The hormone abscisic acid (ABA) plays a central role in suppressing precocious germination in developing maize seeds and in modulating the expression of maturation phase genes. Kernels that are blocked in ABA synthesis do not mature to dormant, desiccation-tolerant seeds, but instead germinate on the ear midway through kernel development. This precocious germination has been widely considered to be a default developmental program, but it is also possible that ABA is required to counteract a hormonal germination signal. Because gibberellins (GAs) and ABA act antagonistically in many aspects of plant development, we hypothesized that ABA antagonizes a positive GA signal that induces precocious germination, and perhaps also suppresses maturation phase gene expression. This model makes three testable predictions: 1) Active GAs should be present in pre-maturation phase embryos, 2) reduced GA levels should suppress precocious germination in ABA-deficient kernels, and 3) inhibition of GA synthesis may induce the expression of maturation phase mRNAs in the absence of exogenous ABA. In a series of experiments, we obtained data in support of each of these predictions.

Using gas chromatography-mass spectroscopy, we measured GA and ABA levels in developing wildtype maize kernels over the course of development. Seven different GAs were identified in developing seeds, two of which are known to have biological activity, GA1 and GA3. As shown in Figure 1, these GAs are present in pre-maturation stage embryos, reaching maximum levels during a developmental window just prior to the peak in ABA accumulation.

To gauge the developmental role of embryo GA, we conducted experiments to manipulate the relative GA and ABA levels over the course of kernel development. Seeds deficient in ABA were created by spraying developing wildtype ears with fluridone, or by using vp5 (viviparous) segregating ears. Reductions in GA levels were achieved through the use of the GA biosynthesis inhibitors paclobutrazol and ancymidol or by genetic blocks with either dwarf1 or dwarf5. We found that vivipary of ABA-deficient kernels was highly suppressed in the dwarf background and in ears that were treated with GA biosynthesis inhibitors prior to stage 2. The resulting seeds are both dormant and desiccation-tolerant.  In contrast, a GA deficit was found not to suppress vivipary in vp1 mutant kernels, which have normal ABA levels, but exhibit no seed-specific ABA responses.

Figure 1. Temporal accumulation of GAs in developing kernels.

When GA biosynthesis inhibitors were applied to cultured embryos, they mimicked the effects of ABA, by suppressing germination and inducing the accumulation of maturation-phase mRNAs. Figure 2 shows the accumulation of maturation mRNAs in pre-maturation embryos cultured for three days in media supplemented with paclobutrazol ± GA or ABA± GA. The ABA-inducible mRNAs in the northern blots are undetectable in pre-maturation phase embryos and are precociously expressed in culture upon treatment with exogenous ABA. As shown, paclobutrazol treatment also induced these mRNAs, while the addition of exogenous GA reduced their steady state levels. The ABA-inducible messages also require the Vp1 gene product, but Vp1 mRNA levels were not affected by these culture treatments (bottom panel).

From these results, we speculate that GA present in the early developing embryo stimulates a developmental program leading to vivipary in the absence of sufficient levels of ABA . When GA levels are reduced, an ABA/GA ratio is established that is appropriate for the suppression of germination and the induction of maturation-phase gene expression in ABA-deficient kernels. The fail ure to suppress vivipary via reduction of GA levels in vp1 kernels suggests that the Vp1 product functions downstream of the sites of GA and ABA action in programming seed development.

Figure 2. GA biosynthesis inhibition mimics ABA effects on cultured embryos.

Return to the MNL 70 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage