The gar2 and rga alleles increase the growth of gibberellin-deficient pollen tubes in Arabidopsis

Abstract

Ectopic expression in Arabidopsis of a pea (Pisum sativum) cDNA (2ox2) encoding a gibberellin (GA) 2-oxidase (PsGA2ox2), involved in the deactivation of biologically active GAs, has been used to establish a role for GAs in promoting pollen tube growth. One line, 35S:2ox2/28c, when homozygous for the transgene, exhibits a novel small fruit phenotype. The 28c transgene reduces pollen tube growth, and this results in a reduced number of fertilized seeds that are only present at the end of the silique nearest the stigma. To confirm that the 28c pollen tube phenotype is due to sense expression of the 2ox2 mRNA, a "hairpin" RNA interface silencing construct, designed to silence 2ox2 expression, has been used to restore pollen tube growth and fruit development. The interaction between 28c and other mutants with increased GA response has also been examined to provide further evidence that GAs play an important role in pollen tube growth. Based on the ability of mutant alleles to suppress the 35S:2ox2/28c phenotype, we define new roles for the gar2-1 and rga alleles in GA signaling during pollen tube elongation in addition to their previously established roles in vegetative tissues. In contrast to the constitutive GA response observed in internodes and leaves lacking RGA and GAI, the rga-2 gai-d5 mutant combination is only a partial suppressor of the 28c phenotype. Because the dominant dwarfing gai-1 allele reduces GA response in vegetative tissues, its effect on plant fertility has been examined. Although gai-1 reduces seed set, this appears to reflect defects in reproductive development other than pollen tube function. Finally, we show that the genetic background (Landsberg erecta or Columbia) modifies the 28c phenotype and that this effect is not due to the ER/er difference between these two ecotypes.

Figure 1.

Suppression of the small fruit phenotype exhibited by homozygous 35S:2ox2/28 plants. A, Treatment of 28c flowers with GA₃ restores fruit size. B, Schematic representation of the construct designed to silence the 35S:2ox2/28c transgene by posttranscriptional gene silencing/RNA interference (RNAi). C, Fruit from self-pollinated WT and 28c plants and from plants homozygous for 28c and the 35S:hp2ox2 construct. The inset shows 2ox2 mRNA levels, measured by semi-quantitative reverse transcriptase-PCR with β-tubulin as a control, in rosettes of WT, 28c+35S:hp2ox2/1n, 28c+35S:hp2ox2/1v, and 28c plants. D, Schematic representation of a self-pollinated hemizygous 35S:2ox2/28c flower at anthesis, indicating the growth of haploid pollen tubes through the transmitting tract to fertilize ovules containing the female gametophyte (egg). WT pollen tubes can fertilize ovules throughout the ovary, whereas 28c pollen tubes can generally only reach ovules near the stigma. E, Suppression of the 28c fruit phenotype by gar2-1 and rga-2, alleles known to increase GA response in other organs.

Figure 2.

Comparison of fruit development on self-pollinated WT (La-er) and 28c plants over the first 10 reproductive nodes on the main stem. n = 8 for each genotype. The inset shows a typical 28c fruit developing from a self-pollinated flower.

Figure 3.

Fruit development at node 3 for self-pollinated plants of various genotypes. n ≥ 8 for each genotype. The inset shows a typical 28c fruit developing from a self-pollinated flower.

Figure 4.

The rga-2 and gar2-1 alleles suppress the 28c phenotype by altering pollen function. Data represent fruit development on WT (La-er) ♀ plants pollinated with pollen of different genotypes. n ≥ 6 for each genotype.

Figure 5.

Pollen tube growth in vitro. Pollen grains were collected from flowers at anthesis, allowed to grow for 24 h, and pollen tubes were measured. n ≥ 25 pollen tubes for each genotype.

Figure 6.

Effect of the rga-2, gai-d5, and rga-2 gai-d5 mutant combinations on 28c fruit development. All plants were allowed to self-pollinate. Data is shown for reproductive node 8 because the flowers at the first several nodes of rga gai-d5 plants are often male sterile. n ≥ 8 for each genotype.

Figure 7.

The dominant dwarfing gai-1 allele reduces plant fertility. All plants were allowed to self-pollinate. Horizontal gray bars represent the mean ± 1 se for the indicated genotype at node 3. La-er plants did not contain unfertilized ovules at the silique base in this experiment. n ≥ 8 for each genotype.

Figure 8.

Aniline blue staining of pollen tubes in vivo. Pistils were stained 20 h after hand pollination.