Genetic characterization and functional analysis of the GID1 gibberellin receptors in Arabidopsis

Abstract

We investigated the physiological function of three Arabidopsis thaliana homologs of the gibberellin (GA) receptor GIBBERELLIN-INSENSITIVE DWARF1 (GID1) by determining the developmental consequences of GID1 inactivation in insertion mutants. Although single mutants developed normally, gid1a gid1c and gid1a gid1b displayed reduced stem height and lower male fertility, respectively, indicating some functional specificity. The triple mutant displayed a dwarf phenotype more severe than that of the extreme GA-deficient mutant ga1-3. Flower formation occurred in long days but was delayed, with severe defects in floral organ development. The triple mutant did not respond to applied GA. All three GID1 homologs were expressed in most tissues throughout development but differed in expression level. GA treatment reduced transcript abundance for all three GID1 genes, suggesting feedback regulation. The DELLA protein REPRESSOR OF ga1-3 (RGA) accumulated in the triple mutant, whose phenotype could be partially rescued by loss of RGA function. Yeast two-hybrid and in vitro pull-down assays confirmed that GA enhances the interaction between GID1 and DELLA proteins. In addition, the N-terminal sequence containing the DELLA domain is necessary for GID1 binding. Furthermore, yeast three-hybrid assays showed that the GA-GID1 complex promotes the interaction between RGA and the F-box protein SLY1, a component of the SCF(SLY1) E3 ubiquitin ligase that targets the DELLA protein for degradation.

Figure 1.

Transcript Levels of GID1a, -1b, and -1c throughout Arabidopsis Development and Their Feedback Regulation by GA and DELLA. (A) Developmental expression profiles of Arabidopsis GID1s in Col-0. Absolute transcript levels of individual genes in each tissue sample, as determined by quantitative RT-PCR, were normalized against three stable endogenous control genes (see Supplemental Methods online) and shown relative to the lowest value, that of GID1b in dry seed, which is set at 1. The means of three biological replicates ± se are shown. The value (±se) for GID1a in dry seeds is shown above the bar. (B) GA treatment downregulates GA3ox1 and GID1a-1c mRNA levels in ga1-3 (in the Ler background). The means of three replicates of quantitative RT-PCR ± se are shown. Relative mRNA levels of individual genes after GA treatment were calculated in comparison to the water-treated control at each time point. Similar results were obtained when quantitative RT-PCR was performed using a second set of samples. (C) Relative GA3ox1 and GID1a-1c transcript levels in the wild type, ga1-3, the triple homozygous mutant rga-24 gai-t6 ga1-3, and the transgenic line carrying P_RGA:rga-Δ17 (all lines are in the Ler background). The means of four replicates of quantitative RT-PCR (using two biological replicates) are shown. Error bars indicate the se of the mean. The expression level in Ler was arbitrarily set to 1.0. For (B) and (C), the housekeeping gene GAPC, whose expression is not responsive to GA (Dill et al., 2004), was used to normalize different samples (see Supplemental Methods online). One-way analyses of variance were performed with least significant difference (LSD) multiple comparison tests at an α level of 0.01 using SPSS version 10.0 (Chicago, IL). When two samples show different letters (A to D) above the bars, the difference between them is significant (P < 0.01).

Figure 2.

Phenotypes of the gid1 Mutants. (A) and (B) Aerial portions of 37-d-old wild-type and homozygous mutant plants. Genotypes are indicated in (A). Close-up view of the gid1a-1 gid1b-1 gid1c-1 (left) and ga1-3 (right) mutants is shown in (B). (C) and (D) Close-up views of Col-0 (C) and gid1a-1 gid1b-1 (D) flowers. Sepals and petals have been removed to reveal the anthers and pistil. (E) to (G) Scanning electron microscopy floral images of gid1a-1 gid1b-1 gid1c-1 (E), ga1-3 (Col-0) (F), and Col-0 (G). Sepals and petals have been removed to visualize the anthers and pistil. Bars = 200 μm in (E) and (F) and 500 μm in (G).

Figure 3.

GA-Insensitive Phenotype of gid1a-1 gid1b-1 gid1c-1. (A) Aerial portions of 35-d-old plants treated with (+) or without (−) 100 μM GA₄ as indicated. Genotypes of plants are indicated below. (B) Representative 5-d-old seedling primary roots of selected genotypes, grown on vertical MS plates and treated with (+) or without (−) 0.2 μM GA₄ as indicated. Bar = 1 cm. (C) and (D) Mean length (±se; n = 20) of 5-d-old primary roots (C) and hypocotyls (D) of wild-type (Col-0), gid1a-1 gid1b-1 gid1c-1, and ga1-3 (Col-0) grown on vertical MS plates and treated with (gray bars) or without (black bars) 0.2 μM GA₄. (E) Transcript levels of GA3ox1. The relative levels of GA3ox1 expression in gid1a-1 gid1b-1 gid1c-1 and wild-type (Col-0) plants treated with 0.2 μM GA₄ (gray bars) or water (black bars) for 2 h as determined by quantitative RT-PCR (see Supplemental Methods online). Total RNA used for quantitative RT-PCR analysis was extracted from 10-d-old seedlings grown on MS plates under continuous light. The means of two technical replicates are shown.

Figure 4.

The Triple gid1 Mutant Accumulates a High Level of RGA and Is Suppressed by an rga Null Allele. (A) Elevated RGA protein levels in gid1a-1 gid1b-1 gid1c-1 with or without GA treatment. Total proteins were extracted from seedlings of Col-0 and homozygous mutants after treatment with water (−) or 1 μM GA₄ (+GA) for 1 h. Protein samples (50 μg each) were separated by 8% SDS-PAGE gel, and the blot was probed with affinity-purified rabbit anti-RGA antibodies. Ponceau staining was used to confirm equal loading. Experiments using two additional biological replicates showed similar results. (B) Partial suppression of gid1a-1 gid1b-1 gid1c-1 by rga-28. Col-0 and the homozygous mutants indicated were grown on soil for 37 d under an LD photoperiod. The rga-28 single mutant is not shown because the phenotype is similar to the wild type (Tyler et al., 2004).

Figure 5.

GA-Dependent Interaction between GID1a-1c and DELLA Proteins via the DELLA Domain. (A) GID1a, -1b, and -1c interact with DELLA proteins but not with SLY1 in yeast two-hybrid assays. The presence of 100 μM GA₃ (+GA) in the media enhanced GID1–DELLA interactions. (B) A schematic showing the full-length and truncated GAI and RGA used in the yeast two-hybrid assays. The numbers represent amino acid positions of the start and/or end of deletion constructs. ST, Poly S/T region. (C) The DELLA domain without the Poly S/T region is essential for Arabidopsis GID1 interaction. All assays in (C) were performed in the presence of 100 μM GA₃. In (A) and (C), interaction of DB and AD fusion proteins in the L40 yeast cells was scored by the relative growth in His⁻ media containing 3-AT (5, 10, 30, and 60 mM) and β-gal activity (means ± se; at least nine colonies were assayed for enzyme activity in triplicates). A dash indicates no growth on His⁻ plates at 5 mM 3-AT.

Figure 6.

GA-Dependent RGA and GID1a Interaction in Pull-Down Assays. (A) Recombinant GST-GID1a or GST was used in pull-down assays with lysates prepared from sly1-10 or sly1-10 rga-24 seedlings in the absence (−) or presence (+) of 100 μM GA₄. The first two lanes contain 3.5 μg of total proteins from sly1-10 and sly1-10 rga-24. (B) RGA was pulled down by GST-GID1a in a GA dose-dependent manner. Pull-down assays were performed as described in (A), except with different GA₄ concentrations (as indicated above the blot). In (A) and (B), pull-down samples were separated by 8% SDS-PAGE gel, and the blots were probed with affinity-purified rabbit anti-RGA antibodies. Experiments were repeated once with similar results.

Figure 7.

Enhanced SLY1-RGA Interaction by GA-GID1 in Yeast Three-Hybrid Assays. (A) A diagram showing that binding of the GA-GID1-(AD-RGA) complex with DB-SLY1 allows reporter gene expression in the yeast three-hybrid assays. (B) SLY1 and RGA interaction was only detected in the presence of both GID1a and 100 μM GA₃ (+GA). Interaction of DB and AD fusion proteins in the PJ69-4A yeast cells was scored by the relative growth in His⁻ media containing 3-AT (5 to 60 mM; middle panel) and by β-gal activity (means ± se; at least nine colonies were assayed for enzyme activity in triplicates) (right panel). C, growth in His⁺ media.