SummaryThree gibberellin (GA) receptor genes (AtGID1a, AtGID1b and AtGID1c), each an ortholog of the rice GA receptor gene (OsGID1), were cloned from Arabidopsis, and the characteristics of their recombinant proteins were examined. The GA-binding activities of the three recombinant proteins were confirmed by an in vitro assay. Biochemical analyses revealed similar ligand selectivity among the recombinants, and all recombinants showed higher affinity to GA 4 than to other GAs. AtGID1b was unique in its binding affinity to GA 4 and in its pH dependence when compared with the other two, by only showing binding in a narrow pH range (pH 6.4-7.5) with 10-fold higher affinity (apparent K d for GA 4 = 3 · 10 )8 M) than AtGID1a and AtGID1c. A two-hybrid yeast system only showed in vivo interaction in the presence of GA 4 between each AtGID1 and the Arabidopsis DELLA proteins (AtDELLAs), negative regulators of GA signaling. For this interaction with AtDELLAs, AtGID1b required only one-tenth of the amount of GA 4 that was necessary for interaction between the other AtGID1s and AtDELLAs, reflecting its lower K d value. AtDELLA boosted the GA-binding activity of AtGID1 in vitro, which suggests the formation of a complex between AtDELLA and AtGID1-GA that binds AtGID1 to GA more tightly. The expression of each AtGID1 clone in the rice gid1-1 mutant rescued the GA-insensitive dwarf phenotype. These results demonstrate that all three AtGID1s functioned as GA receptors in Arabidopsis.

SummaryArabidopsis 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.

To clarify the role of gibberellins in the seed development of Arabidopsis, we investigated the sites where gibberellins are synthesized and induce alpha-amylase genes. The spatial and temporal expression of the genes encoding gibberellin biosynthetic enzymes and alpha-amylases was examined by reverse transcription-PCR (RT-PCR) and in situ hybridization. The mRNAs of AtGA20ox2, AtGA20ox3 and AtGA3ox4 began to be detectable 5-7 d after pollination. In situ hybridization showed that these genes were expressed almost simultaneously around starch granules in the outer integument, preceding the disappearance of those granules. AtGA20ox2 and AtGA3ox4 but not AtGA20ox3 also showed their signals at the rim of the developing embryo. The alpha-amylase gene, Amy3, which responded to gibberellin, was mainly expressed in the developing seed, spatially overlapping with the expression of AtGA20ox2 and AtGA3ox4. These results suggest that gibberellins function in at least two sites of the seed: the outer integument and part of the embryo. We examined the phenotypes of a T-DNA insertion line of AtGA3ox4 and observed the following: (i) a decrease of alpha-amylase gene transcripts in young siliques; (ii) delay of starch degradation in the outer integument; (iii) disarrangement of the seed surface structure; and (iv) abnormal swelling pattern of polysaccharides after imbibition by the mature seed. These characteristics are phenotypes of plants under gibberellin starvation, because the abnormalities could be almost overcome with applied gibberellin, and the gibberellin-treated mutant was indistinguishable from the wild type. These results strongly suggest that gibberellins in the outer integument would be required for the normal formation of the Arabidopsis seed coat.

As the technology of flexible electronics has remarkably advanced, the long-term reliability of flexible devices has attracted much attention, as it is an important factor for such devices in reaching real commercial viability. To guarantee the bending fatigue lifetime, the exact evaluation of bending strain and the change in electrical resistance is required. In this study, we investigated the bending strains of Cu thin films on flexible polyimide substrates with different thicknesses using monolayer and bilayer bending models and monitored the electrical resistance of the metal electrode during a bending fatigue test. For a thin metal electrode, the bending strain and fatigue lifetime were similar regardless of substrate thickness, but for a thick metal film, the fatigue lifetime was changed by different bending strains in the metal electrode according to substrate thickness. To obtain the exact bending strain distribution, we conducted a finite-element simulation and compared the bending strains of thin and thick metal structures. For thick metal electrodes, the real bending strain obtained from a bilayer model or simulation showed values much different from those from a simple monolayer model. This study can provide useful guidelines for developing highly reliable flexible electronics.