Reply: Interaction between Brassinosteroids and Gibberellins: Synthesis or Signaling? In Arabidopsis, Both!

Brassinosteroids (BRs) are plant hormones with versatile roles. Among other functions, the BRs control cell elongation, division, and differentiation events (Gudesblat and Russinova, 2011), fulfill roles in adaptive growth processes (Wang et al., 2012; Fridman and Savaldi-Goldstein, 2013), and participate in abiotic and biotic stress responses (Kagale et al., 2007; De Bruyne et al., 2014; Eremina et al., 2016). Their many roles throughout the plant life cycle are enabled by a central signaling module, which controls the expression of thousands of target genes (Sun et al., 2010; Yu et al., 2011). This module has been heavily studied in the last two decades, thereby considerably improving our understanding, which has been recently summarized in a number of excellent reviews (Guo et al., 2013; Wang et al., 2014; Singh and Savaldi-Goldstein, 2015). Here, we summarize our integrated model of BR-GA crosstalk (Unterholzner et al., 2015) and explain why we consider BRs as “master regulators” of GA biosynthesis.

BR AND GA BIOSYNTHESIS AND SIGNALING ARE INTERTWINED IN ARABIDOPSIS

When the first BR deficient mutants of Arabidopsis thaliana were isolated (Clouse et al., 1996; Li et al., 1996; Szekeres et al., 1996), it became evident that BRs may act synergistically with GAs in aspects of growth control (Steber and McCourt, 2001). In Arabidopsis and other species, BR mutants phenotypically resemble GA-deficient plants in impaired germination (Unterholzner et al., 2015), hypocotyl elongation, dwarfism, darker green leaves, late flowering, and reduced fertility (Clouse, 2011). Numerous studies have shown that BRs and GAs interact at the signaling level (Bai et al., 2012; Gallego-Bartolomé et al., 2012; Li et al., 2012; Bernardo-García et al., 2014). GAs signal through the degradation of the DELLA transcriptional repressors (Schwechheimer, 2012; Davière and Achard, 2013), which is initiated following the perception of bioactive GA by the GA receptor GID1 (Griffiths et al., 2006; Willige et al., 2007). There is good evidence that DELLAs interact with the BR-related transcription factors BES1 and BZR1 and repress their activities on target promoters (Bai et al., 2012; Gallego-Bartolomé et al., 2012; Li et al., 2012).

Because hypocotyl elongation of BR mutant seedlings could not be restored with externally applied GA in Arabidopsis (Li et al., 1996; Szekeres et al., 1996; Bai et al., 2012; Gallego-Bartolomé et al., 2012), it was postulated that BRs do not regulate GA biosynthesis but that the crosstalk is restricted to the signaling level. Therefore, it was surprising that when determining the GA contents in Arabidopsis BR mutants, levels of the major bioactive GA, GA₄, were clearly reduced (Unterholzner et al., 2015). This finding prompted us to readdress a possible role of BRs in GA biosynthesis in this species.

Since an impact of BRs on GA biosynthesis had been excluded in Arabidopsis based solely on experiments performed on seedlings, we expanded the experiments to additional growth stages. This confirmed the earlier results: In the seedling stage, the hypocotyls of the severe BR mutants bri1-1 and cpd responded only weakly to GA when grown on media containing Murashige and Skoog (MS) salts. However, on media without MS salts, the mutants were fully responsive to GA. Moreover, GA almost completely restored the growth defects of these mutants during germination and control of flowering time (Unterholzner et al., 2015). Thus, contrary to what is implied by Ross and Quittenden (2016), externally applied GA can be effective on severe BR mutants, but this depends on the developmental stage and the physiological conditions, suggesting that additional factors influence the process.

When we addressed the molecular mode of BR activity in GA biosynthesis, we found that the expression of multiple genes encoding enzymes of the GA20ox and GA3ox families, in particular GA20ox1 and GA3ox1, were significantly reduced in BR mutants (Unterholzner et al., 2015), confirming earlier reports (Bouquin et al., 2001; Gallego-Bartolomé et al., 2012). Because GA20ox1 expression was enhanced in the dominant bes1-D and bzr1-1D mutants, we studied the role of BES1 in the process. Using in vitro and in vivo DNA binding studies, we showed that BES1 can directly bind to a previously unknown binding motif present in the promoters of several GA biosynthesis genes in a BR-induced manner (Unterholzner et al., 2015). GA measurements in ASKθ-oe, a line in which BR signaling is constitutively repressed (Rozhon et al., 2010), provided evidence that BRs are required for the synthesis of several GAs, namely, GA₁₅, GA₂₄, GA₉, and GA₄. Whereas the bioactive GA GA₄ was reduced 5-fold, GA₉, its biosynthetic precursor and the product of GA20ox activity, was reduced 16-fold in ASKθ-oe compared with that in the wild type (Unterholzner et al., 2015). Therefore, there is evidence that BRs are master regulators of GA biosynthesis in Arabidopsis because they directly regulate not only one but multiple steps of this central metabolic pathway.

With these results, we proposed an expanded model for BR-GA activity in the growth control of Arabidopsis (Unterholzner et al., 2015, Figure 8), where BRs induce GA production to release DELLA-repressive effects on BES1/BZR1 in a feed-forward mode, promoting joined BR-GA responses. In this model, BR-GA cooperation at the signaling level is of equal importance as the function of BRs in GA synthesis because without the former the later is ineffective. In our view, there is no need to envision two separate models; rather, we expanded upon the existing model in which signaling clearly plays a key role.

In Arabidopsis, the function of BRs in GA biosynthetic gene expression is complemented by a role of GAs in BR biosynthetic gene expression (Stewart Lilley et al., 2013). Therefore, BRs and GAs act in a highly interlinked manner in Arabidopsis, impacting each other’s synthesis and cooperating in signaling. That said, it is clear that both hormones have additional roles that do not depend on each other (see Tong and Chu, 2016), and in Arabidopsis, this is supported by the observation that when GA20ox expression is reconstituted in the weak bri1-301 allele, some growth defects are restored whereas others remain (Unterholzner et al., 2015).

DEVELOPMENTAL STAGE AND ENVIRONMENTAL CONTEXT INFLUENCE BR AND GA ACTIVITIES

GA biosynthesis is highly regulated by exogenous stimuli (Stavang et al., 2005, 2010; Weller et al., 2009), which must be taken into account when GA-related phenotypes are studied. For example, in BR mutants, a 4°C treatment fully recovered germination defects (Unterholzner et al., 2015), indicating that cold may act independently of BRs to induce GA biosynthesis during germination.

In addition, GA homeostasis is closely monitored and adjusted via GA signaling. In particular, the DELLAs contribute to the feedback control of GA production because transcript levels of GA20ox and GA3ox genes are low in mutants lacking DELLAs but are strongly elevated in plants containing stabilized DELLA forms (Dill et al., 2001; Zentella et al., 2007; Weston et al., 2008). Interestingly, this feedback control appears to be inactive in BR mutants because despite low levels of bioactive GA, GA20ox and GA3ox expression is not induced (Unterholzner et al., 2015). Therefore, it is possible that BRs participate in the feedback control of GA biosynthesis, and this could be achieved by a regulatory role of DELLAs in BES1/BZR1 activities also in GA biosynthesis. Because DELLAs promote the expression of the BES1 targets GA20ox1 and GA3ox1 (Dill et al., 2001; Zentella et al., 2007; Weston et al., 2008; Gallego-Bartolomé et al., 2012) but repress BES1 activity on the promoters of other GA signaling-responsive genes (Bai et al., 2012; Gallego-Bartolomé et al., 2012; Li et al., 2012; Zentella et al., 2016), it will be interesting to determine how such potential dual activities of BES1/BZR1+DELLA-containing protein complexes may be realized. It is probable that, depending on spatiotemporal requirements, both developmental context and environmental setting govern the assembly of these complexes.

SPECIES- AND TISSUE-SPECIFIC DIFFERENCES

Matters are further complicated by species-specific differences that may occur in the functions of BRs during GA biosynthesis. In tomato (Solanum lycopersicum), GA application rescues aspects of BR mutant phenotypes (Bishop et al., 1999; Nomura et al., 2005; Martí et al., 2006). Micro-Tom, a mutant defective in DWF4 expression, responds to the application of GA₃ with increased growth (Martí et al., 2006), and phenotypes of the severe BR mutant d^x, which is deficient in CYP85A1 expression (Bishop et al., 1999; Nomura et al., 2005), could be partially restored with GA₃. In particular, mean stem height, number of internodes per plant, and mean internode length were significantly increased. In contrast, abnormal leaf morphology was not released in d^x (Nadhzimov et al., 1988). In this respect, it is interesting that, although other phenotypes were rescued, the leaf morphology in the Arabidopsis bri1-301 mutant could not be restored by reestablishing GA20ox expression in the BRI1 expression domains (Unterholzner et al., 2015). When GAs were measured, d^x showed significantly increased GA₂₀ levels; GA₁ was not quantified (Nadhzimov et al., 1988).

Recently, Li et al. (2016) showed that d^{^im}, a weak cyp85a1 mutant of tomato, had increased GA₂₀ levels, whereas GA₁ levels were not altered. Increased GA₂₀ levels were also observed in the pea (Pisum sativum) BR biosynthetic mutant lkb, whereas the level of bioactive GA₁ was slightly reduced to slightly increased, depending on the tissue investigated (Lawrence et al., 1992; Jager et al., 2005). Interestingly, the application of brassinolide (BL) significantly reduced the elevated level of GA₂₀ in lkb and lk; thus, the authors concluded that “the reduction of GA₂₀ levels in BR-deficient mutants after the application of BL indicates a clear effect of BRs on the GA biosynthesis pathway” (Jager et al., 2005). However, how this effect is mediated remains to be investigated.

For rice, there is evidence that BR deficiency impairs GA production because GA₁ was decreased in the BR-deficient mutants d11, GSK2oe, and dlt, whereas it was increased in the BR overaccumulating line m107. When GA intermediates were determined in the BR overaccumulating line, products of GA20ox activities were decreased, whereas GA₁, the product of GA3ox activities, was increased. This correlated with corresponding changes in the GA biosynthetic gene expression in the mutants (Tong et al., 2014). Therefore, there is evidence that in rice, similar to Arabidopsis, BRs control GA biosynthesis through the regulation of GA biosynthetic gene expression, albeit by different means.

These past studies used different tissues for measurements, and the growth conditions strongly varied, ranging from soil growth to hydroponic culture systems, from 10 to 18 h daylength, and from 21 to 30°C daytime temperature. Light levels and spectra remained mostly undefined. Furthermore, the genetic backgrounds used are not easily comparable because they differ in the degree of BR or BR signaling deficiency. In future studies that measure GAs or determine other GA responses in BR mutants, it will be important that the same developmental stages, tissues, and mutants of comparable severity are used. Moreover, it will be critical that the same growth conditions are applied to be able to make statements regarding similarities or differences between plant species.

It is evident that our knowledge of the interplay of BRs and GAs is just beginning to form. At this stage, we know too little regarding the molecular modes of BR and GA activities during gene regulation, be it on promoters of GA biosynthesis genes or targets further downstream during signaling, to postulate dogmatic models. Moreover, linear insulated pathways, as proposed by Ross and Quittenden (2016), do not reflect the complex events of hormonal signaling cascades, which are highly interconnected with other signaling modules. Developmental phase and species-specific regulatory modes yet to be revealed will further expand our knowledge of BR and GA interplay during growth control and other processes, such as stress responses, which are yet to be studied.