Fundamental differences in starch synthesis in the maize leaf, embryo, ovary and endosperm.

Enzymological and starch analyses of various ADP-glucose pyrophosphorylase (AGPase) null mutants point to fundamental differences in the pathways for starch synthesis in the maize leaf, embryo, ovary and endosperm. Leaf starch is synthesized via the AGPase encoded by the small and large subunits shown previously to be expressed at abundant levels in the leaf, whereas more than one AGPase isoform functions in the embryo and in the ovary. Embryo starch content is also dependent on genes functioning in the leaf and in the endosperm. AGPase encoded by shrunken-2 and brittle-2 synthesizes ~75% of endosperm starch. The gene, agpsemzm, previously shown to encode the small subunit expressed in the embryo, and agpllzm, the leaf large subunit gene, are here shown to encode the endosperm, plastid-localized AGPase. Loss of this enzyme does not reduce endosperm starch. Rather, the data suggest that AGPase-independent starch synthesis accounts for ~25% of endosperm starch. Three maize genes encode the small subunit of the AGPase. Data here show that the triple mutant lacking all three small subunits is lethal in early seed development but can be viable in both male and female gametes. Seed and plant viability is restored by any one of the three small subunit genes, including one previously thought to function only in the cytosol of the endosperm. Data herein also show the functionality of a fourth gene encoding the large subunit of this enzyme. Although adenosine diphosphate glucose pyrophosphorylase is shown here to be essential for maize viability, strong evidence for starch synthesis in the endosperm that is independent of this enzyme is also presented. Starch synthesis is distinct in the maize embryo, ovary, leaf and endosperm, and is coordinated among the various tissues.