Selection of the gene set

Information about enzymes known to be involved in starch metabolism in other plant species can be used to identify all or most of the Arabidopsis genes likely to participate in this process. An outline of the general metabolic pathway of starch biosynthesis is shown in Figure 2 (3, 6) . To keep the size of the project practical, this proposal focuses on the polymer assembly and disassembly steps that take place after generation of the glucosyl unit donor ADP-Glucose.

SSs and BEs: Starch synthases (SS) catalyze addition of glucosyl units to growing chains through new a(1®4) linkages (Figure 3). Branching enzymes (BE) introduce a(1®6) linkages into the polymers (Figure 3). Theoretically, these two enzymes alone could be responsible for synthesis of starch, as is the case for glycogen biosynthesis. In this view, evolutionary change in the properties of an SS or a BE may have altered glycogen biosynthesis such that the structural requirements of starch granules were attained. Further evidence, however, suggests that other enzymes are also involved in starch assembly.

Fractionation experiments first characterized these enzymes and revealed multiple isozymes of SS and BE. Three distinct SS activities were fractionated from maize endosperm (7-12). Two SSs were identified in the soluble phase, and a third activity is associated with granules. Three BE isoforms also were detected in the soluble phase of maize endosperm (13-15). Molecular cloning subsequently identified genes coding for SSs and BEs, starting either with amino acid sequence information from purified proteins, or mutations that affect starch structure and coincidentally alter a certain enzyme. Sequence alignments of cloned genes and cDNAs define at least four distinct SS isoforms that are broadly conserved in plant species (7, 12, 16-20). Similarly, cloning of genes coding for BEs revealed two distinct isoforms, generally referred to as BEI and BEII, and that in at least some species including Arabidopsis two very closely related sub-isoforms of BEII exist (21-26).

Additional enzymes involved in starch biosynthesis: Genetic analysis has indicated the involvement of two enzymes in the determination of starch structure that might not have been envisioned solely by considering either the structure of the pathway’s glucan product or the enzymatic activity of the polypeptide. These are starch debranching enzymes (DBE) and disproportionating enzyme (DE). DBEs catalyze hydrolysis of a(1®6) linkages (Fig. 3C). Two DBE isoforms are highly conserved in plants (27-29). Mutations in a DBE are known in maize (30, 31), rice (32, 33), Chlamydomonas (34, 35), and Arabidopsis (36). In all instances the mutation causes reduction or elimination of amylopectin, concomitant with accumulation of an abnormal glucan that is soluble and highly branched, thus resembling glycogen. The specific roles of DBEs in biosynthesis are not yet known. One current hypothesis proposes a direct role in hydrolyzing certain a(1®6) linkages and thus contributing to the clustered arrangement of branches that is critical to the structure and properties of starch (3, 37). Another hypothesis proposes an indirect role for DBEs in elimination of water-soluble polysaccharides synthesized in a nonproductive pathway that competes with amylopectin biosynthesis on the granule (36).

A similar genetic analysis led to the suggestion that DEs have a function in determination of starch structure. DEs catalyze transfer of residues between linear chains through the cleavage and reformation of a(1®4) linkages (Fig. 3D). A mutation of Chlamydomonas was shown to condition loss of a DE activity and not to affect any other known enzyme involved in starch biosynthesis (38, 39). The accumulation of amylopectin is severely reduced in these mutants, and its structure is altered by a significant increase in the frequency of very short linear chains. Whether DEs are universally involved in starch production is not clear, however,, because antisense potato plants in which a DE is reduced do not display a defect in starch biosynthesis (40), nor is an Arabidopsis mutant lacking a DE affected in starch production (41). Nevertheless, the Chlamydomonas results indicate that DE must be considered as a possible component of the starch biosynthetic machinery.

Degradative enzymes: Granules generally are resistant to enzymatic attack, yet can be completely degraded when glucose supply is required. Starches can be mobilized by a combination of phosphorolysis and hydrolytic degradation (42). No single enzyme has been shown to completely convert starch to simple sugars, so multiple enzymes most likely are involved. DBEs and DEs are potential degradative enzymes (the latter releases glucose; Fig. 3D), as well as the a(1®4)-specific hydrolases of the a-amylase and b-amylase classes. a-amylases are endo- enzymes that cleave within a linear chain, whereas b‑amylases are exo-enzymes that work from the nonreducing end and stop their degradative action when encountering a branch linkage. Genetic evidence for involvement of an a-amylase in starch degradation comes from the sex4 mutant of Arabidopsis, which lacks such an enzyme and accumulates abnormally high levels of starch in leaves (43). b-amylase function in Arabidopsis appears to be particularly complex. One isoform is specifically expressed in phloem (44), and another is a chloroplast enzyme (45). At least three b-amylase isoforms are expressed in leaves, with a cytosolic form being the most abundant (42). b-amylases appear to be subject to complex developmental regulation, with expression induced by light, sugars, oligosaccharides, and abscissic acid (46-50). Finally, another enzyme that must be considered in starch degradation is phosphorylase, which inserts phosphoryl groups from inorganic pyrophosphate into the a(1®4) glucoside bond, releasing glucose‑1‑phosphate.

Arabidopsis genes likely to be involved in starch metabolism: The sequences of known enzymes from other plants were used to search for Arabidopsis genes likely to specify starch assembly or disassembly factors. Appendix A-1 shows the result, which is the gene set selected as the starting point for this study. Multiple genes for each enzymatic activity is the rule, with five for SSs, three for BEs, four for DBEs, two for DEs, 3 for a-amylase, 9 for b-amylases, and 2 for phosphorylases. A simple explanation for the multiple loci is gene duplication and functional redundancy. Analysis of SS, BE, and DBE genes in other species, however, shows that particular sequence families are highly conserved in evolution (3). Thus, specific functions, currently unknown, are likely to be provided by many of the duplicated genes. Mutations of individual members of a gene family in other species often result in alterations in starch structure (e.g., 20), which is not consistent with redundant functions. For a complete understanding of the system, each gene must be studied individually. This is now possible using the functional genomic approach in Arabidopsis.