Recent Advances in the Regulation of Brassinosteroid Signaling and Biosynthesis Pathways

Authors


Current address: Department of Genetics, Harvard Medical School, 185 Cambridge St, CPZN7250, Boston MA 02114-2790, USA

Corresponding author Tel: +1 515 294 4816; Fax: +1 515 294 5256; E-mail: yin@iastate.edu

Abstract

inline imageYanhai Yin
(Corresponding author)

Brassinosteroids (BRs) play important roles in plant growth, development and responses to environmental cues. BRs signal through plasma membrane receptor BRI1 and co-receptor BAK1, and several positive (BSK1, BSU1, PP2A) and negative (BKI1, BIN2 and 14–3-3) regulators to control the activities of BES1 and BZR1 family transcription factors, which regulate the expression of hundreds to thousands of genes for various BR responses. Recent studies identified novel signaling components in the BR pathways and started to establish the detailed mechanisms on the regulation of BR signaling. In addition, the molecular mechanism and transcriptional network through which BES1 and BZR1 control gene expression and various BR responses are beginning to be revealed. BES1 recruits histone demethylases ELF6 and REF6 as well as a transcription elongation factor IWS1 to regulate target gene expression. Identification of BES1 and BZR1 target genes established a transcriptional network for BR response and crosstalk with other signaling pathways. Recent studies also revealed regulatory mechanisms of BRs in many developmental processes and regulation of BR biosynthesis. Here we provide an overview and discuss some of the most recent progress in the regulation of BR signaling and biosynthesis pathways.

Introduction

A group of plant steroid hormones, named Brassinosteroids (BRs), regulate many processes in plant growth and development, including cell elongation, cell division, senescence, vascular differentiation, reproduction, photomorphogenesis and, responses to various stresses (Mandava 1988; Clouse and Sasse 1998; Li and Chory 1999; Divi and Krishna 2009). Loss-of-function mutants, deficient in BR biosynthesis or perception, display dwarf phenotypes with reduced cell elongation, dark-green and epinastic leaves, reduced apical dominance, delayed flowering and senescence, altered vascular patterning and male sterility (Clouse et al. 1996; Li et al. 1996; Szekeres et al. 1996). Genetic and molecular studies in Arabidopsis have greatly advanced our understanding of the BR signaling pathway (Figure 1) (Kim and Wang 2010; Li 2010b; Clouse 2011; Yang et al. 2011). BRs are perceived by a membrane-bound receptor called BRI1 (BRASSINOSTEROID INSENSITIVE 1), a leucine rich repeat (LRR) receptor like kinase (Li and Chory 1997; He et al. 2000; Wang et al. 2001; Kinoshita et al. 2005). Many alleles of BRI1 have been identified through various genetic screens (Clouse et al. 1996; Li and Chory 1997; Noguchi et al. 1999; Friedrichsen et al. 2000; Zhao et al. 2009). In the absence of BRs, BRI1's activity is inhibited by the negative regulator, BKI1 (BRI1 KINASE INHIBITOR 1), which exerts its effect by binding the C-terminal tail of BRI1 (Wang and Chory 2006). However, binding of BRs to BRI1 relieves this repression and results in BRI1 association with the co-receptor BAK1 (BRI1-Associated Receptor Kinase 1) and a series of phosphorylation events (Li et al. 2002; Nam and Li 2002; Wang et al. 2005a, 2005b, 2008; Oh et al. 2009b). BRI1 signals through BSK1 (BR SIGNALING KINASE 1) and BSU1 (BRI1-SUPPRESSOR 1) to regulate a negative regulator BIN2 (BRASSINOSTEROID INSENSITIVE 2), a GSK3/Shaggy-like kinase (Choe et al. 2002; Li and Nam 2002; Pérez-Pérez et al. 2002; Mora-Garcia et al. 2004; Tang et al. 2008b; Kim et al. 2009). Other BRI1 substrates such as Arabidopsis TRIP-1 (TGF-β RECEPTOR-INTERACTING PROTEIN-1) and TTL (TRANSTHYRETIN-LIKE PROTEIN) may function to regulate BR signaling as well (Nam and Li 2004; Ehsan et al. 2005). The dephosphoryaltion of BIN2 by BSU1 inhibits BIN2 function, which allows the nuclear accumulation of two transcription factors BES1 (BRI1-EMS SUPPRESSOR 1) and BZR1 (BRASSINAZOLE RESISTANT 1), two bHLH (BASIC HELIX-LOOP-HELIX)-like transcription factors that play major roles in regulating BR target gene expression (Wang et al. 2002; Yin et al. 2002b, 2005; Zhao et al. 2002; He et al. 2005).

Figure 1.

A model for brassinosteroid signaling. Brassinosteroid (BR) are perceived by membrane localized BRI1 receptor-like kinase. In the absence of BRs, BKI1 binds and inhibits BRI1 function; at the same time, GSK3-like kinase BIN2 phosphorylates BES1 and BZR1 family transcription factors and inhibits their functions by several mechanisms including protein degradation, reduced DNA binding, and/or cytoplasmic retention by 14–3–3 proteins. In the presence of BRs, activated BRI1 phosphorylates BKI1, which leads to dissociation of BKI1 from the plasma membrane and association of BRI1 with its coreceptor BAK1. BRI1 phosphorylates and activates BAK1, which in return phosphorylates and further activates BRI1. Activated BRI1 phosphorylates several substrates such as BSKs, TRIP-1, TTL and potentially a proton-ATPase (P-ATPase) to transduce the BR signal to downstream targets. While the targets for TRIP-1 and TTL remain to be identified, BRI1 phosphorylation of BSK1 leads to its association with BSU1 phosphatase, which can function to inhibit BIN2 kinase. At the same time, PP2A phosphatase can dephophorylate BZR1 and likely BES1 as well. These signaling events result in accumulation of unphophorylated BES1 and BZR1 in the nucleus, where they can interact with other transcription regulators such as BIM1, histone demethylases ELF6/REF6 and transcription elongation factor IWS1 to regulate genes for various BR responses. BES1 and BZR1 may interact with additional factors to repress gene expression.

Significant progress has been made in the past few years. After the biochemical confirmation that BRI1 is indeed BR receptor, proteomics and functional studies revealed a sequential phosphorylation model for the activation and signaling mechanisms of BRI1 and its coreceptor BAK1 (Wang et al. 2005a, 2008). In the process, BRI1 and BAK1 have been discovered to have both serine/threonine (S/T) and tyrosine (Y) phophorylation activities, which are important for kinase activation, substrate modification and specific BR responses (Oh et al. 2009a, 2009b, 2010, 2011; Jaillais et al. 2011). Several important signaling components that function between BRI1 and BIN2 have been identified by proteomics and genetic approaches; and the mechanisms of signal transduction have been revealed (Kim and Wang 2010; Tang et al. 2010). The regulatory mechanisms of BES1 and BZR1 by BIN2 phosphorylation appear to be much more complex than originally thought and PP2A phosphatase that dephophorylates BZR1 has been identified (Li and Jin 2007; Ryu et al. 2010a, 2010b; Tang et al. 2011). In addition, the mechanisms and transcriptional network through which BES1 and BZR1 control BR responses have been revealed by identification and characterization of BES1 and BZR1 target genes by ChIP-chip and BES1 partners by protein–protein interaction and genetic screens (Yu et al. 2008, 2011; Li et al. 2009, 2010b; Luo et al. 2010; Sun et al. 2010). Additional transcriptional factors involved in BR signaling have been identified in both Arabidopsis and rice (Li 2010b). BR functions in various developmental processes are covered in detail in two recent reviews (Clouse 2011; Yang et al. 2011). The molecular mechanisms of BR function in pollen and root have been revealed (Ye et al. 2010; Gonzalez-Garcia et al. 2011; Hacham et al. 2011). The studies confirm BR function in cell division (Hu et al. 2000) and BR action from epidermis to control root meristem cell division and expansion, similar to that found in shoots (Savaldi-Goldstein et al. 2007). Finally, several regulators for BR biosynthesis have been identified (Guo et al. 2010; Je and Han 2010; Je et al. 2010; Chung et al. 2011; Poppenberger et al. 2011). In this review, we will discuss the progress and provide some perspectives on the future directions in this fast-evolving field.

BR Receptor BRI1 and its Regulation

BRI1 is a membrane-localized receptor kinase that contains 25 LRRs (leucine-rich repeat) with an island domain between 21st and 22nd repeat in the extracellular region, a single transmembrane domain, a juxtamembrane region, a kinase domain and a C-terminal regulatory region (Li and Chory 1997). BRI1 has been established as the BR receptor by several biochemical and molecular experiments, including domain-swapping analysis with Xa21 receptor kinase, binding of immunoprecipitated BRI1 with radio-labeled brassinolide (BL, the most active BR), and binding assays with recombinant BRI1 protein (He et al. 2000; Wang et al. 2001; Kinoshita et al. 2005). In addition to BRI1, two of three BRI1 homologs were also shown the ability to bind BRs, and genetic studies showed that they played a major role in vascular development and are partially redundant with BRI1 (Caño-Delgado et al. 2004; Zhou et al. 2004).

The identification and functional characterizations of BRI1 phosphorylation sites, its cellular localizations as well as identification and characterization of BRI1 partners have greatly increased our understanding of BRI1 kinase activation and signaling outputs (Kim and Wang 2010). Homo-dimerization of BRI1 was detected in plasma membrane by fluorescence resonance energy transfer analysis, and was confirmed by co-immunoprecipitation experiment with BRI1-GFP and BRI1-FLAG transgenic plants (Russinova et al. 2004; Wang et al. 2005b; Hink et al. 2008). It was further found that this homo-dimerization of BRI1 was promoted or stabilized by BRs (Wang et al. 2005b). In the absence of BRs, two mechanisms operate to inhibit the basal BRI1's activation. The C-terminal domain of BRI1 can function to inhibit BRI1 kinase activity. In addition, BKI1 can bind BRI1 at the plasma membrane and inhibit BRI1 function. BR binding leads to the release of the auto-inhibition by the C-terminal and releasing of BKI1 into cytosol, allowing BRI1 association with its coreceptor BAK1 (Wang and Chory 2006). BAK1, a small LRR-RLK with five LRRs, was identified as BRI1’ coreceptor by activation tagging screen of bri1–5 suppressor and yeast two hybrid screen (Li et al. 2002; Nam and Li 2002). Based on careful mapping of BRI1 and BAK1 phosphorlation sites both in vitro and in vivo as well as their functional studies, a sequential transphorylation model was proposed for the activation of BR signaling (Wang et al. 2005a, 2008). In this model, BR binding to BRI1 leads to autophosphorylation on many of the phosphorylation sites to activate its kinase activity, which then phosporylates and activates BAK1 kinase activity. Activated BAK1 can in return phosphorylate BRI1 at the juxtamembrane and C-terminus to fully activate BRI1 function. The model suggests that BAK1 is involved in the full activation of BRI1 kinase activity, but not required for ligand binding. Recent studies demonstrated that in addition to functioning in BR signaling, BAK1 and its homologs also have roles in cell death and plant defense responses by acting as a coreceptor for flagellin receptor FLS2 (Chinchilla et al. 2007; He et al. 2007; Heese et al. 2007; Kemmerling et al. 2007; Albrecht et al. 2008; He et al. 2008; Kemmerling and Nurnberger 2008; Shan et al. 2008; Chinchilla et al. 2009; Gao et al. 2009; Jeong et al. 2010; Li 2010a).

The phosphorylation site mapping and functional studies also revealed that BRI1 and BAK1 are kinases with dual specificity. Interestingly, while both S/T and Y phosphorylations are important for kinase functions in both BRI1 and BAK1, Y phosphorylation appears to be important for specific BR responses and for the regulation of BKI1 activity.

Unlike several tyrosine residues that are required for BRI1 kinase activity, Y-831 phosphorylation in the juxtamembrane appears to account for some specific BR responses (Oh et al. 2009a, 2009b). Expression of BRI1Y831F rescued bri1 mutant phenotype, but displayed a larger leaf and early flowering phenotypes compared with the expression of wild-type BRI1. The results imply that Y831 control some specific BR responses, likely by altering BR signaling specificity by interacting with specific proteins.

Similarly, a recent study showed that BAK1 Y610 at the C-terminal was a major site for BAK1 tyrosine autophosphorylation (Oh et al. 2010, 2011). Interestingly, functional studies demonstrated that Y610 is required for BAK1 function in BR signaling as the BAK1Y610F transgenic plants in bak1 bkk1 mutant background displayed BR insensitive phenotype including reduced growth, accumulation of unphosphorylated BES1 and reduction of BR-regulated genes. Similarly, Flagellin signaling in defense response is also impaired. However, other BAK1 responses, such as Flagellin-mediated inhibition of plant growth and BAK1-mediated cell death are not affected. It will be interesting to determine how specific tyrosine phosphorylation events differentially affect various responses mediated by BRI1 and BAK1.

Tyrosine phosphorylation also plays an important role regulating the cellular localization of the BRI1 inhibitor BKI1 (Jaillais et al. 2011). The N-terminal lysine-arginine-rich (KR) motif targets BKI1 to the plasma membrane and a C-terminal 20-residue conserved domain mediates the interaction between BKI1 and BRI1. Deletion analysis showed that the KR repeats around motif-3 (amino acids 200–221) was sufficient for BKI1 localization to plasma membrane. Phosphorylation of a conserved tyrosine (Y211) in motif-3 of BKI1 was essential for the dissociation of BAK1 from plasma membrane. Mutation in this tyrosine (Y211F) constitutively targeted BKI1 protein to plasma membrane and overexpression of BKI1Y211F resulting in the plants with dwarf phenotype. In contrast, a phosphorylation-mimicking mutant of BKI1 (Y211D) leads to constitutive cytoplasmic localization of the mutant protein and lost ability in inhibiting BR signaling. Taken together, BR-activated BRI1 phosphorylates BKI1 at Y211, which leads to its disassociation from BRI at the plasma membrane and allows BRI1 association with BAK1.

BRI1 also phosphorylates other positive-acting substrates to transduce BR signal to downstream targets. By proteomic analysis of two-dimensional difference gel electrophoresis, several early BR-response proteins, BSKs, were identified (Tang et al. 2008b). These BSKs are the members of cytoplasmic receptor-like kinases (RLCKs) that are likely associated with plasma membrane through N-myrisylation. BSKs are positive regulators of BR response as overexpression of several members in the family suppressed bri1 phenotype and knockout of one of the members, BSK3, showed a weak BR insensitive phenotype likely due to functional redundancy among BSKs. BSK1 was phosphorylated on Serine residue 230 (S230) by BRI1 in vitro and co-immunoprecipitaion experiments indicated that BSK1 interacted with BRI1 in vivo. BRI1 and BSK1 interaction is reduced upon BR treatment, suggesting that BSKs are disassociated from BRI1 after being phosphorylated. As we will discuss in the next section, BSKs play an important role in transducing BR signal to downstream components.

BRI1 likely has additional substrates in the regulation of various BR responses. Arabidopsis TRIP-1, an essential subunit of the eIF3 eukaryotic translation initiation factor, is phosphorylated by BRI1 at several amino acids in vitro and interacts with BRI1 in vivo (Ehsan et al. 2005). As expected for an essential component in translation initiation, RNAi knockdown of TRIP-1 lead to pleiotropic phenotype with some resemblance to BR loss-of-function mutants (Jiang and Clouse 2001). It remains to be determined if TRIP-1 functions in both translation and transcription, as its mammalian homolog does, and if so, how BRI1 phosphorylation affects its activities. Arabidopsis TTL interacts with BRI1 kinase domain in yeast two-hybrid assays and is phosphorylated by BRI1 in vitro (Nam and Li 2004). Genetic studies indicated that TTL function as a negative regulator for BR-regulated plant growth and the mechanisms by which TTL modulate BR regulated plant growth remain to be determined. It was recently found that BRI-GFP is associated with a plasma-membrane Proton ATPase (P-ATPase) in a BR and BRI1 kinase dependant manner, which accompanies BR-induced cell expansion (Caesar et al. 2011). Since phosphorylation of P-ATPase is known to activate the enzyme activity, it would be interesting to determine if BRI1 can directly phosphorylate P-ATPase (Figure 1).

Regulation of Negative-acting Kinase BIN2

In addition to the large number of BRI1 loss-of-function alleles, screens for BR-insensitive mutants identified a negative regulator, BIN2 (Choe et al. 2002; Li and Nam 2002; Pérez-Pérez et al. 2002). While gain-of-function in BIN2 displayed bri1-like dwarf phenotype, loss-of-function of BIN2 and its homologs displayed a constitutive BR response phenotype (Yan et al. 2009). BIN2 is homologous to GSK3/Shaggy kinase that plays an essential role in WNT signaling pathway that is essential for animal development and is affected in many cancers (Cadigan and Nusse 1997; Polakis 2000). Recent studies suggest that BIN2 is regulated by targeted protein degradation in response to BR signaling as well as by BSK kinase through protein phosphatase BSU1.

The initial evidence that BIN2 is regulated at protein degradation comes from the observation that several gain-of-function mutations of BIN2, localized in a “TREE” domain, stabilizes BIN2 protein (Peng et al. 2008). Further studies clearly indicated that BIN2 protein accumulation is decreased by BR treatment. The decrease of BIN2 is apparently mediated by proteasome-mediated degradation pathway, as treatment with 26S proteasome inhibitor (MG132) reversed the BR-mediated decrease of BIN2 protein.

Although how BR signaling promotes BIN2 degradation remain to be established, a Kelch-repeat containing protein phosphatase, BSU1, has been shown to directly dephosphorylate and regulate BIN2 function (Kim et al. 2009). BSU1 phosphatase, identified by activation tagging with bri1–5, plays a positive role in BR signaling pathway and was found to act downstream of BIN2 (Mora-Garcia et al. 2004). However, recent genetic experiments suggest that BSU1 acts downstream of BRI1 but upstream of BIN2 (Kim et al. 2009). BIN2 can autophosphorylate at Y200, which is required for BIN2 function. BSU1 binds and dephosphorylates BIN2 at Y200, thereby inhibiting BIN2 function. Co-immunoprecipitation and BiFC assay suggested that BSK1 and BSU1 interact with each other in vivo and the phosphorylation of BSK1 at serine 230 (S230) by BRI1 promotes this interaction. The study therefore connected all the components between BRI1, its substrate BSK1, BSU1 and BIN2 in a linear pathway. It remains to be determined if BIN2 is also regulated by additional mechanisms and how different mechanisms coordinate to control this important regulator in the BR pathway.

Regulation of BES1 and BZR1

BES1 and BZR1 are two major transcription factors that are regulated by BIN2 and mediate BR-regulated gene expression (Wang et al. 2002; Yin et al. 2002a, 2005; Zhao et al. 2002; He et al. 2005). BES1 and BZR1 are 88% identical and are composed of DNA binding domain (DBD), BIN2 phsophorylation domain with more than 20 putative BIN2 phosphorylation sites (S/TxxxS/T), and a C-terminal domain (CTD) (Figure 2). BES1 and BZR1 DBDs are predicated to form a bHLH structure although they are not classified as typical bHLH transcription factors. The CTD is required for BES1 function as deletion of this domain leads to accumulation of inactive BES1 that acts as a dominant-negative form (Yin et al. 2005). The C-terminal domain most likely acts as a transcription activation domain as it activates reporter gene expression in yeast when fused with GAL4 activation domain. In addition, the C-terminal domain also contains a 12 amino acid docking motif (DM) that binds BIN2, allowing BIN2 to phosphorylate BZR1 (Peng et al. 2010). Since the same domain is conserved in BES1, it is likely that BIN2 interacts with DM to phosphorylate BES1 as well. BIN2 phosphosphorylates BES1 and BZR1 at their central phosphorylation domain and inhibits their function likely through several different but non-exclusive mechanisms, including targeted protein degradation, nuclear export and cytoplasmic retention by 14–3–3 s and decreased DNA binding of the BIN2-phosphorylated protein (He et al. 2002; Yin et al. 2002b; Vert and Chory 2006; Bai et al. 2007; Gampala et al. 2007; Ryu et al. 2007; Ryu et al. 2008, 2010a).

Figure 2.

BES1 AND BZR1 protein structure and their interacting proteins. BES1 and BZR1, highly related to each other, have bHLH DNA binding domain (DBD), a BIN2 phosphorylation domain that contains 22 BIN2 phosphorylation sites (indicated by *), a PEST motif that is implicated in protein degradation, and a 14–3–3 binding motif that interacts with 14–3–3 when BES1 and BZR1 are phophorylated. The C-terminal domain (CTD) includes a docking motif (DM) for BIN2 binding and a transcription activation function. Histone demethylases ELF6/REF6 and bHLH protein BIM1 interact with BES1 through their basic region and HLH dimerization domain of DBD, respectively. Phosphorylation of T175 for BES1 and T177 for BZR1 are important for interaction with 14–3–3s. While PEST domain is also required for interaction with PP2A, IWS1 and MYB30 interacts with BES1 through a region that contains part of the phosphorylation sites and PEST motif. The amino acid numbers are based on BES1.

BIN2 phosphorylates BES1 and BZR1 at many putative phosphorylation sites (Figure 2). Phosphorylation of different sites could affect different aspects of BES1 and BZR1 functions. For example, phosphorylation at T177 of BZR1 and T175 of BES1 is required to interact with 14–3–3 for cytoplasmic retention (de Vries 2007; Ryu et al. 2010a). Recent studies also suggest that BSU1 and BIN2 preferably act at different locations to regulate BES1 functions. While the cytoplasmic location is more important for BSU1, BIN2 likely phosphorylate BES1 and BZR1 in the nucleus to trigger its nuclear export (Ryu et al. 2010b).

A phosphatase that dephosphorylates BZR1 has been identified by looking for BZR1-interacting proteins through tandem affinity purification (Tang et al. 2011). Protein phosphates 2A (PP2A) is a heterotrimeric serine/threonine phosphatase, which contains as scaffolding subunit A, catalytic subunit C, and a regulatory B subunit that interacts with substrates (Janssens et al. 2008). In vitro and in vivo experiments showed that BZR1 was indeed able to interact with several PP2A B′ isoforms, such as B′α, β and η through the PEST domain of BZR1. While the loss-of-function PP2A mutant accumulated more phosphorylated BZR1, overexpression of B′ components increased both phosphortylated and unphosphorylated BZR1, suggesting that PP2A affects both BZR1 dephosphorylation and protein degradation. Pharmacological studies with GSK3 kinase inhibitor bikinin and PP2A inhibitor okadaic acid also support a role of PP2A in BZR1 dephosphorylation. Immunoprecipitated PP2A can dephosphorylate BZR1 in vitro, which likely affects the inhibitory effects of phosphorylation on BZR1 including the binding of 14–3–3 protein. Finally, deletion of PEST motif in BZR1 leads to the accumulation of phosphorylated protein and a dwarf phenotype, consistent with a negative role of PEST in protein degradation and positive function in recruiting PP2A. Yeast two-hybrid experiments showed that PP2A B′α and β also interacted with BES1, implying that PP2A regulated the function of BES1 as well.

Network and Mechanism for BES1 and BZR1 Regulated Gene Expression

Brassinosteroids affect many growth and developmental processes and much of them are likely due to BR-mediated changes in gene expression. Several genome-wide microarray experiments in Arabidopsis have demonstrated that BRs regulate hundreds to thousands of genes (Goda et al. 2004; Nemhauser et al. 2004, 2006; Guo et al. 2009). Proteomics study also identified many proteins in response to BRs in both Arabidopsis and rice (Tang et al. 2008a; Wang et al. 2010). Understanding how BES1 and BZR1 coordinate with other proteins to control the expression of the large number of genes in a transcriptional network is important to understand how BRs regulate various biological processes at different stages of growth and development under various environmental conditions. Identification and characterization of BES1 and BZR1 partners and target genes can help address this question.

BES1 and BZR1 direct target genes have been recently identified by Chromatin immunoprecipitation followed by genomic tiling array (ChIP-chip) (Sun et al. 2010; Yu et al. 2011). In total, 1609 BES1 targets were identified with 2-week-old bes1-D seedlings and an anti-BES1 antibody, at least 250 of them are regulated by BRs (Yu et al. 2011). On the other hand, 3410 BZR1 target genes were identified with 4-week-old transgenic BZR1-CFP plants treated with BL with an anti-CFP antibody, 953 of them are regulated by BRs (Sun et al. 2010). The different plant ages used (seedlings vs. adult plants) probably account for some of the differences in numbers of target genes identified for BES1 and BZR1. Nevertheless, about half of BES1 target genes are also BZR1 targets, which is consistent with the fact that these two factors function redundantly with distinctive functions (Wang et al. 2002; Yin et al. 2002b). Several important conclusions can be drawn from the characterization of BES1 and BZR1 target genes.

The analysis of enriched promoter elements in BR-regulated BES1 and BZR1 target genes confirmed previously identified BES1 and BZR1 DNA binding sites and changed previous perception about the activation and repression of these two transcription factors. It was previously shown that BES1 binds to E-box sequences to activate a BR-induced gene expression and BZR1 binds to BRRE on CPD (CONSTITUTIVE PHOTOMORPHOGENESIS AND DWARFISM) promoter to repress its expression (He et al. 2005; Yin et al. 2005). However, promoter elements analysis indicated that both BES1 and BZR1 can bind both BRRE and E-boxes (particularly CATGTG and CACGTG that is also termed G-box) in vivo with BRRE mostly enriched in BR-repressed genes and E-boxes are mostly enriched in BR-induced genes (Sun et al. 2010; Yu et al. 2011). Since E-box and BRRE are also enriched in both BR-induced and BR-repressed genes, additional promoter sequence elements and/or BES1 and BZR1 interacting proteins likely determine if these transcription factors either activate or repress gene expression.

BES1 and BZR1 can bind and induce its own expression, probably in a positive feedback loop (Yu et al. 2011). At the same time, BES1 and BZR1 inhibit many genes involved in BR biosynthesis and signaling, likely as a feedback inhibition mechanism (Sun et al. 2010). These observations also point out that at least several regulators in the BR pathway, such as BES1, BZR1, BRI1 and BIN2, can be regulated by BRs at both transcription and post-transcriptional levels.

Both BES1 and BZR1 target gens suggest interactions between BR and other signaling pathways (Clouse 2011). In particular, the genes involved in light and other hormones responses including auxin, gibberellin (GA), abscisic acid (ABA), ethylene, jasmonic acid (JA) and cytokinin are enriched in both BES1 and BZR1 targets. The crosstalk between auxin ABA and BRs have been determined at molecular levels (Hardtke 2007; Zhang et al. 2009b).

Brassinosteroid actions are closely regulated by light, which is directly demonstrated by the constitutive photomorphogenesis/de-etiolation phenotype of BR mutants (Chory et al. 1991; Li et al. 1996; Szekeres et al. 1996). Microarray studies with det2 mutant indicated that BRs negatively regulate genes involved in photomorphogenesis (Song et al. 2009). The interaction between BZR1 and light signaling was further confirmed by the finding that BZR1 represses light signaling components (Sun et al. 2010). The comparison between targets between BZR1 and regulated transcription factors indicated that about one third of target genes are in common between BZR1 and HY5, a transcription factor mediating light regulated gene expression (Oyama et al. 1997; Lee et al. 2007). The crosstalk is further confirmed with one of the BZR1 targets (Luo et al. 2010). Genetic studies indicated that GATA2 is a negative regulator in the BR pathway. The protein level of GATA2 was regulated by light through COP1-dependent proteasome degradation. GATA2 is therefore inhibited by BRs at transcription level and promoted by light at the protein level, providing a link between BR and light signaling pathways. Finally, BES1 and BZR1 repress the expression of two related transcription factors, GLK1 and GLK2, which function redundantly to promote chloroplast development (Waters et al. 2008, 2009; Sun et al. 2010; Yu et al. 2011). It is well known that BR loss-of-function mutants have premature chloroplast development but the mechanisms are not known (Chory et al. 1991). It is conceivable that BRs function through BES1 and BZR1 to repress GLK1 and GLK2 expression and thus chloroplast development in the dark. Consistent with the hypothesis, bes1-D mutant, in which BES1 accumulated to high levels, has reduced expression of GLK1 and GLK2, reduced expression of GLK target genes and altered chloroplast function (Waters et al. 2008, 2009; Sun et al. 2010; Yu et al. 2011). The connection between BR signaling and chloroplast development is further supported by characterization of BPG2 (BRZ-INSENSITIVE-PALE GREEN 2), a chloroplast protein involved in BR response (Komatsu et al. 2009). Taken together, the identification and characterization of BES1 and BZR1 target genes provided more evidence that light and BR pathways have extensive crosstalk in the regulation of plant growth, photomorphogenesis and chloroplast development.

Some of the BES1 and BZR1 target genes were further confirmed by functional studies to mediate BR response or BR-regulated plant growth. In the case of BES1, knockout mutants for 15 BES1-targeted Transcription Factors (BTFs) were tested and 12 of them showed BR response defects in hypocotyl elongation assays (Yu et al. 2011). Interestingly all of these 12 BTFs are also BZR1 targets. With a few exceptions, BES1 and BZR1 in general promote the expression of BTFs that are positive regulators of BR response and repress BTFs that are negative regulators of BR response. In the case of BZR1 targets, overexpression of DREPP, one of the BR-regulated BZR1 targets, increased the cell length of det2 mutant (Sun et al. 2010). On the other hand, the transgenic plants overexpressing BZS1, another BZR1-repressed transcription factor gene, were hypersensitive to BRZ, while co-suppressed lines of BZS1 showed longer hypocotyls grown on BRZ medium compared with wild type plants. Several BES1 and BZR1-targeted receptor-kinases mediate vegetative plant growth in adult plants (Guo et al. 2009; Sun et al. 2010).

Apparently, BES1 and BZR1 target genes depend on developmental stages and tissue specificity. BRs are known to regulate male fertility, but the mechanisms were not well defined. A recent study indicated that BR mutants are defective in several aspects of anther and pollen development, including reduced filament length, fewer numbers of pollen grains, and defects in tapetal development, pollen wall formation and pollen release (Ye et al. 2010). Consistent with the mutant phenotype, several key genes involved in the process, including SPL/NZZ required for microspore mother cell development, TDF1, AMS and AyMYB103 involved in microspore development, MS1/MS2 required for tapetal development and pollen wall formation are reduced in BR mutants. ChIP experiment with chromatin isolated from flower tissues indicated that most of these genes are direct BES1 targets. Since most of these genes are not detected as BES1 targets with 2-week-old seedlings, these results clearly indicated that BES1 targets different genes in different tissues and developmental stages.

Besides BES1 and BZR1 targets, BES1 interacting proteins provide an additional dimension to modulate BR-regulated gene expression in response to developmental and environmental cues. BES1 recruits two related histone demethylases, ELF6 and REF6, to modulate BR-regulated gene expression and BR responses (Yu et al. 2008). ELF6 and REF6 were originally identified as two genes that regulate flowering time (Noh et al. 2004). It was found that ELF6 and REL6 interact with BES1 both in vitro and in vivo through the basic region in the bHLH DNA binding domain (Figure 2). ELF6 and REF6 belong to the JHDM3 subfamily of Jumonji family histone demethylases that function by removing methyl groups from various histone residues (Klose et al. 2006). ChIP assay indeed suggests that histone methylation is elevated at the promoter of BR-induced gene TCH4 in elf6 and ref6 mutants. The study demonstrated that BES1 recruited ELF6 and REF6 to change chromatin structure and regulate genes expression. BRs are known to modulate flowering time as loss-of-function BR mutants have delayed flowering, which is accompanied by increased expression of FLC (Domagalska et al. 2007; Li et al. 2010a). ELF6 and REF6 therefore provide a molecular link between plant growth and reproduction (Clouse 2008).

In addition to histone modifying enzymes, BES1 also recruits IWS1 (INTERACTING-WITH-SPT6–1), a conserved protein implicated in transcription elongation, to regulate BR target genes. IWS1 was identified in a genetic screen for genes required for BES1 function by looking for suppressors for constitutive BR response mutant bes1-D (Li et al. 2010b). The iws1 mutants suppress bes1-D phenotypes, displayed a semidwarf phenotype and reduced BR response in hypocotyl elongation assays. Gene expression studies indicated that about 1/3 of BR-induced genes are compromised in the iws1 mutants. IWS1 interacts with BES1 in vitro and in vivo through the central domain of BES1 (aa 140–271, Figure 2). AtIWS1 is a homolog of IWS1 in the yeast/human and interacts with histone chaperone and transcription elongation factor Spt6. Yeast IWS1, also termed SPN1, is implicated in inducible gene expression in yeast and its function involves histone remodeling complex SWI/SNF (Fischbeck et al. 2002; Zhang et al. 2008). In human cells, IWS1 is required for splicing of HIV gene and global RNA export (Yoh et al. 2007, 2008). Recent structure studies revealed that IWS1 has structure feature similar to transcription elongation factor TFIIS (Pujari et al. 2010). Genomic studies suggest that the expression of up to 30% of genes can be regulated at steps after transcription initiation (Kim et al. 2005; Guenther et al. 2007); but the mechanisms for such regulation are not well defined. The study established IWS1 as a target for BR signaling, providing a potential new mechanism for the regulation of gene expression.

Other Transcription Factors Involved in BR Signaling

Other family transcription factors have been found to be involved in BR signaling in Arabidopsis and rice (Li 2010b; Clouse 2011). Several small and atypical HLH (helix-loop-helix) proteins, ATBS1 (ACTIVATION TAGGED BRI1 SUPPRESSOR 1), its Arabidopsis homologs including KIDARI and PRE1 (PACLOBUTRAZOL RESISTANT 1), and rice orthologs, ILI1 (INCREASED LAMINA INCLINATION 1) and BU1 (BRASSINOSTEROID UPREGULATED 1), were identified as positive regulators for BR response as overexpression of these genes display increased BR responses (Tanaka et al. 2009; Wang et al. 2009a; Zhang et al. 2009a). ATBS1/PRE/ILI cannot bind DNA and therefore likely function by blocking the DNA binding activity of AIF (ATBS1-INTERACTING FACTOR)/IBH1 (ILI1–BINDING bHLH) bHLH proteins that function as negative regulators of the BR pathway. Overexpression of AIF1/IBH1 resulted in the plants with BR-like dwarf phenotype. Interestingly, AIF1/IBH1 and PRE1 are BZR1 targets that are repressed and induced by BZR1, respectively. These results suggest AtBS1/PRE/ILI1/BU family proteins are positive factors for BR pathway by sequestering AIFs/IBH1, the negative regulators of the BR pathway. How AIFs/IBH1 inhibits BR response remains to be established. AtBS1/PRE/ILI1/BU family proteins appear to be involved in other signaling processes as well (Clouse 2011).

In rice, BES1 and BZR1 homolog, OsBZR1, functions as a positive regulator of BR response (Bai et al. 2007). Interestingly, rice DLT (DWARF AND LOW TILLERING), a member of unique GRAS family transcription factors, also act as positive regulators of BR response as loss-of-function mutants display BR-like dwarf phenotype and have increased expression of BR biosynthesis genes (Tong et al. 2009). Several MADS box proteins, OsMDP1, OsMADS22 and OsMADS55 are negatively regulated by BRs and function as negative regulators in the BR pathway (Duan et al. 2006; Lee et al. 2008).

Regulation of BR Biosynthesis

Brassinosteroid biosynthesis pathway is well established (Fujioka and Yokota 2003; Asami et al. 2005). It is well known that BR signaling inhibits BR biosynthesis through BES1 and BZR1 inhibition of the expression of DWF4, CPD and other biosynthesis genes (Noguchi et al. 1999; Choe et al. 2002; Mora-Garcia et al. 2004; Sun et al. 2010). Recent studies expanded our understanding how BR biosynthesis is positively regulated (Figure 3). BRX (BREVIS RADIX) was identified as a gene required for optimal root growth by promoting the expression of CPD gene expression; and was later found to promote shoot growth as well (Mouchel et al. 2004, 2006; Beuchat et al. 2010a, b). Since BRX expression is induced by auxin and feedback inhibited by BR signaling, BRX apparently regulates BR level by coordinating auxin signaling and BR feedback pathway in the regulation of root growth (Mouchel et al. 2006). Interestingly, BRX protein also translocates from membrane to nucleus in the presence of auxin, suggesting that auxin regulate BRX activity through multiple mechanisms (Scacchi et al. 2009). Auxin signaling was recently found to induce the expression of DWF4, possibly through an Auxin Responsive Element (AuxRE) and its interacting protein, which may reduce BZR1/BZR1 binding to DWF4 promoter (Chung et al. 2011).

Figure 3.

Regulation of brassinosteroid biosynthesis. Brassinosteroid (BR) biosynthesis pathway is shown from cycloartenol to brassinolide (BL). The enzymes catalyze most of the steps are shown (see text for references). Different color indicates enzymes/proteins from different species (black: Arabidopsis; red: rice and green: pea). BES1 and BZR1, OsBZR1, and DLT function to inhibit many of the BR biosynthetic genes. In Arabidopsis, DWF4 is induced by TCP1; CES and auxin (likely through an unidentified ARF protein) and CPD expression is promoted by BRX and CES. In rice, RAVL1 promotes the expression of D2, D11 and BRD1 as well as BR receptor gene OsBRI1. BL and its precursor, castasterone, can be inactivated by cytochrome P450 enzymes BAS1 and its homolog SOB7/CHI2 /SHK1.

T-DNA activation tagging identified two bHLH transcription factors, TCP1 and CTSTA (CES), both of which positively regulate BR biosynthetic gene expression. The T-activation tagged allele of TCP1, tcp1-D, suppresses the dwarf phenotype of weak allele of BRI1, bri1–5, but not bri1–4, a null BRI1 allele, suggesting that TCP1 functions upstream of BRI (Guo et al. 2010). Loss-of-function of TCP1 created by TCP1 fusing with EAR transcription repressor domain (SRDX) resulted in a dwarf phenotype in adult plant and short hypocotyls that can be recovered by exogenous BL. BR measurement in the TCP1 mutants suggest that DWF4 was the target of TCP1. Gene expression and ChIP assay confirmed that DWF4, but not other biosynthetic genes, was directly regulated by TCP1. TCP1 is a member of a unique family of bHLH proteins that have an additional conserved region (R-domain) (Cubas et al. 1999). The characterization of TCP1 binding sites will be crucial to understand exactly how it regulates DWF4 expression.

Similarly, overexpression of CESTA (CES) leads to increased plant growth that resembles BR gain-of-function mutants (Poppenberger et al. 2011). In contrast, loss-of-function mutant created by CES-SRDX caused a BR-related dwarf phenotype. BR levels are increased, and BR biosynthesis genes, DWF4 and CPD, are increased in ces-D. In addition, CES protein was shown to bind CPD promoter in vitro and in vivo likely through G-box (CACGTG). Interestingly, CES appears to interact with BEE1, BEE2 and BEE3, which were previously shown to be induced BR and play a positive role in BR-regulated growth (Friedrichsen et al. 2002). Although the CES gene is not regulated by BRs, CES-YFP protein accumulates as distinct nuclear bodies in response to BR treatment. These results raise a possibility that CES protein is regulated by BR signaling. It would be interesting to determine how BES1 and BZR1, TCP1, and CES/BEEs function together in the regulation of BR biosynthesis.

In rice, B3-domain containing transcription factor, RAVL1 (RELATED TO ABI3/VP1, ABA INSENSITIVE 3/VIVAPARIOUS 1), was identified as a positive regulator of BR response in rice (Je and Han 2010; Je et al. 2010). RAVL1 regulates the expression of BR receptor OsBRI1 and several BR biosynthesis genes, including D2, D11 and BRD1 through E-box sequences (Yamamuro et al. 2000; Hong et al. 2002, 2003; Tanabe et al. 2005).

In addition to the regulation of BR biosynthesis genes at transcription level, regulatory proteins can directly regulate a couple of BR-biosynthesis enzymes. A small G-protein from Pea, Pra2, is expressed in the dark and interacts with BR biosynthesis enzyme DDAWF1 to promote BR level in the dark (Kang et al. 2001). GSR1 (GA STIMULATED TRANSCRIPT IN RICE 1, a small cystein-rich protein) binds to DIM1/DWF1 to regulate its enzyme activity and BR level in rice, providing another mechanism for the regulation of BR biosynthesis (Wang et al. 2009b). OsGSR1 is regulated by GA and thus provides a link between BR and GA pathways.

In addition to the regulation of biosynthesis, BRs such as castasterone and BL can be inactivated by conserved P450 moonooxygenases, CYP734A1/BAS1 and CYP72C1/SOB7/CHI2/SHK1, through C26 or C22 oxidations (Neff et al. 1999; Nakamura et al. 2005; Takahashi et al. 2005; Turk et al. 2005; Ohnishi et al. 2006; Thornton et al. 2010; Sakamoto et al. 2011).

Conclusion

Many new regulatory components have been identified in the BR signaling or biosynthesis pathways. Identification of these components is just the beginning for the understanding how BRs function to regulate a large number of biological processes at different stages, tissues and environmental conditions. Identifications of interacting proteins for many of the BR signaling components have provided important insight into the functions and regulation in BR signaling and will continue to do so in the future. As we learn from some of the components, regulation for each protein can be very complex involving multiple mechanisms. Identifications of both in vitro and in vivo phosphorylation sites and mapping the functional domains in combination with genetic studies will reveal new mechanisms in the regulation of BR pathways. Although a large number of BES1 and BZR1 target genes have been identified, it is still a daunting task to figure out how all of these genes are regulated and more importantly, how they function together to control specific processes. BES1 and BZR1 activate and repress approximately equal numbers of genes; how they interact with different partners to perform the opposite functions in transcription remains to be defined. An initial focus on BES1 and BZR1 targeted transcription factors and their transcriptional partners seem to be reasonable to move forward. Eventually, computational modeling, in combination with genetic and genomic studies, are needed to put hundreds to thousands of BR target genes in order to fully understand how a simple hormone can lead to fundamental changes in plant growth and development, not only by itself, but also with many other hormonal and environmental pathways.

(Co-Editor: Li-Jia Qu)

Acknowledgement

The research in author's lab is supported by United States National Science Foundation (IOS 0546503).

Note added in proof: A recent study demonstrated that PP2A, regulated by BR-induced SBI1 leucine carboxylmethyltransferase, dephosphorylates BRI1 and promotes its degradation (Wu et al. 2011). PP2A therefore has dual roles in BR pathway: termination of BR signaling by regulating BRI1 and promotion of BR signaling by dephosphorylating BZR1 and BES1 (Di Rubbo et al. 2011).

Di Rubbo S, Irani NG, Russinova E (2011) PP2A Phosphatases: The “On-Off” Regulatory Switches of Brassinosteroid Signaling. Sci. Signal. 4, pe25.

Wu G, Wang X, Li X, Kamiya Y, Otegui MS, Chory J (2011) Methylation of a phosphatase specifies dephosphorylation and degradation of activated brassinosteroid receptors. Sci. Signal. 4, ra29.

Ancillary