3.

Cell-free degradation assay

Fourteen-day-old seedlings expressing the 35S::PP2AB′α-YFP and 35S::PP2AB′β-YFP plasmids were harvested, and total proteins were extracted as described by Wang et al. (39). Protein extracts were then incubated with 20 μM MG132 or mock solution at 22°C. Samples were collected at the indicated time intervals for determination of PP2A abundance by Western blotting.

Protoplast transient assays of gene transcription

To test the overall transcriptional activation activity of RGA, full-length RGA cDNA was cloned into the pMN6 vector, and the resulting RGA-pMN6 plasmid was cotransformed with a reporter plasmid pGLL and an internal control vector pRLL into the protoplasts of bzr1-1D or Col-0 plants (46). After incubation at 22°C for 15 hours, the protoplasts were treated with 10−6 M BL for 20 min before samples were collected for luciferase activity assay.

In another set of transient assays, the promoters of SCL3 (1.4 kb) and SAUR-AC (0.78 kb) were each cloned into the pGreen II 0800-LUC vector (49) to generate reporter constructs. Each reporter construct, together with either 35S::BZR1 or 35S::RGA, was cotransformed into Col-0 protoplasts for transcriptional activity assay. In the case of the SAUR-AC promoter, 35S::BIM1 was also included to improve BZR1’s transcriptional activation activity. The signals of Firefly and Renilla luciferase were assayed with the dual luciferase assay reagents (Promega).

Western blotting for DELLA, BZR1, and PP2A abundance

Seedlings were collected from the MS plates and either processed directly or immersed in solutions containing GA3, EtOH, MetOH, PAC, LiCl, KCl, DMSO (dimethyl sulfoxide), or OA. The entire seedlings of 12-day-old Arabidopsis plants were collected and ground into powder in liquid nitrogen. SDS sample buffer (2×) was added in the ratio of 1:1 (1 μl of buffer for 1 mg of tissue powder) to extract the proteins. The extracted proteins were then heated at 70°C for 10 min, followed by centrifugation at 12,000g for 10 min. The resulting supernatants were transferred to a new microfuge tube. SDS–polyacrylamide gel [10% (w/v)] electrophoresis (SDS-PAGE) was performed to resolve the protein extracts. After electrophoresis, proteins were transferred to a PVDF (polyvinylidene difluoride) membrane (Millipore) with a semidry electrophoretic transfer cell (Bio-Rad) and immunodetected with antibodies recognizing GFP (Clontech) or myc (Sigma).

Data processing

Bands on Western blots were quantified with the software ImageJ. All the quantification of blots was done by normalizing each band for loading and then quantified across three independent experiments. The relative integrated density of pBZR1 in the first sample of each blot was set to 1 as a common data point for the experiment. The relative integrated density of other bands on the same blot was calculated as a ratio of their density to the common data point and was shown below the blot. Each experiment was repeated at least three times, and one representative result was shown. Hypocotyl and root lengths were also measured with software ImageJ. For experiments with single pair-wise comparison, the Student’s t test was used to determine the level of significance. For experiments with multiple comparisons, the data were analyzed by ANOVA with Bonferroni’s correction.

Supplementary Materials

Fig. S1. BES1 interacts with RGA in a yeast two-hybrid assay.

Fig. S2. The LHR1 domain of RGA and the BIN2 phosphorylation domain of BZR1 mediate the interaction of RGA and BZR1.

Fig. S3. BiFC analysis reveals a BES1 and RGA interaction in planta.

Fig. S4. The gai mutant (with the GAIΔ17 mutation) partially suppresses root elongation of bzr1-1D.

Fig. S5. RGA and GAI overexpression fails to alter the sensitivity of bzr1-1D to BRZ.

Fig. S6. The gai mutation enhances the dwarf phenotypes of bri1-5 in a root elongation assay.

Fig. S7. The RGA and GAI loss-of-function double mutant (rga-24 gai-t6) has increased sensitivity to BL but reduced sensitivity to BRZ.

Fig. S8. The double mutant rga-24 gai-t6 has reduced sensitivity to BRZ in a hypocotyl elongation assay.

Fig. S9. bzr1-1D reduces and bri1-5 enhances sensitivity to the GA inhibitor PAC.

Fig. S10. bzr1-1D reduces and bri1-5 enhances sensitivity to PAC in a hypocotyl elongation assay.

Fig. S11. DELLAs and BZR1 and BES1 have limited effects on each other’s gene expression.

Fig. S12. Quantitative analysis of the effect of GA or inhibition of GA biosynthesis or of BIN2 or PP2A activity on BZR1 abundance and phosphorylation status.

Fig. S13. DELLAs inhibit growth of BZR1-overexpressing plants.

Fig. S14. RGA and BIN2 do not interact.

Fig. S15. RGA does not compete with PP2A for phosphorylated BZR1.

Fig. S16. Inhibition of GA signaling stimulates the expression of PP2AB′α and PP2AB′β.

Fig. S17. GA signaling inhibits SCL3 expression.

Fig. S18. Some BZR1 target genes and RGAΔ17 responsive genes are common.

Table S1. Primers used for plasmid construction in yeast two-hybrid assays.

Table S2. Primers used for RT-PCR and qRT-PCR analyses.

Table S3. Primers used for plasmid construction in other experiments.

References and Notes

Acknowledgments: We thank Z.-Y. Wang for the pBiFC vectors; W. Tang for the BIN2-pPAL7 construct; W. A. Laing for the pGreenII-0800-LUC vector; T. Nakagawa for the pGWB gateway vectors; Y. Sun for the seeds of 35S::mBZR1-Myc, 35S::PP2AB′ α-YFP, and 35S::PP2AB′ β-YFP plants; and X. W. Deng for the 35S::TAP-RGL3 seeds. The seeds of gai, rga-24 gai-t6, 35S::TAP-RGA, 35S::TAP-GAI, 35S::TAP-RGL1, 35S::TAP-RGL2, and pRGA::GFP-RGA were ordered from Arabidopsis Biological Resource Center (Ohio). Funding: This work was supported by the Hong Kong RGC General Research Fund (CUHK codes 465009 and 465410), a grant from the National Natural Science Foundation of China (no. 91125027), and the Direct Grants from the Chinese University of Hong Kong (to J.-X.H.) and by the National Transgenic Project of China (no. 2009ZX08009-001B), the Hong Kong UGC AoE Center for Plant and Agricultural Biotechnology Project AoE-B-07/09, and a special fund from the Resource Allocation Committee, The Chinese University of Hong Kong (to S.S.M.S.). Author contributions: Q.-F.L., C.W., L.J., and S.L. performed the experiments. Q.-F.L., S.S.M.S., and J.-X.H. designed the experiments, analyzed the data, and wrote the manuscript. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article