Update on Gibberellin Signaling. A Tale of the Tall and the Short

Phenotype of GA-deficient mutants in Arabidopsis and in barley. A, Five-week-old wild-type (WT) Arabidopsis and the GA-deficient ga1-3 mutant. B, Two-week-old WT Himalaya and a GA-deficient mutant grd2 (Chandler and Robertson, 1999). In recent years our understanding of GA metabolism has advanced rapidly with the cloning of most biosynthetic and catabolic genes (Hedden and Phillips, 2000; Olszewski et al., 2002). In Arabidopsis and other plant species, GA metabolism is tightly regulated by the GA-response pathway. This homeostatic mechanism involves the feedback and feedforward regulation of the respective biosynthetic and catabolic genes. In contrast to GA metabolism, much less is known about the perception of bioactive GAs, the downstream signaling components, and the changes in gene expression that elicit physiological responses. The cereal aleurone has provided an ideal system to study GA response, with GAs specifically promoting the rapid expression of genes encoding hydrolases, such as α-amylases (Lovegrove and Hooley, 2000; Olszewski et al., 2002). In addition, aleurone studies have provided evidence for a plasma membrane localized GA receptor and have identified GAMYB as an early GA-response gene that is important in activating expression of α-amylase. Although the cereal aleurone system has vastly improved our knowledge of GA response, it has helped to identify only a few genes encoding primary signaling components. The identification of GA-response mutants in many plants has proven instrumental in elucidating the GA-signaling pathway (Richards et al., 2001; Olszewski et al., 2002). The model plant Arabidopsis initially provided the opportunity to use molecular genetics to isolate genes encoding GA-signaling components (Table I Arabidopsis GA-response mutants Arabidopsis GA-response mutants The conserved domains in DELLA proteins. Among the GRAS protein family, the DELLA proteins contain a unique N-terminal DELLA domain with two highly conserved motifs (named DELLA and VHYNP), and a divergent Poly S/T region. The C-terminal region (named GRAS domain) is conserved among all GRAS family members. Poly S/T, polymeric Ser and Thr; LZ, Leu zipper; NLS, nuclear localization signal. The importance of DELLA proteins as repressors of GA signaling is illustrated by the observations that DELLA loss-of-function mutations suppress the phenotype of ga1-3. In Arabidopsis, the DELLA proteins have overlapping as well as specific roles in controlling plant growth and development. For example, an rga null mutation (rga-24) partially suppresses many of the defects of ga1-3, including those affecting leaf expansion, abaxial trichome initiation, flowering time, stem elongation, and apical dominance (Silverstone et al., 1997). By contrast, the gai-t6 null mutation does not dramatically suppress the ga1-3 phenotype, but in combination with rga-24 completely suppresses the above defects; in fact, the triple mutant (ga1-3/rga-24/gai-t6) has a slender phenotype that resembles GA-overdosed wild type (Dill and Sun, 2001; King et al., 2001). However, the triple mutant seed remains nongerminating and the mutant plants still show GA-deficient floral defects, suggesting a role for the RGL genes in controlling these additional developmental processes. This hypothesis was confirmed by recent studies showing that rgl2 null mutations suppress the germination defect of ga1-3 (Lee et al., 2002). Furthermore, a combination of rga, rgl1, and rgl2 null mutations dramatically rescued the floral development defect and fertility of ga1-3 (Cheng et al., 2004; Tyler et al., 2004). Bioactive GAs control plant growth and development, at least in part, by relieving the repression imposed by DELLA proteins. Unlike the loss-of-function gai-t6 allele, the gai-1 mutant is a GA-insensitive dwarf (Table I). This dominant gai-1 allele is predicted to encode a mutant gai-1 protein lacking 17 amino acids within the highly conserved DELLA motif (Peng et al., 1997). Transgenic plants carrying rga-Δ17 or rgl1 mutant gene that encodes rga or rgl1 proteins lacking the 17 amino acids of the DELLA motif, as in gai-1, also have a GA-insensitive dwarf phenotype (Dill et al., 2001; Wen and Chang, 2002). The DELLA domain deletions appear to produce constitutively active repressors that are not inactivated by GA. It has now been demonstrated that GA induces rapid degradation of the wild-type RGA protein, providing a molecular basis for the GA relief of DELLA repression (Silverstone et al., 2001). In fact, RGA protein levels are significantly reduced 5 to 10 min after GA treatment (S.G. Thomas and T.-p. Sun, unpublished data), indicating that this is a very early event in the GA signaling cascade. Interestingly, the rga-Δ17 mutant protein is resistant to GA-induced degradation (Dill et al., 2001), suggesting that the GA-insensitive dwarf phenotype of the rga-Δ17 plants is caused by a constitutively active rga-Δ17 protein that cannot be targeted for degradation by GA. Other DELLA proteins are also likely to be subject to GA-induced degradation, although recent studies of transgenic Arabidopsis expressing both GAI and RGL1 as green fluorescent protein (GFP) fusion proteins did not observe such an effect (Fleck and Harberd, 2002; Wen and Chang, 2002). By contrast, studies using polyclonal antibodies raised against RGA have demonstrated rapid GA-induced degradation of both endogenous GAI and RGL2 proteins (Dill et al., 2004; Tyler et al., 2004). The early events in response to many plant hormones involve the destruction of key regulatory proteins via the ubiquitin-proteasome pathway (Hare et al., 2003; Vierstra, 2003). Proteins destined for degradation by the 26S proteasome are polyubiquitinated by an E3 ubiquitin (Ub) ligase enzyme complex. Pioneering work in the fields of auxin, jasmonate, and ethylene signal transduction have illustrated the importance of Skp1/cullin/F-box (SCF) E3 Ub ligase complexes in these pathways (Gray et al., 1999; Xu et al., 2002; Guo and Ecker, 2003; Potuschack et al., 2003). SCF E3 Ub ligases are complexes consisting of four main protein components: Skp1, cullin/Cdc53, Rbx1/Hrt1/Roc1, and F-box. Studies in yeast (Saccharomyces cerevisiae) have defined the F-box subunit as providing substrate specificity of the SCF E3 Ub ligase complex. The F-box proteins are characterized by an N-terminal F-box motif, which has a structural role for mediating binding to the Skp1 component of the complex. The C termini of F-box proteins are often required for binding and targeting the substrate for ubiquitination and in many cases contain a conserved protein-protein interaction domain for this purpose. Studies in yeast and mammalian cells have demonstrated that proteins targeted for ubiquitination are posttranslationally modified, usually by phosphorylation. It is the modified form of the target protein that is recognized by the F-box component of the SCF E3 Ub ligase. The importance of SCF E3 Ub ligases in controlling plant growth and development is illustrated by recent predictions of almost 700 genes encoding F-box proteins in Arabidopsis (Gagne et al., 2002). A role for the ubiquitin-proteasome pathway, and more specifically SCF E3 Ub ligases, in GA signaling comes from work with SLY1, which encodes a predicted F-box protein (McGinnis et al., 2003). SLY1 is a small protein of 151 amino acids, containing an N-terminal F-box motif. Although its C terminus does not have a characteristic protein-protein interaction domain, this region is clearly important for SLY1 function as both of the loss-of-function sly1 alleles are predicted to encode sly1 proteins with C-terminal truncations. The Arabidopsis genome encodes a predicted protein, MIF21.6, which is 30% identical to AtSLY1 over a 124-amino acid region. MIF21.6 contains an N-terminal F-box domain but demonstrates higher sequence identity over the C-terminal domain (51% identity over 51 amino acids). Thus, MIF21.6 may also be involved in GA signal transduction. Several lines of evidence indicate that SLY1, presumably as part of the SCFSLY1 complex, targets the DELLA proteins for GA-induced degradation by the ubiquitin-proteasome pathway (McGinnis et al., 2003; Dill et al., 2004). In the sly1 mutants, RGA and GAI protein levels are dramatically elevated and protein levels are not reduced by GA treatment. Furthermore, the dwarf phenotype of sly1 is completely suppressed by a combination of rga and gai null alleles (Dill et al., 2004). The triple sly1/rga/gai mutant still demonstrates germination and floral defects, suggesting that SLY1 also plays a role in GA-induced degradation of RGL proteins. A direct interaction of SLY1 with RGA and GAI was found in a yeast two-hybrid system, providing further support for a role of SLY1 in recruiting DELLA proteins to the SCFSLY1 E3 Ub ligase complex. Yeast two-hybrid assays have also been performed to closely map the domains of GAI necessary for the interaction with SLY1 (Dill et al., 2004). Surprisingly, the DELLA motif of GAI or RGA is not required for their interaction with SLY1, whereas the GAI GRAS domain alone is sufficient to interact with SLY1 in yeast. The DELLA domain may therefore be important for the GA-induced posttranslational modification of the DELLA proteins. Future studies will be needed to verify this hypothesis, as well as to identify additional components in the SCFSLY1 E3 Ub ligase complex and the regulation of assembly and/or activity of this complex by GA. It has long been known that there is cross talk between different phytohormone signaling pathways (Gazzarrini and McCourt, 2003). A recent microarray study aimed at identifying GA-responsive genes in germinating Arabidopsis seeds demonstrated that GA may affect both ethylene biosynthesis and response, as well as auxin biosynthesis and transport (Ogawa et al., 2003). In addition, both auxin and ethylene can affect GA-regulated root and hypocotyl growth by modifying the stability of RGA (Achard et al., 2003; Fu and Harberd, 2003). The molecular mechanism(s) involved in the interactions among these hormones remain to be determined. Considering the central role of SCF E3 Ub ligase-mediated protein degradation in each of the auxin, ethylene, and GA signaling cascades, it will be intriguing to see whether these SCF complexes provide the molecular link to this phytohormone cross talk. Biochemical and proteomics approaches should help to provide answers to these questions. The importance of SPY as a negative regulator of GA signaling is clearly illustrated by the finding that spy mutations partially suppress all of the GA-related growth defects of the ga1 mutants (Jacobsen and Olszewski, 1993; Silverstone et al., 1997). SPY encodes a novel Arabidopsis protein that shows sequence similarity to animal O-linked N-acetylglucosamine (O-GlcNAc) transferases (OGTs) over its C terminus (Thornton et al., 1999). Support for SPY being a functional OGT is provided by the findings that SPY has in vitro OGT activity and that spy mutants exhibit a reduction in O-GlcNAcylated proteins. The SECRET AGENT (SEC) gene encodes a second Arabidopsis OGT, which also demonstrates OGT activity when expressed in an Escherichia coli system (Hartwek et al., 2002). Although the sec mutant has a wild-type phenotype, the spy/sec double mutant is synthetic lethal in gamete and seed development, indicating that OGT is essential in these developmental stages and that SPY regulates other cellular pathways in addition to GA signaling. It is unclear whether SEC plays a minor role in GA signaling. The N termini of both SPY and SEC have tetratricopeptide repeat (TPR) domains that appear necessary for mediating protein-protein interactions. Overexpression of the SPY TPR domain in Arabidopsis confers a spy-like phenotype (Tseng et al., 2001), suggesting that the overexpressed TPR domain may block SPY function by forming inactive protein complexes and/or by interacting with SPY substrates and preventing modification. O-GlcNAcylation of target proteins in animal systems is specific to Ser or Thr residues and is a dynamic modification analogous to phosphorylation that can affect the activity of target proteins (Wells et al., 2001). In addition, O-GlcNAcylation may compete with phosphorylation for identical or close-by target residues. Target proteins of SPY have not been identified, but potential candidates include the DELLA proteins (Olszewski et al., 2002). To improve our understanding of the roles of GAs in regulating Arabidopsis growth and development, it is essential to study the GA-mediated changes in spatial and temporal patterns of gene expression. DNA microarray analysis provides a powerful tool to identify GA-responsive genes. A recent study identified 138 up-regulated and 120 down-regulated genes after a 12-h GA4 treatment of imbibed ga1-3 seeds, using the oligo-based microarrays representing about 8,200 Arabidopsis genes (Ogawa et al., 2003). A subset of the GA-regulated genes in ga1-3 was identified in germinating wild-type seeds, the expression of which demonstrated a direct correlation with increases in endogenous levels of the bioactive GA4. Previous studies in Arabidopsis have identified potential promoter elements that may have a role in GA responsiveness (Blazquez and Weigel, 2000). An 8-bp cis-element was identified in the promoter of the floral meristem identity gene LEAFY as being necessary for GA-responsive expression. This element is a potential MYB transcription factor binding site. Furthermore, the expression of AtMYB33, a GAMYB-like gene in Arabidopsis, was detected in the shoot apex prior to LEAFY expression (Gocal et al., 2001). Only a single GA-responsive gene identified in the study by Ogawa et al. (2003) contained the GA-responsive LEAFY cis-element. This indicates that this element does not serve as a common cis-element for GA-responsive genes. Interestingly, a previously identified abscisic acid-responsive element was found in many of the promoters of the GA down-regulated genes. Because DELLA proteins are thought to be transcriptional regulators, it is critical to identify the genes they control. The use of DNA microarray studies to establish global gene expression profiles in DELLA loss-of function and gain-of-function mutants should help to identify candidate target genes of DELLA proteins. In the search for improved crops, geneticists and plant breeders have generated and characterized many mutants. These collections include GA-deficient and -response mutants that have proven immensely important in both furthering our understanding of GA action as well as increasing agronomic yields. In particular, many GA-response mutants have been identified in cereals. These GA-insensitive mutants fall into the same categories as those seen in Arabidopsis, showing either a GA-dwarf or -overdose (slender) phenotype. In contrast to Arabidopsis GA-overdose mutants, the corresponding cereal slender mutants display a more dramatic, easily observed phenotype (Olszewski et al., 2002). The importance of GA-response mutations in producing commercially important crops is clearly illustrated by the Rht semidominant dwarfing alleles of wheat (Triticum aestivum; Perkins, 1997). These dwarfing mutations have been essential in producing higher yielding lines that are also more resistant to wind damage; these dwarfing alleles have been a crucial component of the development of Green Revolution varieties and are critical for the success of the Green Revolution. In addition to Rht in wheat, dominant mutations producing GA-insensitive dwarf plants have also been identified in other species, including maize (Zea mays; D8), rice (Oryza sativa; slr1), barley (Hordeum vulgare; sln1-d), Brassica rapa (dwf2), and grape (Vitis vinifera; Vvgai1; Peng et al., 1999; Ikeda et al., 2001; Boss and Thomas, 2002; Chandler et al., 2002; Muangprom and Osborn, 2004). Identification of DELLA protein genes in Arabidopsis has facilitated the cloning of orthologs in other plant species. Interestingly, the dwarfing cultivars described above all contain gain-of-function mutations in genes encoding DELLA protein orthologs (Peng et al., 1999; Boss and Thomas, 2002; Gubler et al., 2002). Moreover, these dwarfing alleles are predicted to encode mutant DELLA proteins that are altered in the N-terminal region encompassing the highly conserved DELLA domain. Similar to Arabidopsis, GA opposes the action of SLR1 and SLN1 in rice and in barley, respectively, by targeting their rapid degradation. The DELLA motif-mutated slr1 and sln1-d proteins are resistant to GA-induced degradation (Gubler et al., 2002; Itoh et al., 2002), with of the rga-Δ17 protein in These observations that the role of the DELLA proteins in controlling GA signaling is conserved between and to the dwarf phenotype caused by the dominant slr1 and mutations in these a GA-insensitive slender phenotype in both rice and barley et al., 2001; Chandler et al., 2002). to contain only a single DELLA protein gene et al., 2001). this gene in a constitutively active GA response and a dramatically phenotype. By contrast, a single DELLA protein gene in Arabidopsis only GA to functional among the DELLA family (Dill and Sun, 2001; King et al., 2001). The of GA-signaling components, with the of the aleurone system, an system to GA signal transduction. The study of DELLA proteins in has the between GA signaling in and in addition to our understanding of the role of DELLA proteins. To further the role of SLR1 in GA functional domain analysis was performed by of SLR1 as in transgenic rice et al., 2002). in the highly conserved N-terminal DELLA and motifs GA-insensitive dwarf plants in which the fusion proteins not in response to GA. This other studies described above that show the N terminus of DELLA proteins is essential for the GA signal. Interestingly, of an N-terminal region S/T to the GRAS domain, a dwarf that has reduced to GA. However, the Poly S/T motif does not to GA-induced degradation or an elevated mutant protein It was that this domain may be modified by phosphorylation or O-GlcNAcylation to the activity of in the C-terminal GRAS domain of SLR1 appear to which the findings that null alleles of DELLA proteins often contain or deletions in the GRAS region (Silverstone et al., Ikeda et al., 2001; Gubler et al., 2002). In addition, the mutant proteins with the C-terminal are not in response to GA. These findings support a role for the GRAS domain as both functional and regulatory The C terminus also contains a Leu motif, which may a role in of SLR1 of a C-terminal form containing the Leu motif a dominant negative producing a slender phenotype in the wild-type It was that the form this effect by forming inactive with wild-type SLR1 et al., 2002). Studies described above have clearly illustrated that DELLA proteins have a highly conserved role in GA signaling in different species. The controlling GA-mediated degradation of DELLA protein also appear to be highly at least in Arabidopsis and An important study in rice by et al. (2003) identified the alleles that produce GA-insensitive dwarf The gene encodes an F-box protein that is highly to and support a role for as part of a E3 Ub ligase complex, in GA-induced degradation of SLR1 in The SLR1 protein is elevated in the and of SLR1 are observed after GA treatment. Furthermore, GA phosphorylation of SLR1 in the mutant et al., and SLR1 protein specifically to in an in vitro et al., 2004). These that phosphorylation of SLR1 is necessary for its degradation. Studies of plants other Arabidopsis have novel potential GA-signaling components. In the alleles have a GA-insensitive dwarf phenotype, suggesting that is a regulator of GA signaling et al., The gene encodes a of the protein et al., 1999; et al., 1999). This finding suggesting a role for proteins in GA response et al., However, it is important to that also has additional roles in other signaling including et al., 2002). In Arabidopsis, is by a single In contrast to the loss-of-function mutants not display a dwarf phenotype, although they are less to GA germination et al., 2002). of the mutant a role for in regulating signaling In was identified as an up-regulated gene in the et al., 2001). Transgenic plants expressing an a phenotype. These plants reduced response to GA and higher levels of endogenous GAs to control These findings support a role as a regulator of GA signaling. Furthermore, to of GA treatment induces of the fusion protein to the in The predicted protein has been to contain a which is in a of Ub E3 ligase proteins et al., 2003). It will be to whether plays a role in GA-mediated degradation of DELLA proteins. It is the of this to a of GA signaling and response in cereal aleurone However, there are recent this (Lovegrove and Hooley, 2000; Olszewski et al., 2002). The aleurone has a of that it an ideal system to study GA signaling and response using Although aleurone cells not produce transcription of a of genes encoding is by GA treatment. promoters of these when to are of GA response in the aleurone In addition, GA-responsive can be easily from aleurone cells for expression GAMYB is a transcription factor that regulates expression of genes in cereal aleurone cells (Gubler et al., 1999). In barley, GAMYB has been to specifically to the GA-response element of genes and their expression. The identity of primary GA-signaling components in has provided the opportunity to their roles in controlling GAMYB of expression et al., Gubler et al., 2002; et al., 2002). For example, and are regulators, whereas DELLA proteins and SPY are negative of the GA response in aleurone Studies aimed at understanding the of GA response have that GAMYB expression is after aleurone cells with GA. By contrast, SLN1 degradation within 10 min (Gubler et al., 2002). It will be important to the molecular events that after degradation. The use of aleurone cells in DNA microarray provides an ideal opportunity to these early transcriptional changes caused by GA. A model of GA-signaling pathway in and the indicate and are by and plasma Although our understanding of the GA signal transduction has improved in recent there are still in our approaches have proven highly in identifying GA-signaling components in a of plant species. GA signaling components identified in other plants can now be in For example, Arabidopsis has genes encoding closely proteins to et al., 2003). the of Arabidopsis genetics and the cereal aleurone system will to to the of known GA-signaling components and to novel in the GA-response pathway. The DELLA proteins are highly conserved in higher plants and are central of GA-mediated growth and developmental These findings provide an important opportunity to the GA-signaling pathway, both downstream and of the DELLA proteins. DNA microarray studies should help in the transcriptional events that DELLA protein degradation and may also identify potential DELLA target genes. In addition, and proteomics approaches have the potential to further the events promoting DELLA protein degradation. It is also that these approaches may the GA Phillips, and Silverstone for the are also to Chandler for providing the in for not being to all of the studies and for to the