- Top of page
- Materials and Methods
- Supporting Information
Interactions between hormones have been demonstrated for several plant developmental stages, from hypocotyl elongation and the formation of the apical hook to the determination of leaf and shoot form, tropic and nastic movements of the shoot, and the control of flowering (Vandenbussche & Van Der Straeten, 2004; Achard et al., 2006, 2007). Crosstalk between gibberellins (GAs) and ethylene, as well as with other hormones such as auxins, has been demonstrated (Achard et al., 2003; Fu & Harberd, 2003; Vriezen et al., 2004; De Grauwe et al., 2007). DELLA proteins, which act as nuclear repressors of GA signalling, appear to be key integrators in the GA–ethylene crosstalk. They are members of the GRAS family of transcriptional regulators – gibberellin insensitive (GAI), repressor of ga1-3 (RGA), and SCARECROW – which are characterized by the conserved N-terminal (DELLA) domain (Richards et al., 2001). In Arabidopsis, the DELLA family comprises GAI, RGA and RGA-like 1/2/3 (RGL1/2/3) (Peng et al., 1997; Silverstone et al., 1998, 2001; Lee et al., 2002; Wen & Chang, 2002). These proteins are rapidly destabilized after GA treatment through degradation by the 26S proteasome (Silverstone et al., 1998; Dill & Sun, 2001; Fu et al., 2002). The destabilization is an effect of polyubiquitination and subsequent degradation by means of the ubiquitin proteasome system involving the Skp1/cullin/F-box (SCF) E3 ubiquitin ligase enzyme complex. In Arabidopsis, DELLA proteins are targeted for degradation by the F-box protein Sleepy 1 (SLY1) (McGinnis et al., 2003; Dill et al., 2004; Fu et al., 2004), binding to which is enhanced by interaction with the soluble GA receptors gibberellin-insensitive dwarf 1a/b/c (AtGID1a/b/c) in the presence of GA (Griffiths et al., 2006; Nakajima et al., 2006). The N-terminal DELLA domain is required for binding of the DELLA proteins to the GA–GID1 complex (Griffiths et al., 2006).
Ethylene has been shown to delay the disappearance of a green fluorescent protein (GFP)–RGA fusion protein from root cell nuclei via a constitutive triple response 1 (CTR1)-dependent signalling pathway (Achard et al., 2003). Furthermore, ethylene was shown to control the maintenance and exaggeration of the apical hook by modifying DELLA degradation (Achard et al., 2003; Vriezen et al., 2004). More recently, it was demonstrated that active ethylene signalling results in decreased GA content, thus stabilizing DELLA proteins (Achard et al., 2007). Hence, ethylene appears to affect DELLA stability primarily via changes in GA concentrations.
In order to reveal how the modulation of ethylene and GA pathways affects global plant growth, the effects of simultaneous alterations in these pathways were assessed at the whole-plant level. In this work, we present the characterization of the gibberellin-insensitive (gai), ethylene-overproducing 2-1 (eto2-1) double mutant, which is negatively affected in GA signalling and positively affected in ethylene biosynthesis. These two effects (reduced GA signalling and enhanced ethylene accumulation) both limit cell expansion. The (semidominant) gain-of-function mutant gai encodes a protein lacking a segment of 17 amino acids from close to the N-terminus, resulting in a reduced response to GAs which cannot be rescued by application of the hormone, as well as an elevated endogenous GA concentrations. The latter arises from loss of feedback regulation, inducing transcription of GA-biosynthesis genes (GA 20-oxidase and GA 3β-hydrolase) (Talon et al., 1990; Cowling et al., 1998). The ethylene-overproducing eto2-1 mutant has a single base pair insertion, disrupting the C-terminal 12 amino acids of 1-aminocyclopropane-1-carboxylic acid (ACC) synthase 5 (ACS5), one of the nine ACS genes in Arabidopsis (Vogel et al., 1998; Chae & Kieber, 2005). The C-terminal region is necessary for proteolytic targeting of ACS5 in wild type; hence, the truncated form of ACS5 in eto2-1 is more stable. Although dark-grown seedlings produced approximately 20 times more ethylene than the wild type, adult tissues from eto2-1 mutants and light-grown eto2-1 seedlings had a concentration of ethylene that was only slightly higher than that of wild-type tissues (Vogel et al., 1998). Therefore, both the developmental stage and the presence of light appear to affect ACS5 activity.
In order to determine the combined effects of the gai and eto2-1 mutations, the growth and development of the double mutant were compared with those of the single-mutant parents and the wild type throughout the entire life-cycle, from hypocotyl growth until flowering. Root and shoot growth were assessed by in vivo imaging, revealing an enhanced shoot:root ratio in the double mutant. Growth inhibition seems to be synergistic in the double mutant, as does floral induction. In addition, ethylene and GA concentrations were measured in all genotypes. Our results indicate that altered GA sensitivity in the gai eto2-1 double mutant, which has an elevated GA content as a consequence, may cause the observed growth alterations through changes in ethylene responsiveness.
- Top of page
- Materials and Methods
- Supporting Information
The phenotypic plasticity of a plant is achieved by integration of environmental and internal signals, collectively orchestrating output at the transcriptome and proteome level, resulting in an optimal growth response (Vandenbussche & Van Der Straeten, 2004). Several of these responses involve crosstalk between the ethylene and GA signalling pathways. The most conspicuous example of such an interaction is the extension growth of deepwater rice (Oryza sativa) upon submergence (Kende et al., 1998; Vriezen et al., 2003). In addition to stress-induced responses, ethylene and GAs also concertedly control germination (Karssen et al., 1989; Ogawa et al., 2003; Chiwocha et al., 2005), cell elongation in hypocotyls (Collett et al., 2000; Saibo et al., 2003) and roots (Achard et al., 2003), formation of stomata (Saibo et al., 2003) and maintenance of the apical hook (Achard et al., 2003; Vriezen et al., 2004). More recently, evidence was found for a role of ethylene in floral transition (Achard et al., 2006, 2007). It was shown that ethylene causes a delay in flowering time in wild-type Arabidopsis and that this effect is almost entirely abolished in a quadruple loss-of-function DELLA mutant (rga-24 gai-t6 rgl1-1 rgl2-1) (Achard et al., 2006).
In this study, we have tried to extend our understanding of interactions between the ethylene and GA pathways by analysis of growth and development of the gai eto2-1 double mutant. Both single mutations lead to negative effects on growth.
Temporal gene expression patterns and protein stability determine the stage of development at which a mutant phenotype will be observed. In wild-type Arabidopsis, ACS5 (At5g65800) expression is very high at the seedling stage, but drops dramatically after development of the first leaf (Schmid et al., 2005; AtGenExpress: http://web.uni-frankfurt.de/fb15/botanik/mcb/AFGN/atgenex.htm). In the eto2-1 background, ACS5 mRNA accumulates to the same levels as in the wild type, but, as a result of an increase in protein stability, ethylene production and responses such as inhibition of cell elongation are enhanced (Vogel et al., 1998). GAI (At1g14920) is highly expressed during germination, but at 6 DAG its expression has already decreased (Schmid et al., 2005); moreover, the mRNA levels in gai were comparable to those in the wild type, while protein stability was enhanced (Peng et al., 1997). Because of the repressor character of GAI, increased protein stability leads to a stronger inhibition of GA responses in the mutant. As GAs are responsible for cell elongation, the lack of GA response causes a dwarfed phenotype in the gai mutant.
Combined gain-of-function mutations in both GAI and ETO2 resulted in strong phenotypes at all stages of development. Although at the seedling stage (until 8 DAG) the double mutant behaved intermediate between the two single mutants, the growth parameters of roots and shoots of older plants (from day 8 onwards) indicated an additive or even synergistic growth inhibition in gai eto2-1 (Figs 1, 4). One possible cause for this double-mutant phenotype could be an additional enhancement of ethylene biosynthesis by GAI. However, measurements of ethylene emanation showed that, in contrast to eto2-1, ethylene biosynthesis was not increased in gai eto2-1. This observation implies that a functional GA-response pathway is required for increased ethylene biosynthesis in the eto2-1 single mutant. It further indicates that the absence of active GA signalling caused by the dominant gain-of-function mutation in GAI negatively influences ethylene biosynthesis in the double mutant. The nature of this interaction remains to be investigated, but, as eto2-1 is affected in the stability of the ACS5 protein, and gai has altered stability in one of the DELLA proteins (Peng et al., 1997; Vogel et al., 1998), it may occur at the protein level, possibly at the 26S proteasome. Our findings support the existence of an additional mode of ACS degradation, which is not mediated by ETO1. Whether or not this pathway is the same as the ETO1-independent cytokinin-inhibited pathway suggested by Chae & Kieber (2005) remains to be investigated. Alternatively, the gai mutation could result in lower ACS5 (eto2-1) gene expression in the double mutant. In contrast to ethylene production, the ethylene responsiveness of gai eto2-1 at low concentrations of ethylene was enhanced in the light as compared with both the wild type and eto2-1 (Fig. 8). Although differential responsiveness was not found in darkness, double mutants displayed a complete triple response in 4-d-old etiolated seedlings in the absence of ethylene (data not shown). Taking these findings together, it can be concluded that, despite the lower ethylene production, the growth inhibition of the double mutant compared with the eto2-1 single mutant may – at least in part – be caused by enhanced ethylene responsiveness. However, as the morphological phenotype of the enhanced ethylene response is subtle, a more detailed analysis is required. To that end, the expression of an ethylene-inducible reporter gene (e.g. EBS::GUS; Stepanova et al., 2007) in the presence of different concentrations of ethylene could be compared in the wild type and the single and double mutants at different developmental stages.
The phenotype of the double mutant was either reminiscent of that of one of the parental lines (eto2-1-like in lateral root formation (Fig. 3), or gai-like in ethylene-emanation (Fig. 7)), additive (in inflorescence length; Supplementary Material Fig. S2) or even synergistic (in shoot surface enlargement; Fig. 4), indicating that the nature of the interaction between ethylene and GA is both spatially and temporally dependent. Ethylene was shown to antagonize the GA response by stabilizing DELLA proteins, RGA in particular (Achard et al., 2003, 2006; Vriezen et al., 2004). However, recent evidence suggests the existence of a DELLA-independent GA pathway. Cao et al. (2006) demonstrated that, even in processes known to be controlled in a DELLA-dependent way, such as seed germination and floral development, only half of all GA-response genes are regulated by DELLA proteins. Moreover, within this group of genes, differences were observed related to the developmental process, as a different set of genes was influenced in germination compared with floral development. It can therefore be speculated that stage-dependent forms of ethylene–GA crosstalk may exist, either based on the described interaction at the level of DELLA proteins, or on a DELLA-independent mechanism. To test this hypothesis, it might be interesting to study the expression pattern of both DELLA-dependent and DELLA-independent GA response genes upon GA treatment of ethylene mutants and a combination with multiple DELLA loss-of-function mutants in parallel with wild type or ga1-3 mutants treated with GA in combination with an ethylene action inhibitor, at several stages of development.
Communication between shoots and roots is a crucial factor determining overall plant architecture. Although gai eto2-1 had the smallest roots and shoots, the shoot:root ratio was the highest for the double mutant (55 and 75% increase compared with wild-type at 10 and 12 DAG, respectively; Fig. 5), indicating stronger relative growth of the shoot as compared with all other genotypes. The enhanced shoot:root ratio indicates a redirection of the energy of the plant to shoot growth, suppressing root development, as indicated by the decrease in root growth rate from day 8 onwards. As more lateral roots were produced in gai eto2-1 plants older than 8 d, a loss of primary root dominance was evident. Shoot growth reached a maximum shortly thereafter (between days 10 and 12). Long-distance source-to-sink transport of components such as sucrose and cytokinins has been shown to be important in floral induction. Although there is no direct evidence yet, it is possible that long-distance transport of GA is also important in these processes (Bernier & Périlleux, 2005). Our observation that the ethylene-induced delay in flowering was increased in the double mutant as compared with both single mutants indicates that an interaction between ethylene and GA pathways may be involved in this process, although gai remains the major contributor to delay in floral initiation.
Measurements of GA content showed clear differences in gai eto2-1 as compared with both single mutants, especially for GA4, which is known to be the most active GA in the vegetative stage of plant development in Arabidopsis (Hooley et al., 1991). It has been reported previously that the GA content in gai is higher as a result of a feedback mechanism on GA biosynthesis (Talon et al., 1990; Phillips et al., 1995; Xu et al., 1995; Cowling et al., 1998). The higher GA concentrations in the gai single mutant and the double mutant were concomitant with higher expression levels of GA-biosynthesis genes, supporting the feedback hypothesis. The enhanced concentrations of GA4 in the double mutant corresponded to higher expression of the three GA20ox genes examined in combination with higher expression of GA3ox1. In the double mutant, the enhanced responsiveness to ethylene could be responsible for further stabilization of DELLA proteins, hyperactivating the feedback mechanism, resulting in higher bioactive GA concentrations. However, this was not reflected in our expression analysis. In particular, the similar levels of GA3ox1 and GA20ox expression in gai and gai eto2-1 appear inconsistent with the strong difference in GA4 content in these genotypes. However, in Arabidopsis, the GA 3-oxidase family consists of at least four members (Hedden et al., 2002), so other GA3ox genes may be responsible for the increased GA4 concentrations in gai eto2-1. Alternatively, another explanation for this discrepancy might be found in a modulation of GA enzyme stability by ethylene. Importantly, not only the concentration of bioactive GA, but also the GA response was enhanced in gai eto2-1 as compared with the gai single mutant. This response may be controlled by a DELLA-independent mechanism, or may suggest that the mutant form of GAI in gai eto2-1 is subject to an ethylene-driven degradation mechanism, probably by post-translational modification of the mutant GAI protein.
In conclusion, our data indicate that the absence of an active GA-signalling cascade suppresses the higher ethylene biosynthesis seen in eto2-1 while the responsiveness to ethylene is slightly enhanced. The suppression of ethylene biosynthesis in the double mutant suggests that the absence of active GA signalling may affect the stability of ethylene-biosynthesis enzymes in a negative feedback mechanism. The enhanced sensitivity to GA in the double mutant indicates a reciprocal influence of the two pathways on one another. This is also corroborated by earlier data demonstrating that ACC enhances the activity of the GA-biosynthesis gene GA1 (Vriezen et al., 2004).