Present state of knowledge in the field and relation to work in progress elsewhere

       Isoform function.  Research in this field for the most part addresses the specific functions of the multiple isoforms of each starch metabolic enzyme.  Work has focused on starch metabolizing enzymes in storage organs, in plants such as maize, rice, pea, and potato. Studies employing biochemical purifications of native and recombinant enzymes have attempted to distinguish different enzymatic properties of each of the SS and BE isoforms (13, 15).  Defining the specificity of individual BEs, for example, involves consideration of the chain lengths transferred, as well as reaction rates with respect to cleavage distance from a reducing end, a non-reducing end, or a branch linkage.  Such functional enzymatic differences among maize BE isoforms recently were identified in our laboratory, following expression of each BE in transgenic yeast (51).  Similarly, details about the distinct enzymatic properties of SS isoforms are emerging through characterization of purified native or recombinant enzymes (7, 12, 52-54).  Considering, however, that the native primers are not characterized and that isoform specificities overlap to varying degrees, it remains difficult to define individual SS functions.  Recently, functional isoform differences were investigated regarding the regulation of SS activity.  Isoform-specific regulation of a particular maize SS, zSSIII/DU1, via 14-3-3 binding was suggested in a recent report that also indicated starch synthesis defects in Arabidopsis plants with reduced expression of 14-3-3 proteins (55).  Because little is known about the either in vivo substrates or regulatory factors for starch granule metabolizing enzymes, the current proposal will be useful in that every isoform from one species will be made available for biochemical study.

Genetic analysis indicates that isoforms in general do not overlap in function.  A clear example is the wx- mutation found in many species, which eliminates the conserved granule-bound SS isoform, GBSSI, and also prevents synthesis of amylose (19, 56).  Thus, none of the other conserved isoforms, i.e., SSI, SSII, nor SSIII, can substitute for this function of GBSSI.  Another example is provided by the du1- mutation of Zea mays, which results in loss of zSSIII/DU1 without affecting GBSSI or SSI, but still causes specific alterations in starch structure (57, 58).  Structural alterations also were observed in starch from pea embryos bearing a mutation in the rug5 gene coding for SSII (59), in a Chlamydomonas mutant lacking a specific SS (60), and in potatoes in which SSIII and/or SSII are targeted by antisense RNA (53).  Similarly, mutations that affect one isoform of BEII, i.e., ae- in maize and r in pea, also have significant affects on starch structure (21).  Specific functions extend even to isoforms that apparently are not present in abundance, as evidenced recently from analysis of transgenic plants in which expression of the maize enzyme zSSIIa is prevented (61).  The zSSIIa isoform codes for only a very minor percentage of the total soluble SS activity (7), yet these antisense plants were drastically affected in starch structure.  This research problem is one to which genome-scale studies in Arabidopsis are particularly applicable, because mutations in every isoform can be obtained and combined in congenic plant lines.

Pleiotropic genetic effects.  Current research also addresses the pleiotropic genetic effects on multiple enzymes in the network.  Several lines of evidence indicate that starch metabolism should not be considered as a set of reactions each operating as an independent kinetic process.  For example, mutations in the maize du1 gene, which codes for a specific SS (7, 20), also cause a decrease in the activity of BEIIa (8).  Genetic interactions in the form of a synthetic double mutant phenotype indicate a functional relationship between this SS and the isoamylase-type DBE coded for by the su1 gene (62).  Additionally, a recent report from the James/Myers laboratory demonstrated allele-specific effects of alterations in SU1 on two different BEs (63).  These observations and others indicating functional interactions between multiple proteins in starch metabolism are suggestive of protein-protein interactions in vivo.  The Arabidopsis genome information again provides a major advance by allowing comprehensive descriptions of the effects of eliminating any single enzyme in the pathway on every other enzyme in the group.  Also, comprehensive recombinant protein approaches allow the most thorough possible testing for direct interactions between proteins in the study set.

Transient starch metabolism.  Few research efforts have been directed toward understanding the synthesis or degradation of transient starch.  Studies of starch granules in tobacco and pea leaves show they contain both amylose and amylopectin, although relative to storage starches, the percentage of amylose is lower and the distribution of branch chains in amylopectin is different (26, 64).  This suggests that factors determining amylose production and amylopectin structure differ among tissues.

Altered carbon partitioning in Arabidopsis.  Mutations of Arabidopsis are known that eliminate or severely reduce starch production (65-68), as are mutations that cause starch excess and thus are likely to affect starch degradation (41, 69, 70).  Recently, the starch excess mutant sex1 was shown to be defective in the R1 protein, a putative pyruvate phosphate dikinase postulated to serve as a regulator of leaf starch mobilization (71).  Such mutants have been useful for research into the roles starch plays in gravity response (72, 73), carbon partitioning and photosynthetic capacity (74), and floral induction (5).  The relative proportions of oil and carbohydrate in seeds have been altered by at least two mutations in Arabidopsis, providing initial clues to connections between these two energy storage pathways.  The wri1 mutation seemingly alters the conversion of glucose into fatty acid precursors, resulting in a low accumulation of oil and high amounts of either starch or sucrose in the seed (75).  Another mutant, sse1, also results in accumulation of starch over oil in the seed (76).  The SSE1 cDNA predicts a polypeptide similar in sequence to a membrane-associated protein required for peroxisome assembly and protein trafficking.  From these data the hypothesis was proposed that starch accumulation is a default storage deposition pathway, which is repressed by normal oil body production.  The dynamic relationship between oil and starch production in seeds indicates a broad metabolic network, and points to the need to understand individual gene function for each isozyme in the pathway.

Relation to other work in the field.  The attractive features of applying Arabidopsis whole-genome studies to starch metabolism make it likely that other laboratories will propose similar comprehensive approaches.  To date, the only specific mutations in the gene network that have been published are in dbe1 (AtDBE-1) and dpe1 (AtDE-1) (36, 41).  Also, a mutation is known that affects an a‑amylase, but the specific mutated gene has not been identified (43).  An essential goal is a complete mutant collection in which each gene listed in Appendix A-1 is inactivated.  Towards that end we will seek actively to coordinate with other laboratories so that duplication is avoided.  The practical expectation is that several laboratories will work simultaneously towards this goal, and each will share resources so that a comprehensive tool set is available to all.  This resource will engender a wide variety of experiments that can be undertaken from various physiological, molecular, and biochemical approaches, certainly more than one laboratory could expect to address.

The need to coordinate research in several laboratories causes us to propose procedures for sharing data obtained from the current proposal.  A web page will be established that will comprehensively list information about each gene noted in Appendix A-1.  Sequence information will be formatted to allow easy reference to the size and amino acid composition of the protein, and this will shortly be expanded to show pairwise comparisons to all known isoforms from other species.  In this way the evolutionary relationships will be evident between each Arabidopsis protein and the conserved isoforms throughout the plant kingdom.  All literature references to any gene or gene product will be included on the page for the appropriate member of the network, along with links to published studies of conserved isoforms in other plant species.  As work proceeds on isolating mutants, progress will be posted on the page so that all labs will simultaneously be aware of ongoing progress.  Other laboratories will be invited to contribute information to the page for each gene.  Progress towards recombinant expression, isoform-specific antibody production, and results from biochemical characterization also will be posted on an ongoing basis.  We anticipate that this web page resource for the Arabidopsis starch metabolism gene network will be linked to all central Arabidopsis bioinformatics resources.  All experimental tools listed in specific aim #1, including Arabidopsis seed, antibodies, and bacterial expression strains, will be made freely available as soon as their nature is verified definitively.  Prior publication will not be a requirement for release of strains and reagents to other laboratories.