Arabidopsis RGL1 encodes a negative regulator of gibberellin responses

Abstract

In Arabidopsis, the DELLA subfamily of GRAS regulatory genes consists of GAI, RGA, RGA-LIKE1 (RGL1), RGL2, and RGL3. GAI and RGA are known to be negative regulators of gibberellin (GA) responses. We found that RGL1 is a similar repressor of GA responses, as revealed by RGL1 gain-of-function and loss-of-function phenotypes. Repression of GA responses in Arabidopsis was conferred by a dominant 35S-rgl1 transgene carrying a DELLA domain deletion analogous to the GA-insensitive gai-1 mutation. As in GA-deficient Arabidopsis, the transgenic plants were dark green dwarfs with underdeveloped trichomes and flowers. Expression levels of GA4, a feedback-regulated GA biosynthetic gene, were increased correspondingly. Conversely, a loss-of-function rgl1 line had reduced GA4 expression and exhibited GA-independent activation of seed germination, leaf expansion, flowering, stem elongation, and floral development, as detected by resistance to the GA biosynthesis inhibitor paclobutrazol. RGL1 plays a greater role in seed germination than do GAI and RGA. The expression profile of RGL1 differed from those of the four other DELLA homologs. RGL1 message levels were predominant in flowers, with transcripts detected in developing ovules and anthers. As with RGA, green fluorescent protein (GFP)-tagged RGL1 protein was localized to the nucleus, but unlike GFP-RGA, there was no degradation after GA treatment. These findings indicate that RGL1 is a partially redundant, but distinct, negative regulator of GA responses and suggest that all DELLA subfamily members might possess separate as well as overlapping roles in GA signaling.

Figure 1.

Sequence Alignment of the Arabidopsis DELLA Subfamily of the GRAS Transcriptional Regulator Family. The names RGL2 and RGL3 were proposed by Dill and Sun (2001). Symbols shown above and below the sequences denote the following: asterisks, the 17 deleted residues in rgl1Δ17; closed triangles, the DELLA motif; open circles, leucine heptad repeats; open triangles, the putative nuclear localization sequence; closed diamonds, the VHIID motif; closed circles, the LXXLL motif; open diamonds, a potential tyrosine kinase phosphorylation site.

Figure 2.

Phenotypes of the 35S-rgl1Δ17 Transgenic Lines. (A) Comparison of soil-grown plants at the same age. Ecotype Columbia (Col-0) is the wild type for the transgenic lines. gai-1 is a semidominant GA-insensitive mutant (in the Landsberg erecta [Ler] background). 35S-rgl1Δ17 carries the dominant transgene. rgl1Δ17-R is an rgl1 cosuppressed line that was derived from a 35S-rgl1Δ17 line. Bar = 5 cm. (B) Typical 35S-rgl1Δ17 rosette (T1 generation) next to a wild-type (Col-0) rosette of the same age, both grown on soil. Bar = 0.5 cm. (C) Leaf trichome of the wild type (Col-0) grown on soil. Bar = 0.2 mm. (D) Typical underdeveloped leaf trichome of a 35S-rgl1Δ17 transformant grown on soil. Bar = 0.2 mm. (E) Wild-type (Col-0) flower from a soil-grown plant. Bar = 0.5 mm. (F) Flower of a wild-type (Col-0) plant grown on Murashige and Skoog (MS)–agar medium plus the GA biosynthesis inhibitor PAC. Bar = 0.5 mm. (G) Weak phenotype of a 35S-rgl1Δ17 flower (from a soil-grown plant) is comparable to the phenotype shown in (F). Bar = 0.5 mm. (H) Severe phenotype of a 35S-rgl1Δ17 flower from a soil-grown plant. Two of the sepals have been removed to show the underdeveloped petals and anthers. Bar = 1 mm. (I) Wild-type (Col-0) inflorescence from a soil-grown plant. (J) Inflorescence of the wild type (Col-0) grown on MS-agar medium plus PAC. (K) Inflorescence of a 35S-rgl1Δ17 plant grown on soil.

Figure 3.

RGL1 Is Expressed Predominantly in Flowers. RNA gel blots containing ∼12 μg of total RNA from the indicated wild-type tissues were probed individually with the five DELLA subfamily members (RGL1, RGA, GAI, RGL2, and RGL3). The seedlings were etiolated. The lane labeled siliques + GA consists of siliques of mixed ages collected 2 hr after spraying with 100 μM GA₃. At bottom is shown the amount of 28S rRNA per lane in an ethidium bromide–stained gel before blotting. (No signal was detected for the RGL2 probe.)

Figure 4.

RGL1 Is Expressed in Developing Ovules and Anthers. Results are shown for in situ hybridizations to wild-type floral sections using an RGL1 antisense strand probe. The RGL1 sense strand probe, which was used as a negative control, did not give specific signals (data not shown). Floral stage designations are according to Smyth et al. (1990). (A) Longitudinal section of a carpel containing developing ovules at floral stage 11 to early stage 12 shows hybridization to the ovules, with the strongest signals in the inner and outer integuments. (B) Longitudinal section of a stamen primordium (with constricted base) at floral stage 7, including truncated portions of two flanking stamen primordia surrounded by sepal tissue. The hybridization signal is in the developing anther region. (C) Cross-section of a stage 8 flower with signal in the differentiated anthers. (D) Close-up view of a longitudinal section of the locules of differentiated anthers at floral stage 8, with signal in the microspore mother cells.

Figure 5.

RGL1 Transcript Levels Are Reduced Specifically in Line rgl1Δ17-R (a Phenotypic Revertant of 35S-rgl1Δ17), Consistent with Cosuppression. (A) RNA gel blots indicating a specific reduction of RGL1 signal in rgl1Δ17-R. Total RNA samples from inflorescences and rosette leaves of a dominant mutant line (35S-rgl1Δ17), a phenotypic revertant line (rgl1Δ17-R), and the untransformed wild type (Col-0) were probed with RGL1, RGA, GAI, RGL2, RGL3, and GA4 (a GA biosynthetic gene used as a marker for GA response). 35S-rgl1Δ17 and GA4 transcripts are increased in inflorescences and leaves of the 35S-rgl1Δ17 line, whereas expression levels of the other genes are relatively constant among all of the different RNA samples. (B) DNA gel blot of RT-PCR indicating that RGL1 and GA4 transcripts are increased in inflorescences and leaves of the 35S-rgl1Δ17 line but decreased in inflorescences and leaves of the rgl1Δ17-R line compared with those of the wild type. RT-PCR products for RGL1, GA4, and RGA transcripts (from the same RNA samples indicated at the top in [A]) were visualized by DNA gel blot hybridization to their respective probes. RGA and GAI results serve as amplification controls. GA4 message levels also are reduced in GA-treated inflorescences compared with those of the wild type. Inflorescence tissue was collected 2 hr after spraying with 100 μM GA₃. (C) RT-PCR indicating no obvious reduction in RGL2 and RGL3 transcripts in rgl1Δ17-R inflorescences. RT-PCR products were generated using RGL2 and RGL3 gene-specific primers, run on an agarose gel, and visualized by ethidium bromide staining. Total RNA (before reverse transcription) was used as a control template (data not shown).

Figure 6.

Cosuppression of the RGL1 Gene Confers PAC-Resistant Seed Germination. (A) Germination frequency in the rgl1 cosuppressed line (rgl1Δ17-R) compared with that of the constitutive GA response mutant spy-3 (in the Col-0 ecotype) (Jacobsen et al., 1996) and the wild type (Col-0). Error bars show standard deviations. Germination was scored 6 days after stratification. (B) Germination frequency in the rgl1 cosuppressed line (rgl1Δ17-R) compared with that of the gai-t6 null mutant (in the Ler background) and Ler. Error bars show standard deviations. Germination was scored 6 days after stratification.

Figure 7.

Cosuppression of the RGL1 Gene Confers PAC-Resistant Leaf Expansion, Flowering, Stem Elongation, and Flower Development. (A) The wild type (Col-0), rgl1Δ17-R, and spy-3 were grown on MS-agar medium plus 1 μM PAC, except for Col-0 (left), which was grown without PAC. All plants are 2 months old. Bar = 1 cm. (B) Flowers and siliques from plants in (A). Bar = 2 mm.

Figure 8.

PAC Resistance in the Cosuppressed RGL1 Line, the rga-24 Null Mutant, and the gai-t6 Null Mutant. (A) The wild type (Col-0), rgl1Δ17-R, rga-24, and gai-t6 were grown on MS-agar medium plus 0.75 μM PAC. All plants are 50 days old. rga-24 and gai-t6 both are in the Ler background. Bar = 2.5 cm. (B) Inflorescences of rgl1Δ17-R, rga-24, and gai-t6 from plants in (A). A wild-type (Col-0) inflorescence in the absence of PAC is shown for comparison (left). Bar = 2 mm.

Figure 9.

Nuclear Localization of GFP-RGL1 in Transgenic Arabidopsis. (A) GFP-RGL1 detected in the nucleus of a root cell (top), with DAPI staining of the nucleus (bottom) for comparison. (B) GFP alone in a root cell (top), with DAPI staining of the nucleus (bottom) for comparison. (C) GFP-RGL1 localized in the nucleus of a trichome. (D) GFP-RGL1 localized in guard cell nuclei. (E) GFP-RGL1 at 5 and 180 min after GA treatment. (F) GFP-RGL1 at 5 and 180 min after mock GA treatment.