Nucleocytoplasmic trafficking and turnover mechanisms of BRASSINAZOLE RESISTANT1 in Arabidopsis thaliana

Significance

Upon brassinosteroid (BR) treatment, BRASSINAZOLE RESISTANT1 (BZR1) is dephosphorylated and accumulate in the nucleus. However, it was unclear whether these two processes are interdependent. Here, we show that BR signaling first recruits phosphorylated BZR1 to the nucleus where it is dephosphorylated by protein phosphatase 2A (PP2A). BZR1 is constitutively degraded via a BR- and 26S proteosome–independent mechanism; however, BR treatment activates a different 26S proteosome–dependent pathway and increases BZR1 turnover rate. These results help to untangle the intertwined mechanisms governing BZR1 phosphorylation/dephosphorylation, nucleocytoplasmic trafficking, and protein turnover during BR signaling and clarify the role of PP2A in the BR signaling pathway. Our results will inspire new studies on BR signaling and in other fields of plant research.

Abstract

Regulation of the nucleocytoplasmic trafficking of signaling components, especially transcription factors, is a key step of signal transduction in response to extracellular stimuli. In the brassinosteroid (BR) signal transduction pathway, transcription factors from the BRASSINAZOLE RESISTANT1 (BZR1) family are essential in mediating BR-regulated gene expression. The subcellular localization and transcriptional activity of BZR1 are tightly regulated by reversible protein phosphorylation; however, the underlying mechanism is not well understood. Here, we provide evidence that both BZR1 phosphorylation and dephosphorylation occur in the nucleus and that BR-regulated nuclear localization of BZR1 is independent from its interaction with, or dephosphorylation by, protein phosphatase 2A. Using a photoconvertible fluorescent protein, Kaede, as a living tag to distinguish newly synthesized BZR1 from existing BZR1, we demonstrated that BR treatment recruits cytosolic BZR1 to the nucleus, which could explain the fast responses of plants to BR. Additionally, we obtained evidence for two types of protein turnover mechanisms that regulate BZR1 abundance in plant cells: a BR- and 26S proteosome–independent constitutive degradation mechanism and a BR-activated 26S proteosome–dependent proteolytic mechanism. Finally, treating plant cells with inhibitors of 26S proteosome induces the nuclear localization and dephosphorylation of BZR1, even in the absence of BR signaling. Based on these results, we propose a model to explain how BR signaling regulates the nucleocytoplasmic trafficking and reversible phosphorylation of BZR1.

Through various signal transduction mechanisms, living cells sense changes in their immediate environment and respond accordingly. During signal transduction, two physical barriers must be crossed: the plasma membrane, which separates the internal cellular components from the outside, and the nuclear envelope, which separates an organism’s genetic information from the cytosolic contents. Many studies have focused on revealing how cells perceive and transmit extracellular signals across their plasma membrane to induce appropriate cellular responses. However, much less attention has been paid to understanding how extracellular signals regulate the nucleocytoplasmic trafficking of signaling components, especially transcription factors.

Brassinosteroids (BRs) are essential hormones in regulating many plant growth and developmental processes as well as plant responses to environmental signals. BRs are perceived by the plasma membrane–localized receptor kinase BRASSINOSTEROID INSENSITIVE1 (BRI1) and its co-receptor BRI1-ASSOCIATED KINASE1 (BAK1) (1–3). BR-activated BRI1 then inactivates and promotes the degradation of BRASSINOSTEROID INSENSITIVE2 (BIN2), a cellular protein kinase, via sequential phosphorylation and dephosphorylation (4–7). When the BR signaling pathway is inactive, BIN2 phosphorylates BRASSINAZOLE RESISTANT1 (BZR1) family transcription factors (BZRs) (8, 9), thereby preventing them from regulating the expression of their target genes (10). With the help of 14–3-3 proteins, BIN2 phosphorylation also promotes the export of BZR1 from the nucleus (11), resulting in the accumulation of phosphorylated BZR1 in the cytosol. Upon activation of BR signaling, BIN2 is inactivated and degraded, promoting the dephosphorylation of BZR1 by protein phosphatase 2A (PP2A) (12). Dephosphorylated BZR1 accumulates in the nucleus, where it binds to the promoters of downstream BR-regulated genes, modulating their transcription (13).

The regulation of BZRs dephosphorylation and nuclear localization is therefore a key feature of BR signaling, and these two phenomena are closely related. For example, BIN2-mediated phosphorylation of BZR1 at Ser-173 initiates an interaction between BZR1 and 14–3-3 proteins that promotes the nuclear export of BZR1. Substituting Ser-173 with alanine (BZR1^S173A) or treating seedlings with an inhibitor that disrupts the interaction between 14–3-3 proteins and its target protein increases the nuclear localization of BZR1 (11). However, it is currently unknown how BR signaling recruits cytosolic BZR1 into the nucleus and whether this process depends on the dephosphorylation of BZR1.

In this study, we used transgenic Arabidopsis (Arabidopsis thaliana) seedlings that express various mutated versions of BZR1 and a photoconvertible fluorescent protein tag to distinguish existing BZR1 from newly synthesized BZR1 to reveal the mechanisms governing the turnover and nucleocytoplasmic trafficking of BZR1. Our findings clarified the intertwined means by which BRs regulate nuclear localization and dephosphorylation of BZR1 and helped understanding of the mechanism regulating the proteolysis of BZR1, which together should inspire new studies in the field.

Results

BZR1 Phosphorylation and Dephosphorylation Occur in the Nucleus.

BRs induce the dephosphorylation and nuclear accumulation of BZR1. However, it is unclear whether phosphorylated BZR1 is first dephosphorylated in the cytosol and then translocated into the nucleus or whether BRs first induce the nuclear import of phosphorylated BZR1 and then promote its dephosphorylation in the nucleus. Previously, we showed that nuclear-localized PP2A positively regulates BR signaling by dephosphorylating BZR1, while cytosolic PP2A negatively regulates BR signaling by dephosphorylating BRI1 (12, 14). Given these results, we hypothesized that phosphorylated BZR1 might be dephosphorylated in the nucleus.

To test this hypothesis, we first analyzed the BZR1 protein sequence online (http://www.moseslab.csb.utoronto.ca/NLStradamus/) for the presence of nuclear localization signals (NLSs) and identified one such sequence between amino acids 20 and 43 (²⁰AARRKPSWRERENNRRRERRRRAV⁴³). Next, we generated transgenic plants expressing BZR1 variants as fusions with the yellow fluorescent protein (YFP), driven by the BZR1 promoter or the cauliflower mosaic virus (CaMV) 35S promoter: the 14–3-3 binding deficient form BZR1pro:BZR1^S173A-YFP; BZR1pro:BZR1^TTTT-YFP, in which four arginine residues in the putative NLS of BZR1 were mutated to threonine (RRRERRRR to RTTETRTR); and 35Spro:BZR1^∆PEST-YFP, in which the PP2A-binding PEST domain was removed. All seedlings were grown on growth medium containing 0.2 μM of the BR biosynthesis inhibitor propiconazole (PCZ) for 1 wk and then treated with 1 μM epi-brassinolide (eBL) for 30 min. Immunoblot analysis showed that wild-type BZR1-YFP was phosphorylated and localized to the cytosol in the absence of BR. BR treatment led to dephosphorylation of BZR1-YFP and its accumulation in the nucleus (Fig. 1 A and E). In comparison, BZR1^S173A-YFP constitutively localized to the nucleus, while BZR1^TTTT-YFP constitutively accumulated in the cytosol, regardless of BR treatment (Fig. 1 B and C). Moreover, like wild-type BZR1-YFP, BZR1^∆PEST-YFP moved from the cytosol to the nucleus in response to BR (Fig. 1D). Because BZR1^∆PEST cannot interact with PP2A (12), this result suggests that the BR-regulated nuclear translocation of BZR1 is independent of its interaction with PP2A.

Fig. 1.

BZR1 phosphorylation and dephosphorylation take place in the nucleus. (A–D) Confocal microscopy examination of the subcellular localization of BZR1-YFP, BZR1^S173A-YFP, BZR1^TTTT-YFP, and BZR1^∆PEST-YFP. One-week-old transgenic seedlings grown on 1/2 MS medium containing 0.2 μM PCZ were visualized directly for YFP fluorescence or treated with 1 μM eBL for 30 min and then visualized. (Scale bars, 20 μm.) (E) Immunoblots of BZR1-YFP (BZR1), BZR1^S173A-YFP (S173A), BZR1^TTTT-YFP (TTTT), and BZR1^∆PEST-YFP (∆PEST) using anti-GFP antibody and protein extracts from the whole seedlings used in A–D. pBZR1: phosphorylated BZR1; BZR1: nonphosphorylated BZR1.

Immunoblotting showed that the predominantly nucleus-localized BZR1^S173A-YFP still exhibits a wild-type–like phosphorylation/dephosphorylation mobility shift in response to BR. By contrast, BZR1^TTTT-YFP and BZR1^∆PEST-YFP proteins failed to exhibit the BR-dependent phosphorylation and dephosphorylation cycle characteristic of BZR1. BZR1^TTTT-YFP consistently migrated to the same position as dephosphorylated BZR1-YFP, regardless of BR stimulation (Fig. 1E). Given that BZR1^TTTT-YFP and BZR1-YFP differ in only four amino acids and that BZR1^TTTT-YFP retained the ability to interact with and be phosphorylated by BIN2 (SI Appendix, Fig. S1), our confocal and immunoblot analyses suggested that the constitutively cytosolic form of BZR1^TTTT-YFP remains nonphosphorylated even in the absence of BR signaling. As previously reported (12), BR treatment did not promote the dephosphorylation of BZR1^∆PEST-YFP, possibly because this BZR1 variant cannot interact with PP2A (Fig. 1E).

Next, we generated transgenic plants harboring 35Spro:BZR1^∆NLS-YFP, which accumulate a truncated version of BZR1 (BZR1^∆NLS) lacking the N-terminal 41 amino acids. As with BZR1^TTTT-YFP, BZR1^∆NLS-YFP consistently localized to the cytosol of root cells in the presence of BR. Similarly, BZR1^∆NLS-YFP can be phosphorylated by BIN2, but it remains nonphosphorylated when the BR signal is not turned on (SI Appendix, Fig. S2). Taken together, these results suggest that both BZR1 phosphorylation and dephosphorylation steps take place in the nucleus; upon activation of the BR pathway, cytosolic phosphorylated BZR1 translocates to the nucleus, where it is then dephosphorylated by nucleus-localized PP2A.

BR Treatment Induces the Relocation of Cytosolic BZR1 to the Nucleus.

BR treatment promotes the nuclear accumulation of BZR1. It remains unclear, however, whether this nuclear accumulation reflects existing cytosolic BZR1, de novo translated BZR1, or both. Determining this requires methods to distinguish newly biosynthesized BZR1 (following BR treatment) from preexisting BZR1 (from before BR treatment) and to separately track each pool of BZR1 inside the cell.

For this purpose, we used Kaede, a photoconvertible fluorescent protein isolated from the stony coral Trachyphyllia geoffroyi. In its baseline state (Kaede^green), Kaede emits green fluorescence when excited at 488 nm. Irradiation of Kaede^green with ultraviolet (UV) light at 405 nm induces the cleavage of three amino acids (His⁶²-Tyr-Gly⁶⁴) from the protein and irreversibly converts it to a red fluorescent variant (Kaede^red) with an excitation wavelength at 543 nm (15). This photocontrollable character means Kaede is a powerful tool for studying organelle fusion events and tracking protein dynamics (16, 17).

Accordingly, we generated transgenic plants harboring a BZR1pro:BZR1-Kaede-FLAG transgene. Confocal microscopy and immunoblotting showed that, like wild-type BZR1-YFP, the BZR1-Kaede^green fusion protein was phosphorylated and localized to the cytosol in the absence of BR but was dephosphorylated and localized to the nucleus after 30 min of BL treatment (Fig. 2A and SI Appendix, Fig. S3). When seedlings were excited at 543 nm, only background fluorescence was observed (Fig. 2B). After irradiating seedlings with a UV light at 405 nm for 5 min, the BZR1-Kaede^green fluorescence signal decreased sharply (Fig. 2C) and a strong BZR1-Kaede^red fluorescence signal developed (Fig. 2D). Because BZR1-Kaede^red localized to the cytosol (Fig. 2D), this indicated that 5 min of UV irradiation does not alter the subcellular localization of BZR1-Kaede. Therefore, we next treated UV-irradiated seedlings with 1 μM eBL for 30 min: We observed strong accumulation of both green (a mixture of residual nonconverted and newly synthesized BZR1-Kaede^green protein) and red (converted BZR1-Kaede^red protein only from before BR treatment) fluorescence signals in the nucleus (Fig. 2E and F). This experiment demonstrates that cytosolic BZR1 can be recruited to the nucleus following activation of the BR pathway.

Fig. 2.

BR recruits cytosolic BZR1 to the nucleus. Confocal microscopy examination of the subcellular localization of BZR1-Kaede. Transgenic seedlings harboring BZR1pro:BZR1-Kaede-FLAG were grown on 1/2 MS medium containing 0.2 μM PCZ for 1 wk (−BR). The fluorescence signal of BZR1-Kaede was then examined using excitation at 488 nm and 543 nm, before (−UV) (A and B) and after UV illumination at 405 nm for 5 min (C and D). After UV conversion of Kaede^green to Kaede^red, the same seedling was treated with 1 μM eBL for 30 min and Kaede fluorescence was observed (E and F). (Scale bars, 20 μm.)

BZR1 Protein Degradation is Promoted by BRs.

Monitoring the change in signal intensity of BZR1-Kaede^green and the half-life of BZR1-Kaede^red allowed us to investigate de novo protein biosynthesis and protein degradation of BZR1 simultaneously and in real time, providing invaluable information about in vivo BZR1 protein dynamics in plant cells. Using 7 d old seedlings grown on PCZ-containing medium, we first examined the fluorescence signal derived from BZR1-Kaede excited at 488 nm and 543 nm before and after UV irradiation (Fig. 3 A and B). After UV conversion, we returned the seedlings to a medium containing 0.2 μM PCZ for different time intervals and monitored fluorescence intensity upon excitation at 488 nm and 543 nm. BZR1-Kaede^red fluorescence signal gradually decreased, while BZR1-Kaede^green fluorescence continually increased, indicating the constitutive degradation of existing BZR1-Kaede^red and the accumulation of newly translated or maturated BZR1-Kaede^green in the absence of BR (Fig. 3 A and C). In parallel, we also transferred seedlings to medium containing 1 μM eBL after UV conversion for various time intervals. Surprisingly, as the length of the BR treatment increased, the corresponding fluorescence signal from BZR1-Kaede^red decreased, while the fluorescence signal from BZR1-Kaede^green did not rise as high as it had in seedlings treated with 0.2 μM PCZ (Fig. 3 B and C). Continuously multiple scans of the same root (SI Appendix, Fig. S4) and a one-time scan experiment (SI Appendix, Fig. S5) showed the gradual decrease in BZR1-Kaede^red fluorescence was not caused by photobleaching associated with repeated laser scanning. Together, these data suggest that BZR1-Kaede undergoes constitutive degradation, even in the presence of BR. Based on the half-life of BZR1-Kaede^red fluorescence signal, BR treatment appears to promote BZR1 degradation (Fig. 3C).

Fig. 3.

BZR1 protein turnover is promoted by BR. (A and B) Transgenic seedlings harboring BZR1pro:BZR1-Kaede-FLAG were grown on 1/2 MS medium containing 0.2 μM PCZ (−BR) for 1 wk under LD conditions. BZR1-Kaede fluorescence was monitored before (−UV) and after UV illumination, under excitation wavelengths of 488 nm and 543 nm. After conversion of Kaede^green to Kaede^red, the seedlings were transferred to a new 1/2 MS medium containing 0.2 μM PCZ (−BR) (A) or 1 μM eBL (+BR) (B). After the indicated time, the Kaede fluorescence signal from the same root was examined. (Scale bars, 20 μm.) (C) Quantification of BZR1-Kaede fluorescence signal following the treatments described in A and B. (D) Quantification of the half-life of the BZR1-YFP fluorescence signal after the seedlings were transferred to a fresh 1/2 MS medium containing various chemicals and treated for indicated times. In C and D, error bars indicate mean ± SD, and statistical differences are indicated by ns (not significant), *P < 0.05, **P < 0.01, and ***P < 0.001 (Student’s t test, two tailed).

The finding that BR promotes BZR1 degradation was surprising, as BRs are generally considered to stabilize BZR1 protein (8). To independently validate this result, we transferred 1-wk-old BZR1pro:BZR1-Kaede-FLAG and BZR1pro:BZR1-YFP transgenic seedlings grown on a medium containing 0.2 μM PCZ to fresh medium containing eBL and treated for various time intervals. Real-time quantitative PCR (qPCR) showed the BZR1-YFP transcript remained unchanged or slightly increased after BR treatment (SI Appendix, Fig. S6A), while immunoblots showed that abundance of both BZR1-Kaede and BZR1-YFP proteins significantly decreased after treatment with 1 μM eBL for 4 h or longer (SI Appendix, Fig. S6 B–E and G). We also monitored yellow fluorescence signal in the root cells of BZR1pro:BZR1-YFP transgenic seedlings after transfer to eBL-containing medium, either alone or with 200 μM of the protein synthesis inhibitor cycloheximide (CHX). Again, we observed a continuous, gradual decrease in fluorescence derived from BZR1-YFP upon extended BR treatment, which was accelerated by CHX treatment. Quantifying the fluorescence signal indicated that the half-life of BZR1-YFP is around 2 h in the presence of eBL and CHX, which is close to that determined for BZR1-Kaede^red after BR treatment (Fig. 3 C and D).

To rule out the possibility that YFP- and Kaede-tagged BZR1 respond differently to BR than the endogenous BZR1 protein, we examined the protein stability of endogenous BZR1. Like BZR1-Kaede and BZR1-YFP, endogenous BZR1 protein levels continued to decrease with prolonged BR treatment (SI Appendix, Fig. S6 F and G). Together, these results support the finding that BR promotes BZR1 degradation, not BZR1 stabilization.

BZR1 Proteolysis and BZR1 Nuclear Localization Are Regulated by 26S Proteasome.

To investigate whether BZR1 degradation is mediated by the 26S proteasome, we treated UV light–converted seedlings harboring the BZR1pro:BZR1-Kaede-FLAG transgene with 50 μM of the proteasome inhibitor MG132 in the presence of 0.2 μM PCZ or 1 μM eBL and examined the fluorescence signal of Kaede^green and Kaede^red in the same seedlings over time. Treating seedlings with MG132 when the BR pathway is inactive (+PCZ) led to a slight but insignificant delay in the degradation of BZR1-Kaede^red, as indicated by a reduction in Kaede^red fluorescence signal (Fig. 4A and SI Appendix, Fig. S7). In comparison, BR-promoted BZR1-Kaede^red degradation appeared to be more sensitive to MG132 treatment (Fig. 4A andSI Appendix, Fig. S8). To confirm this result, we treated 1-wk-old transgenic seedlings harboring the transgenes BZR1pro:BZR1-β-glucuronidase (GUS), BES1pro:BES1-GUS or BEH4pro:BEH4-GUS grown on PCZ-containing medium with 50 μM MG132 in the presence of PCZ or eBL. Indeed, the GUS histochemical staining signal decreased in all transgenic seedlings treated with eBL for 3 h. This BR-induced reduction in GUS signal was alleviated by MG132 (SI Appendix, Fig. S9). Taken together, these results suggest that BRs promote the degradation of BZR1 via the 26S proteasome proteolytic pathway.

Fig. 4.

MG132 treatment regulates the turnover and nucleocytoplasmic trafficking of BZR1. (A) Quantification of fluorescence signal of BZR1-Kaede^red after seedlings were transferred to fresh medium containing various chemicals and treated for the indicated times. Error bars indicate mean ± SD, and statistical differences are indicated by ns (not significant), *P < 0.05, **P < 0.01, and ***P < 0.001 (Student’s t test, two tailed). (B and C) Confocal microscopy images of the subcellular localization patterns of BZR1-Kaede^green (B), BZR1-YFP or BZR1^∆NLS-YFP (C) in the absence of BR (−BR) after treatment with DMSO (mock) or MG132 for 7 h. (D) Immunoblot analysis of BZR1-Kaede-FLAG in the seedlings shown in A. M: +MG132, D: +DMSO. (E) qPCR analysis of the expression of CPD and DWF4 in the seedlings used in A (n = 3). Error bars indicate mean ± SD. Statistically significant differences are indicated by different lowercase letters (P < 0.05, two-way ANOVA). (F) A proposed model of the regulation of BZR1 phosphorylation/dephosphorylation and BZR1 nucleocytoplasmic trafficking in response to BR signaling. (Scale bars in B and C, 20 μm.)

Unexpectedly, we also discovered that MG132 treatment induces the accumulation of both BZR1-Kaede^green and BZR1-Kaede^red in the nucleus, even in the presence of PCZ (Fig. 4B and SI Appendix, Figs. S7B and S10). To rule out the possibility that it is an artifact of Kaede fusion, we repeated the experiment with BZR1pro:BZR1-YFP transgenic seedlings and confirmed that nuclear accumulation of BZR1-YFP can be induced by the treatments with MG132 and MG115, a different proteasome inhibitor (Fig. 4C and SI Appendix, Fig. S11A). The fact that MG132 induced the nuclear localization of BZR1-YFP but not BZR1^∆NLS-YFP (Fig. 4C) suggests that the NLS in BZR1 is critical for the change in BZR1 localization. After MG132 and MG115 treatment, immunoblotting revealed that BZR1-Kaede and BZR1-YFP were slightly dephosphorylated (Fig. 4D and SI Appendix, Fig. S11B), and qPCR showed relative transcript levels of DWARF4 (DWF4) and CONSTITUTIVE PHOTOMORPHOGENIC DWARF (CPD), two BR biosynthesis genes down-regulated by BR-activated BZR1, were reduced (Fig. 4E). Taken together, these results suggest that the nuclear localization of BZR1 induced by MG132 treatment partially activates BZR1, possibly via the action of nucleus-localized PP2A (14).

Discussion

BZRs are plant-specific transcription factors with essential roles in mediating BR-regulated gene expression. These roles have been elucidated through a combination of genetic, biochemical, and molecular biological approaches. BRs regulate the transcriptional activity and nucleocytoplasmic trafficking of BZRs via protein phosphorylation/dephosphorylation. However, the underlying mechanism is not well understood. Using confocal microscopy and immunoblot analyses, we discovered that 1) BR treatment recruits cytosolic BZR1-Kaede^red to the nucleus, 2) a constitutively nucleus-localized form of BZR1 (BZR1^S173A) is dephosphorylated when BR signaling is activated and phosphorylated in the absence of BR, 3) constitutively cytosolic forms of BZR1 (BZR1^TTTT and BZR1^∆NLS) remain nonphosphorylated, regardless of activation of BR signaling, and 4) BZR1^∆PEST, which cannot interact with PP2A, is phosphorylated and localizes to the cytosol in the absence of BR signaling but migrates to the nucleus and remains phosphorylated when BR signaling is active. Together, these results suggest that the recruitment of cytosolic BZR1 to the nucleus by BRs is not dependent on PP2A binding or PP2A-mediated BZR1 dephosphorylation. Instead, both the phosphorylation and dephosphorylation of BZR1 occur in the nucleus. These findings concur with a previous study showing that only nucleus-localized PP2A positively regulates BR signaling by dephosphorylating BZR1 (14). Our results also concur with another study indicating that a form of BIN2 targeted to the nucleus through the addition of a heterologous NLS was more efficient in shutting down BR signaling than wild-type BIN2, possibly through an ability to phosphorylate BZR1 and other BZRs in the nucleus (10).

We were surprised to discover that newly translated BZR1-Kaede^green accumulates more rapidly in the absence than in the presence of BR (Fig. 3), suggesting that de novo BZR1 biosynthesis or maturation of newly translated BZR1-Kaede^green protein is partially inhibited by BR signaling. These results explain why the pp2ab′;α pp2ab′β double mutant, which lacks two of the B′ subunits of PP2A that are responsible for BZR1 dephosphorylation in the nucleus, does not show the typical dwarf or semidwarf phenotypes normally seen in other BR signaling mutants (12). Newly synthesized and nonphosphorylated BZR1 can probably move into the nucleus of the pp2ab′α pp2ab′β mutant, which is independent of PP2A binding, and regulate the expression of downstream target genes when BR signal is turned on.

Using multiple approaches, we discovered two mechanisms regulating BZR1 protein degradation: one is independent of the 26S proteasome and constitutively degrades BZR1 regardless of the status of the BR signal and the second, a 26S proteasome-dependent mechanism, is evoked after prolonged (>1 h, Fig. 3C) BR treatment. These results contradict a previous study, which reported that in the absence of BR signaling, BZR1 undergoes phosphorylation and degradation via the 26S proteasome proteolytic pathway, and BR treatment dephosphorylates and stabilizes BZR1 (8). Different methods used to reach each conclusion might explain this discrepancy. Using immunoblots, He et al. (8) showed that MG132 and BR treatment promote the accumulation of BZR1 proteins. As phosphorylated BZR1 is prone to degradation during sample preparation, the differences in BZR1 abundance before and after BR treatment might have been caused by differences in sample handling. Indeed, BZR1 protein levels have been reported to increase, decrease, or remain constant after BR treatments (10, 18, 19). Alternatively, BZR1 may be phosphorylated at multiple sites (up to 25 putative BIN2 phosphorylation sites) in the absence of a BR signal, and these phosphorylated BZR1 forms might then migrate to different positions in the SDS-PAGE gel and diffuse the overall immunoblot signal of phosphorylated BZR1 (20). By contrast, BR treatment promotes the dephosphorylation of BZR1, thus focusing most of BZR1 to a single molecular weight range; this might explain why we typically see stronger BZR1 immunoblot signals after short (<2 h) BR treatments.

It seems like a paradox that BR activates BZR1 but at the same time promotes BZR1 degradation. A similar mechanism has been reported for transcription factors NPR1 and FIT in Arabidopsis. Pathogen challenge and iron deficiency promote nucleus localization of NPR1 and stimulate the expression of FIT, respectively, which is required for the induction of the expression of system-acquired resistance–related or iron-uptake machinery–related genes. However, pathogen challenge and iron deficiency could also accelerate the turnover of both proteins (21, 22). This stimulus-promoted degradation of activated transcription factors mechanism was thought to continuously deliver “fresh” transcription factor to the gene promoter and was necessary for maintaining high-transcription efficiency of the target genes (21, 23).

Besides slowing down BR-promoted BZR1 degradation, MG132 and MG115 treatment also recruited cytosolic BZR1 to the nucleus in the absence of BR, a phenomenon requiring the intact NLS of BZR1 (Fig. 4C). In eukaryotes, NLS-dependent protein import into the nucleus is usually mediated by proteins of the importin family (24). The Arabidopsis genome encodes nine importin-α (IMPA) and 18 importin-β (IMPB) proteins (25). In a yeast two-hybrid assay, both BZR1 and BES1 interact directly with different IMPAs and IMPBs (SI Appendix, Fig. S12). However, it is unknown whether BRs regulate the stability of these importins and whether these importins compete with 14–3-3 proteins to regulate the nucleocytoplasmic trafficking of phosphorylated BZR1.

Based on our results, we propose a nucleocytoplasmic trafficking mechanism to explain the regulation of BZR1 phosphorylation/dephosphorylation and subcellular localization by BRs (Fig. 4F). In this model, when in vivo BR levels are low, BZR1 abundance inside the cell is balanced by de novo protein biosynthesis and 26S proteasome–independent degradation. Newly translated BZR1 is first imported into the nucleus where it is phosphorylated by BIN2. In turn, this initiates the interaction between BZR1 and 14–3-3 proteins and promotes the export of phosphorylated BZR1 from the nucleus with the assistance of 14–3-3s via an unknown mechanism. We also postulate that a currently unknown regulator X, which controls the nuclear import of phosphorylated BZR1, is degraded by the 26S proteasome in low-BR conditions, allowing phosphorylated BZR1 to accumulate in the cytosol. Activation of the BR signaling pathway inactivates BIN2 and promotes its degradation by the 26S proteasome. At the same time, BR signaling stabilizes regulator X, which might then compete with 14–3-3s to interact with phosphorylated BZR1 and help recruit cytosolic phosphorylated BZR1 to the nucleus where it is dephosphorylated by PP2A in the absence of BIN2. However, the recruitment of phosphorylated BZR1 back to the nucleus is not essential for BR signaling because nonphosphorylated BZR1 can be produced continuously, even in the absence of BR. This de novo translated BZR1 can be imported into the nucleus where it regulates the expression of downstream target genes when the activity and protein abundance of BIN2 are reduced by BR. Because de novo BZR1 biosynthesis takes time, recruitment of cytosolic phosphorylated BZR1 to the nucleus and PP2A-mediated dephosphorylation of BZR1 in the nucleus might represent a fast-response mechanism for plant cells when BRs are first perceived.

This model incorporates most of the key known features of BR signaling (11, 12, 14) and the data generated from this study. Several questions remain unanswered. For example, what is regulator X? Where and how can this model integrate the recent discovery that BZR1 sumoylation promotes the nuclear accumulation of BZR1 and stabilizes BZR1 (26)? These points will require thorough exploration in future studies.

Materials and Methods

Plant Materials and Growth Conditions.

All seedlings were first grown on half-strength Murashige and Skoog (1/2 MS, pH 5.7) agar medium containing 0.2 μM of the brassinosteroid biosynthesis inhibitor PCZ in a growth chamber at 22 °C under long-day (LD) conditions (16 h light and 8 h dark) for 7 d before observation under the microscope or further treatment.

Microscopy.

To determine the correlation between the subcellular localization of BZR1 and its phosphorylation status, primary roots from 1-wk-old seedlings grown on PCZ-containing 1/2 MS medium were mounted in a drop of 0.2 μM PCZ water solution (with or without 1 mg/mL propidium iodide), and their fluorescence was observed using a LSM 710 confocal laser scanning microscope (Zeiss, Jena, Germany). For BL and MG132 treatments, seedlings were transferred to a new 1/2 MS agar plate containing 0.2 μM PCZ, 1 μM eBL, 50 μM MG132 for the indicated times, mounted in a drop of solution containing the same concentration of PCZ, eBL, MG132 for confocal microscopy, or harvested for immunoblots.

For the study of BR- and/or MG132-regulated nucleocytoplasmic trafficking and turnover of BZR1-Kaede, BZR1pro:BZR1-Kaede-FLAG transgenic seedlings were first grown on 0.2 μM PCZ-containing agar medium for 1 wk. Primary roots were mounted in a drop of 0.2 μM PCZ water solution, and the subcellular localization pattern of BZR1-Kaede^green was determined using a Fluoview FV3000 confocal microscope (Olympus, Tokyo, Japan) with an excitation wavelength of 488 nm (5% power with a 700 V master gain). While on the microscope, the same root was then scanned continuously at 405 nm (30% power with a speed of 10 μs/pixel) for 3 to 5 min to convert Kaede^green to Kaede^red. Immediately after the conversion, Kaede^red fluorescence signal was observed using an excitation wavelength of 543 nm (5% power with a 600 V master gain). After image acquisition, UV-converted seedlings were carefully retrieved and placed onto 1/2 MS agar medium containing 0.2 μM PCZ or 1 μM eBL (with or without 50 μM MG132).

For time course quantification of Kaede fluorescence signal, the same seedling roots were repeatedly returned to agar medium after image acquisition at the indicated time intervals. Then they were being mounted in a drop of solution containing the same concentration of PCZ or eBL (with or without 50 μM MG132) as in the agar medium to image Kaede^green and Kaede^red fluorescence, using the same microscope settings.

Data Availability

All study data are included in the article and/or supporting information.