Plant Biotechnology Journal (2003) 1, pp. 59–70 © 2003 Blackwell Publishing Ltd 59 Blackwell Science, Ltd Transcriptional control of nutrient partitioning during rice grain filling Tong Zhu*, Paul Budworth, Wenqiong Chen, Nicholas Provart 1 , Hur-Song Chang, Sonia Guimil, Wenpei Su, Bram Estes, Guangzhou Zou 2 and Xun Wang Torrey Mesa Research Institute, Syngenta Research and Technology, 3115 Merryfield Row, San Diego, CA 92121, USA Summary Cereal grains accumulate carbohydrates, storage proteins and fatty acids via different pathways during their development. Many genes that participate in nutrient partitioning during grain filling and that affect starch quality have been identified. To understand how the expression of these genes is coordinated during grain development, a genomic approach to surveying the participation and interactions of all the pathways is necessary. Using recently published rice genome information, we designed a rice GeneChip microarray that covers half the rice genome. By monitoring the expression of 21 000 genes in parallel, we identified genes involved in the grain filling process and found that the expression of genes involved in different pathways is coordinately controlled in a synchronized fashion during grain filling. Interestingly, a known promoter element in genes encoding seed storage proteins, AACA, is statistically over-represented among the 269 genes in different pathways with diverse functions that are significantly up-regulated during grain filling. By expression pattern matching, a group of transcription factors that have the potential to interact with this element was identified. We also found that most genes in the starch biosynthetic pathway show multiple distinct spatial and temporal expression patterns, suggesting that different isoforms of a given enzyme are expressed in different tissues and at different developmental stages. Our results reveal key regulatory machinery and provide an opportunity for modifying multiple pathways by manipulating key regulatory elements for improving grain quality and quantity. Received 6 September 2002; revised 23 September 2002; accepted 1 October 2002. * Correspondence (fax +1 858 812 1097; e-mail tong.zhu@syngenta.com) Present addresses: 1 Department of Botany, University of Toronto, 25 Willcocks Street, Toronto, Ontario, M5S 3B2, Canada; 2 ActivX Biosciences, 11025 N. Torrey Pines Road, La Jolla, CA 92037, USA . Keywords: gene expression, grain development, microarray, Oryza sativa , regulatory elements. Introduction The most important economic characteristics of agricultural grain crops are their yield, nutritional characteristics and culi- nary quality. Yield and nutritional value are mostly deter- mined by the synthesis and storage of carbohydrates, proteins and minerals during grain filling, and culinary quality is affected by the interaction of various enzymes to produce the final structure of the starch at the molecular and granule levels. The manipulation of these pathways can result in significant improvements in nutritional value (Mazur et al ., 1999; Ye et al ., 2000). For example, reducing the amount of even one enzyme in the starch biosynthetic pathway (granule-bound starch synthase) can dramatically affect the culinary quality (Singh et al ., 2000; Umemoto et al ., 1995), resulting in softer, less sticky cooked rice. Many genes that participate in nutrient partitioning during grain filling and that affect starch quality have been identified in rice and other cereals (e.g. Umemoto et al ., 1995). How- ever, these genes and their transcriptional controls are poorly understood, especially in cereal grains. Major biological processes such as cereal grain filling are believed to require a close coordination of gene expression among many important pathways. However, direct experi- mental evidence for this hypothesis has been lacking due to the absence of genomic sequence information and func- tional genomic technologies. To prove this hypothesis and to identify key regulators in grain development, a parallel examination of gene expression on a genome scale is necessary. Recent studies have examined gene expression