Multiple loss-of-function of Arabidopsis gibberellin receptor AtGID1s completely shuts down a gibberellin signal


(fax +81 3 5841 8025; email or fax +81 29 836 9053; email


Arabidopsis carries three receptor genes for the phytohormone gibberellin (GA), AtGID1a, AtGID1b and AtGID1c. Expression of each gene in the rice gid1-1 mutant for GA receptors causes reversion of its severely dwarfed phenotype and GA insensitivity to a normal level, even though each loss-of-function mutant shows no clear phenotype in Arabidopsis (Nakajima et al., 2006). In this paper, we report the functional redundancy and specificity of each AtGID1 by analyzing the multiple mutants for loss of function. Seeds of the double knockout mutants atgid1a atgid1b, atgid1a atgid1c and atgid1b atgid1c germinated normally. The double knockout mutant atgid1a atgid1c showed a dwarf phenotype, while other double mutants were of normal height compared to the wild-type. The stamens of the double knockout mutant atgid1a atgid1b were significantly shorter than those of the wild-type, and this leads to low fertility. A severe disarrangement of the pattern on its seed surface was also observed. The triple knockout mutant atgid1a atgid1b atgid1c did not germinate voluntarily, and only started to grow when the seed coat was peeled off after soaking. Seedlings of the triple knockout mutants were severe dwarfs, only a few millimeters high after growing for 1 month. Moreover, the triple knockout seedlings completely lost their ability to respond to exogenously applied GA. These results show that all AtGID1s function as GA receptors in Arabidopsis, but have specific role(s) for growth and development.


One of the classical plant hormones, gibberellin (GA), has multiple functions in the growth, development and environmental response of plants. Several GA signaling factors have been identified by a genetic approach (Sun and Gubler, 2004). Recently, a rice cytosolic GA receptor, OsGID1, was identified by positional cloning of a causal gene of GA-insensitive dwarf 1 (gid1) (Ueguchi-Tanaka et al., 2005), and, using this rice clone, the Arabidopsis orthologs AtGID1a, AtGID1b and AtGID1c were successively identified (Nakajima et al., 2006). Functional potency of the three gene products as GA receptors was demonstrated by detecting their GA-binding activity in vitro and by recovering the wild phenotype by the introduction of each AtGID1 into the gid1 mutant. We characterized some knockout lines of AtGID1s, but could not find any abnormal phenotypes in these lines, probably because of the functional redundancy of these AtGID1s in Arabidopsis (Nakajima et al., 2006), and expression of the AtGID1 genes was indeed found to overlap in various organs of Arabidopsis (Nakajima et al., 2006). The lack of a clear phenotype in single knockout lines is not unusual in the case of functional redundancy of the genes of interest. For example, in the case of the ethylene receptor, the Arabidopsis genome has five genes, ETR1, ETR2, EIN4, ERS1 and ERS2, and single knockout mutant lines of these ethylene receptors do not show the ethylene-insensitive phenotype at all (Hua and Meyerowitz, 1998). Phenotypic abnormality of knockout lines mainly depends on the expression pattern and functional specificity of the redundant genes. In the case of the ethylene receptor again, double knockout lines carrying loss-of-function alleles of ETR1 and ERS1 show a severe constitutive ethylene response phenotype including rosette size and flower development in the light, and a moderate phenotype of etiolated seedlings (Wang et al., 2003), while other double knockout lines do not show any clear phenotypes. Similarly, a clear phenotype of cytokinin, the inhibition of the growth and development of roots, leaves, and influorescence meristems, was seen only in a triple knockout mutant of the cytokinin receptor genes AHK2, AHK3 and AHK4/CRE1/WOL (Nishimura et al., 2004; Riefler et al., 2006). These examples of ethylene and cytokinin receptor mutants show that, if we wish to observe a knockout phenotype of the Arabidopsis GA receptor mutant, we will have to produce double or triple knockout lines, and that the characterization of double knockout mutants of AtGID1 genes will reveal the redundancy of each AtGID1 function in Arabidopsis. Thus we prepared double knockout lines carrying all combinations of AtGID1a, AtGID1b and AtGID1c, and also a triple mutant of these three genes.

Results and discussion

Ds or T-DNA insertion lines for AtGID1

In this study, we obtained two mutant lines for AtGID1a (atgid1a-1 and atgid1a-2), one for AtGID1b (atgid1b-1) and one for AtGID1c (atgid1c-1) by in silico screening of various T-DNA or Ds element tagging libraries. As these mutant lines are derived from two different ecotypes, Nossen (atgid1a-1 and atgid1b-1) and Columbia (atgid1a-2 and atgid1c-1), we used these two ecotypes as control plants. Figure 1(a) shows the insertion positions of T-DNA or the Ds element in each AtGID1 gene. To characterize the genotype of each mutant, we amplified the fusion sequences carrying T-DNA/Ds and AtGID1 or the sequences of AtGID1 alone by PCR using primers corresponding to the T-DNA/Ds sequences and each AtGID1 sequence. Complete repression of the target expression in each knockout line was confirmed by RT-PCR using total RNA prepared from seedlings as the template (Figure 1b). Seeds of single knockout lines germinated and grew normally. We observed no clear differences in plant height between each mutant line and its parental wild-type plant after analyzing over 10 seedlings of each line at both the young seedling and fruiting stages. Figure 1(c) shows whole mutant and parental wild-type plants at the fruiting stage.

Figure 1.

 Characteristics of atgid1 single knockout mutants.
(a) Location of T-DNA or Ds element insertions in each line for AtGID1. The line Ds13-1770-1 (ecotype Nossen) for AtGID1a (At3g05120) was designated as atgid1a-1, SALK_044317 line (ecotype Columbia) for AtGID1a was designated as atgid1a-2, line Ds11-3970-1 (ecotype Nossen) for AtGID1b (At3g63010) was designated as atgid1b-1, and SALK_023529 line (ecotype Columbia) for AtGID1c (At5g27320) was designated as atgid1c-1. Black boxes indicate exons.
(b) RT-PCR expression analyses of AtGID1 genes in the leaves of each line. Wn, wild-type Nossen; Wc, wild-type Columbia; a-1, atgid1a-1; a-2, atgid1a-2; b-1, atgid1b-1; c-1, atgid1c-1. The numbers of PCR cycles were 35 for all AtGID1 genes and 25 for Act2 (At3g18780).
(c) No clear phenotype of atgid1 single knockout seedlings was apparent at the fruiting stage. Over ten seedlings were observed and showed similar tendencies to the figure. Wn, wild-type Nossen; Wc, wild-type Columbia; a-1, atgid1a-1; a-2, atgid1a-2; b-1, atgid1b-1; c-1, atgid1c-1. Bar = 5 cm.
(d) Reduced response of atgid1b single knockout seeds to a lower level of GA. Imbibed seeds (= 35–40) after a cold treatment (4°C) were incubated in a mixed solution of uniconazole (50 μm) plus increasing GA4 (0–10 μm) for 3 days at 23°C, then radicle emergence was determined under an optical microscope. The germination rate was evaluated on the basis of the rate in the presence of 10 μm GA4. Values are means ± SD for three independent experiments.

Reduced response of the atgid1b single mutant at a lower level of GA4

It is well known that seed germination is a process that is sensitive to GA, and it is necessary to break the dormancy of seeds. In fact, GA-deficient seeds or wild-type seeds treated with GA biosynthesis inhibitors do not germinate without exogenous treatment with GA. We examined the effect of exogenously applied GA4 on the germination of single knockout mutant seeds in the presence of uniconazole, a GA biosynthetic inhibitor. Imbibition and chilling of the seeds were carried out at 4°C for 2 days in a mixture of uniconazole (50 μm) and various concentrations of GA4 (0–10 μm). The relative germination rate was evaluated on the basis of the rate in the presence of 10 μm GA4 after cultivating for 3 days at 23°C on filter paper moisturized with the same solution used for imbibition (Figure 1d). This experiment was repeated three times using 35–40 seeds per batch. When 0.01 μm GA4 was added to the medium, the germination rate of most of the single knockout mutants was almost equivalent to that of the parental wild-type (data not shown), but atgid1b-1 showed a lower germination rate than the wild-type and other mutants, which was not observed above 0.1 μm of GA4.

The reduced frequency of atgid1b-1 seed germination at a lower concentration (0.01 μm) of exogenously applied GA4 but not at an intermediate or higher concentration (0.1–1 μm) of GA4 corresponds well with our previous observation that AtGID1b has about a sixfold higher affinity for GA4 compared with AtGID1a and AtGID1c (Nakajima et al., 2006). Based on the in vitro GA binding results, this suggests that AtGID1b functions as a dominant GA receptor at lower GA concentrations in germination as previously suggested (Nakajima et al., 2006).

Phenotypic analysis of three double knockout mutants

We prepared all possible double knockout mutant lines, i.e. atgid1a-1 atgid1b-1, atgid1a-2 atgid1c-1 and atgid1b-1 atgid1c-1, by crossing the homozygous single knockout AtGID1 mutants. Only the atgid1a-2 atgid1c-1 double knockout mutant showed a dwarf phenotype, while the other double mutants did not show such a phenotype (Figures 2a,b). As dwarfism of the atgid1a-2 atgid1c-1 double mutant was recognizable at a young seedling stage (Figure 2d) and as expression of AtGID1b was detected in this double knockout seedling (data not shown), we conclude that AtGID1b has a minor contribution to the stem growth of Arabidopsis. In contrast, the other double knockout mutants suppressing the expression of AtGID1a and 1b or AtGID1b and 1c did not show dwarfism, indicating that AtGID1a or AtGID1c is sufficient for GA perception for normal growth of Arabidopsis seedlings. This conclusion does not take into account the fact that there may be subtle growth defects in the atgid1b atgid1c double mutant, which is the product of two different ecotypes. To assess this possibility, we selected three independent F3 lines of the two-ecotype hybrid at random, and detected no clear differences between their growth and that of the parental Nossen or Columbia wild-type plants (data not shown). With regard to the root growth of these double knockout mutants, we could not discern any clear differences between them (data not shown). We also examined the GA sensitivity of double mutants by the feedback repression of GA 3-oxidase expression, which is negatively regulated by the level of GA. GA-dependent suppression of GA 3-oxidase expression was observed in all double mutants, wild-type plants (Nossen or Columbia; Figure 2e), and the three F3 lines of the two-ecotype hybrid (data not shown). Similarly, we also detected an increase in the expression of GA-responsive expansin (At-EXP1) following GA treatment in all lines (Figure 2e). These expression results indicate that all the double mutants retain the ability to perceive GA.

Figure 2.

 Characteristics of atgid1 double knockout mutants.
(a) Gross morphology of double knockout seedlings. Seeds were imbibed in distilled water (DW) for 2 days at 4°C, and grown at 23°C. The picture was taken at 56 days after sowing. Wn, wild-type Nossen; Wc, wild-type Columbia; ab, atgid1a-1 atgid1b-1 (Ns); ac, atgid1a-2 atgid1c-1 (Col); bc, atgid1b-1 atgid1c-1 (Ns/Col). Bar = 5 cm. Five seedlings were observed and showed similar tendencies to the figure.
(b) The atgid1a-2 atgid1c-1 double mutant shows a dwarf phenotype. The height (mean ± SD) for each line was calculated from five seedlings of each line at 56 days after sowing.
(c) The atgid1a-1 atgid1b-1 double mutant produces fewer seeds. The value ‘seeds per plant’ was calculated by weighing seeds from at least five seedlings on the basis of the average weight of 400 seeds of each line. The mean ± SD was calculated from three independent experiments.
(d) Time course of the double knockout growth. The length of the bolting stem was measured, and the mean ± SD was calculated from five seedlings of each line.
(e) Normal GA response of gene expression in the double knockout mutant. Young seedlings were incubated on a filter containing 50 μm uniconazole in 0.2% v/v DMSO for 3 days, and then transferred onto a filter containing GA4 (10 μm in 0.2% v/v DMSO; right lane, +G) or 0.2% v/v DMSO as a reference (left lane). Total RNA was prepared from 10 seedlings of each line 12 h after applying GA. The numbers of PCR cycles were 35 (GA3ox1; At1g15550), 31 (At-EXP1; At1g69530) and 24 (Act2). This experiment was repeated twice.
(f) Abnormal structure of atgid1a-1 atgid1b-1 double knockout flowers. (f1–f3) atgid1a-1 atgid1b-1 (Ns); (f4–f6) wild-type (Ns). Bars = 1 mm.

At the reproductive stage, the atgid1a-1 atgid1b-1 mutant showed a lower yield of seeds per plant than the other mutants (Figure 2c). We observed the morphology of the mutant flowers using optical microscopy. Most pollen adhered to an incorrect region below the stigma, which was probably caused by the incomplete elongation of stamens, resulting in low fertility (Figure 2f). Such abnormalities were not detected in other double and single knockout mutants (data not shown). It has been reported that GA is important for flower development, especially for stamen development (Fleet and Sun, 2005). One of the primary hormones responsible for flower induction and subsequent development in vitro of several species was exogenously applied GA (Taylor and Staden, 2006). A high amount of GA has been detected in the anthers (Kobayashi et al., 1988) and pollen (Hasegawa et al., 1995) of rice, and a decrease in the GA level by over-expression of a GA catabolic enzyme causes a reduction in the elongation of pollen tubes in Arabidopsis (Singh et al., 2002). To examine whether pollen of atgid1a-1 atgid1b-1 was fertile or not, we performed artificial self-pollination in some atgid1a-1 atgid1b-1 flowers using a cotton bud, and found that the yield of seeds was restored to normal levels (data not shown). Thus, the lower yield of seeds in the atgid1a-1 atgid1b-1 mutant is not caused by lower pollen fertility, but is probably due to a defect in filament elongation.

In the process of seed maturation of Arabidopsis, starvation of GA influences the normal development of the seed. For example, in GA-deficient mutants such as ga1-3 and in a knockout line for GA 3-oxidase, the hexagonal pattern of the seed surface is severely disarranged, but this is rescued by the application of GA (Kim et al., 2005). The lower yield of atgid1a-1 atgid1b-1 seeds (Figure 2c) led us to speculate that the process of seed development in this mutant does not proceed normally, and that the pattern on the seed surface may be disarranged similarly to GA-deficient mutants. As expected, an irregular pattern on the seed surface was observed in the atgid1a-1 atgid1b-1 mutant (Figures 3a–e), while other mutant lines and wild-types showed a normal pattern (Figure 3 and Figure S1). This demonstrates that AtGID1c has a minor contribution to the formation of seed surface pattern in Arabidopsis.

Figure 3.

 Disarranged surface of the atgid1a atgid1b double knockout seeds by scanning electron microscopy.
(a–e) atgid1a-1 atgid1b-1 (Ns); (f–j) atgid1a-2 atgid1c-1 (Col); (k–o) atgid1b-1 atgid1c-1 (Ns/Col); (p–t) atgid1a-1 (Ns); (u–y) atgid1b-1 (Ns). (a, f, k, p, u) Images of two whole seeds selected at random from each pool (bars = 100 μm). (b, d, g, i, l, n, q, s, v, x) Magnified images of the two squares in (a, f, k, p, u; bars = 50 μm). (c, e, h, j, m, o, r, t, w, y) Further magnified images of the squares in (b, d, g, i, l, n, q, s, v, x; bars = 10 μm). The scanning electron microscope observations were repeated with over ten seeds of each line, and showed similar tendencies to these images.

Phenotype of the triple mutant

By crossing the atgid1a-1 atgid1b-1 double mutant with the atgid1c-1 single mutant, we obtained several seedlings carrying the genotype atgid1a-1/AtGID1a atgid1b-1 atgid1c-1 (Aabbcc), and harvested their seeds by self-pollination. The frequency of the triple knockout seeds (aabbcc) is expected to be a quarter of the progeny, and the triple knockout seeds cannot germinate because the GA signal is essential for seed germination. Thus, we examined the germination rate of the progeny from the Aabbcc lines. We used approximately 200 seeds for this experiment, and found that the ratio of germinated and ungerminated seeds was 3:1 (χ2 = 1.40, P < 0.05) as expected. To directly examine the genotype of germinated and ungerminated seeds, we performed PCR analysis. All germinated seeds had either the Aabbcc or AAbbcc genotype but not aabbcc, while ungerminated seeds had all three genotypes (Figure 4a). These results demonstrate that the triple knockout mutant does not germinate, as for GA-deficient mutants, providing genetic confirmation that AtGID1a, 1b and 1c function redundantly as GA receptors in Arabidopsis.

Figure 4.

 Characteristics of the atgid1a-1 atgid1b-1 atgid1c-1 triple knockout mutant.
(a) No germination of triple knockout seeds. Fifty to 60 seeds from an atgid1a-1/AtGID1a atgid1b-1 atgid1c-1 double knockout (Aabbcc) line were used per experiment. The genotype of each seed was determined by nested PCR with genomic DNA prepared either from part of a germinated seedling or from an ungerminated seed without seed coat. The mean ± SD was calculated from three independent experiments.
(b) Severely dwarfed phenotype of the triple knockout seedling. The picture was taken 34 days after the decoating treatment. Aabbcc seedlings were used as references. The genotype of each seedling was determined by nested PCR using one leaf each. Over ten seedlings were observed and showed similar tendencies. Numbers on the scale indicate millimeters. The inset shows a magnified image of a triple knockout seedling using an optical microscope. Bar = 1 mm.
(c) The more severe phenotype of the triple knockout mutant than of the GA-deficient mutant ga1-3. The images were taken 50 days after the decoating treatment. Eight seedlings were observed and showed similar tendencies. Each interval of the scale indicates 1 mm.
(d) No enhancement of seed germination from an Aabbcc line by the application of GA. The seeds (= 30–40) were pre-incubated in 0.2% v/v DMSO (−), 50 μm uniconazole (+Uni) or 50 μm uniconazole plus 10 μm GA4 (+Uni + GA) for 3 days at 4°C, then incubated on wet filter paper for 3 days. The mean ± SD was calculated from three independent experiments.
(e) Loss of GA responsiveness of the triple knockout seedlings. Decoated seedlings of each line (= 8) were placed on an MS plate, then GA4 solution (10 μm in 0.2% v/v DMSO) was applied after 3 days (right lane), while 0.2% v/v DMSO was applied as a reference (left lane). All pictures were taken 1 week after the GA treatment. Bar = 1 cm.
(f) Response failure of GA-responsive gene expression in the triple knockout seedlings. Total RNA was prepared from 10 seedlings of the triple mutant or thw double mutant (AAbbcc/Aabbcc) 12 h after applying GA (10 μm GA4 in 0.2% v/v DMSO). The numbers of PCR cycles were 36 (GA3ox1), 31 (At-EXP1), 34 (AtGID1a) and 25 (Act2). This experiment was repeated twice, and very similar tendencies were shown for all PCR runs.

There are some reports that, although seeds of severe GA-deficient mutants do not germinate under normal conditions, they can germinate after artificial abrasion of seed coats (Baskin and Baskin, 1971; Debeaujon and Koornneef, 2000). Thus, we removed the seed coat of ungerminated seeds under an optical microscope. The treated seeds germinated under continuous light for 3 days on an MS plate, and developed into seedlings. This artificial abrasion of seed coat was essential for germination of the ungerminated seeds, and none of them germinated without this treatment even after 6 days incubation. These results show that, even though the triple knockout mutant lacks the ability to germinate under normal conditions, this lack of germination ability is not caused by the lethality of seeds.

The triple mutant seedlings showed very severe dwarfism, while the seedlings of the parental double mutant (Aabbcc) showed a weak phenotype (Figure 4b). Such severely dwarfism is similar to the gid1 mutant of rice, which has only one GID1 gene in its genome (Ueguchi-Tanaka et al., 2005), and therefore GID1 protein(s) are very important for GA perception both in Arabidopsis and rice. As for the triple mutant seed, we removed the seed coat of the GA-deficient mutant ga1-3, which does not germinate in the absence of GA, and compared its growth with that of the triple mutant (Figure 4c). We determined that the phenotype of the triple mutant was more severe than that of ga1-3, which may be caused by the entire loss of the GA-independent interaction of AtGID1, especially of AtGID1b with DELLA in the triple mutant (Nakajima et al., 2006). Moreover, we chilled both seedlings at 4°C for 1 month, which is effective to induce flowering of ga1-3 without GA treatment (Michaels and Amasino, 1999). However, we could not detect any bolting of the stem or flowering of the triple mutant even after 40 days of cold treatment (data not shown).

Next, we examined the reactivity of the triple knockout mutant with exogenously applied GA4 in order to investigate whether GA perception in Arabidopsis depends only on the three GID1 proteins or not. To clarify this, three approaches were used. First, we studied the effect of GA4 on germination of seeds from the Aabbcc plants. The seeds were germinated on culture medium only, or with 50 μm of uniconazole or a mixture of 50 μm uniconazole and 10 μm GA4 for 4 days, using 30–40 seeds for each treatment. When added alone, uniconazole completely inhibited seed germination, while the addition of GA recovered the germination frequency to a similar level (approximately 75%) as under control conditions (Figure 4d). As seeds carrying the triple mutant alleles (aabbcc) theoretically form 25% of this seed group derived from the Aabbcc plants, the result strongly suggests that seeds carrying triple mutant alleles did not respond to the GA treatment and failed to germinate. Second, we examined the effect of GA on the seedlings of the triple knockout mutant. The seedlings of triple mutant (aabbcc) and the parental double mutant (AAbbcc) were grown on MS medium with or without 10 μm GA4 for 1 week. Internode elongation was clearly detectable in the double mutant, while no morphological changes in the growth of the stem, leaf, petiole or root recognize were observed in the triple mutant seedlings (Figure 4e). Finally, we examined the expression level of genes regulated by GA in the triple knockout mutant. For this experiment, we selected two genes, namely At-EXP1, whose expression is positively regulated by GA, and GA3ox1, which is negatively regulated by GA. We prepared total RNAs from the seedlings of the triple or double (a mixture of AAbbcc and Aabbcc) mutants 12 h after 10 μM GA4 treatment. In the double mutants in which AtGID1a functioned, expression of GA3ox1 and At-EXP1 decreased and increased, respectively, following the application of GA, as expected (Figure 4f). In contrast, their expressional levels did not change in the triple mutant. Taken together, we conclude that the triple knockout mutant is insensitive to GA. In addition, we found that the expression level of AtGID1a was negatively regulated by GA in the mixture of AAbbcc and Aabbcc plants. To understand the regulatory mechanism of expression of the AtGID1 genes, we analyzed the GA responsiveness of the three AtGID1 genes in imbibed seeds and in stems. As shown in Figure 5(a,b), all AtGID1 genes were clearly negatively regulated by the application of GA in both organs. This strongly suggests that the expression of AtGID1 genes is regulated by the endogenous GA level.

Figure 5.

 GA responsiveness of AtGID1 genes.
(a) The expression of AtGID1 genes during seed germination. Imbibed seeds (Columbia, = 30–40) after a cold treatment (2 days, dark) were incubated in water (DW), 10 μm uniconazole (+Uni), 50 μm GA4 (+GA), or both (+Uni + GA), for the indicated number of hours at 23°C, and then total RNA was prepared. The numbers of PCR cycles were 28 (for all AtGID1 genes and GA3ox1) and 22 (Act2).
(b) Expression of AtGID1 genes during growth of bolting stems. A solution of water (DW), 10 μm uniconazole (+Uni), 50 μm GA4 (+GA), or both (+Uni + GA) was sprayed onto 10 bolting stems directly, and total RNA was prepared after the indicated number of hours. The numbers of PCR cycles were 31 (AtGID1a, AtGID1c and GA3ox1), 33 (AtGID1b) and 22 (Act2). Both experiments were repeated twice, and very similar tendencies were observed.

Closing remarks

In this paper, we have dissected the functional differences among AtGID1s, i.e. AtGID1b is the least responsible for stem elongation while AtGID1c is not vital for flower development, especially for filament growth. In Figure 6, we summarize the possible relationships between functional GID1 proteins and functional DELLA proteins according to organ. The information about DELLA has been fully elucidated following analyses of multiple loss-of-function mutants (Cao et al., 2005; Cheng et al., 2004; Dill and Sun, 2001; Fu and Harberd, 2003; King et al., 2001; Lee et al., 2002; Tyler et al., 2004; Yu et al., 2004). Our next goal is to elucidate the mechanisms of a functional GID1–DELLA combination in planta in order to explain the diverse and complicated mechanisms of GA signal transduction in Arabidopsis. Furthermore, all experiments performed with the triple knockout mutant strongly support the notion that AtGID1s solely or at least dominantly function in GA perception in various GA-related events in Arabidopsis, such as seed germination, stem and leaf elongation, and GA-regulated gene expression. GID1 is also essential in GA perception in rice, which has only one GID1 gene whose loss-of-function phenotype can be observed much more easily than in Arabidopsis. Thus, we can conclude that GID1 protein(s) are the most important factor in GA perception both in dicotyledonous and monocotyledonous plants. However, the results presented above do not exclude the possibility that an alternative GA receptor(s) functions to perceive GA and transduce such GA signals to DELLA proteins or other unknown proteins in Arabidopsis. Further precise studies on the triple mutants are necessary to evaluate the existence of an alternative GA receptor in Arabidopsis.

Figure 6.

 Deduced specificities for GA perception and for signal transduction to DELLA protein(s) in Arabidopsis. The specificity of AtGID1 genes is elucidated in this paper. Analyses of multiple loss-of-function mutants for Arabidopsis DELLA genes have elucidated the roles of specific DELLA protein(s) in specific organs: (i) RGL2 and RGA, and to a lesser extent RGL1, in flower development (Cheng et al., 2004; Tyler et al., 2004; Yu et al., 2004); (ii) RGA and GAI in stem elongation (Dill and Sun, 2001; King et al., 2001); (iii) predominantly RGL2, and to a lesser extent RGL1, in seed germination (Cao et al., 2005; Lee et al., 2002); (iv) RGA and GAI in root growth (Fu and Harberd, 2003). It is strongly expected that specific AtGID1(s) transduce a signal to specific DELLA protein(s) for the development of an organ in Arabidopsis.

Experimental procedures

Plant materials

The Ds insertion lines in the Nossen (Ns) background were obtained from the RIKEN BioResources Center (Tsukuba, Japan), where they are designated Ds13-1770-1 (for AtGID1a-1) and Ds11-3970-1 (for AtGID1b-1). Detailed information on the mutants is available online ( The T-DNA insertion lines in the Columbia (Col) background were obtained from the Arabidopsis Biological Resource Center (Columbus, OH, USA) as were the SALK T-DNA insertions designated SALK_044317 (for AtGID1a-2) and SALK_023529 (for AtGID1c-1). Detailed information on the mutants is available online ( To check the Ds or T-DNA insertions, we used a PCR-based approach with the specific primers: for AtGID1a, 5′-TACCTCGGGTTCGAAATCGAT-3′ (Ds region for atgid1a-1, forward), 5′-TGGTTCACGTAGTGGGCCATCG-3′ (T-DNA region for atgid1a-2, forward), 5′-ATTGGTGTTTTCTGGGTTTGATGCAG-3′ (1st intron, forward) and 5′-CATTCCGCGTTTACAAACGCC-3′ (2nd exon, reverse); for AtGID1b, 5′-TACCTCGGGTTCGAAATCGAT-3′ (Ds region, forward), 5′-GAAGAGGAAATCTAAGACGTCCCATC-3′ (1st intron, forward) and 5′-CCAGACTCTGGACTTGACCCAG-3′ (2nd exon, reverse); and for AtGID1c, 5′-TGGTTCACGTAGTGGGCCATCG-3′ (T-DNA region, forward), 5′-ATGGCTGGAAGTGAAGAAGAAGTTAATC-3′ (1st exon, forward) and 5′-TCAGAGAACTTTCAAACACAAAACC-3′ (1st intron, reverse). To determine the genotype of AtGID1a of each seed from the Aabbcc F1 progeny, we performed nested PCRs with primers for the region between the 1st intron and the 2nd exon, 5′-TCTGGGTTTGATGCAGACAG-3′ (forward) and 5′-CGTTTCAGAATCCAGCGTTT-3′ (reverse), and then 5′-CCTCTCAATACATGGGTTTTAATATCCAAC-3′ (forward) and 5′-CATTCCGCGTTTACAAACGCC-3′ (reverse); and for the region between the Ds intron and the 2nd exon, 5′-TACCTCGGGTTCGAAATCGAT-3′ (forward) and 5′-CGTTTCAGAATCCAGCGTTT-3′ (reverse), and then 5′-TACCTCGGGTTCGAAATCGAT-3′ (forward) and 5′-CATTCCGCGTTTACAAACGCC-3′ (reverse). All plants were cultivated in a temperature-controlled chamber at 23°C under continuous light or under a 16 h photoperiod after 2 days of cold treatment (4°C).

Scanning electron microscopy

Samples for scanning electron microscopy (JSM 5410LV, JEOL, were mounted on a conductive carbon tape (Ted Pella, Scanning conditions were: accelerating voltage, 15 kV; vacuum, 0 Pa; spot size, 10; filament current, 52 μA.


The procedure for semi-quantitative RT-PCR was performed almost exactly as described by Kim et al. (2005). The correct number of PCR cycles was determined for each gene by trial runs by using various different numbers of cycles at an interval of two cycles, to confirm that the amount of visualized product was within a dynamic range. We adopted the cycle number for each gene for total RNA from the wild-type plant (Columbia) without GA treatment as representative for all lines. All primers were previously checked using positive controls and their sequences are: for GA3ox1 (At1g15550), 5′-CATCCCATTCACCTCCCACACTCTCAC-3′ and 5′-CGATTCAACGGGACTAACCAGCTTC-3′; for At-EXP1 (At1g69530), 5′-GACGTCACATGTCAATGGTTACGC-3′ and 5′-GTTGTAACTTTGAATGAGAGAGATTGTCCG-3′; for AtGID1a (At3g05120), 5′-CAGATCAAGAGCAACCTCCTAG-3′ and 5′-CCACAGGCAATACATTCACCTGTGTG-3′; for AtGID1b (At3g63010), 5′-GAACCCTCGAGCTAACCAAACCTCTC-3′ and 5′-GGAGTAAGAAGCACAGGACTTGACTTGC-3′; for AtGID1c (At5g27320), 5′-CTGGCACTTCACCAAGTATTACTG-3′ and 5′-GCCAATAGTGGCTTGCTCCAAG-3′; for Act2 (At3g18780), 5′-CGTGTGTGACAATGGTACCGGTATGG-3′ and 5′-CTGTGAACGATTCCTGGACCTGCCTC-3′. All experiments were repeated at least twice independently, with very similar results.


We are grateful to Dr N. E. Olszewski, University of Minnesota (MN, USA), for technical advice. We thank Dr T.-P. Sun, Duke University (NC, USA), for providing ga1-3 seeds, and the Arabidopsis Biological Resource Center for providing T-DNA insertion mutants for AtGID1a and AtGID1c. We also thank Ms S. Kawamura, Ms F. Mori and Ms S. Kitabayashi for their assistance in the establishment and maintenance of single, double and triple loss-of-function mutants. This work was supported in part by grants to M.N. and M.K. from the Ministry of Education, Culture, Sports, Science and Technology of Japan. Seeds of Ds insertion mutants of AtGID1a (line Ds13-1770-1), AtGID1b (Ds11-3970-1) and the atgid1a atgid1b double knockout mutant (established by crossing Ds13-1770-1 and Ds11-3970-1) are available from RIKEN BioResource Center on request.