Developmental Regulation of Oil Deposition in Arabidopsis Seeds

The Arabidopsis transcriptional factors LEAFY COTYLEDON1 (LEC1), LEAFY COTYLEDON2 (LEC2), FUSCA3 (FUS3), ABSCISIC ACID INSENSITIVE3 (ABI3), and ABSCISIC ACID INSENSITIVES (ABI5) are known to regulate multiple aspects of seed development. In an attempt to understand the developmental control of storage product accumulation, we observed the expression time course of the five transcripts. The sequential expression of these factors during seed fill suggests differentiation of their normal responsibilities. By extending the expression periods of the two early genes LEC1 and LEC2 in transgenic seeds, we demonstrated that the subsequent timing of FUS3, ABI3, and ABI5 transcripts depends on LEC1 and LEC2. Because a delayed onset or reduced level of FUS3 mRNA coincided with reduction of seed oil content in the transgenic seeds, the role of FUS3 in oil deposition was further examined. Analysis of published seed transcriptomic data indicated that FUS3 transcript increased together with nearly all the plastidial fatty acid biosynthetic transcripts during development. The ability of FUS3 to rapidly induce fatty acid biosynthetic gene expression was confirmed using transgenic Arabidopsis seedlings expressing a dexamethasone (DEX)-inducible FUS3 and Arabidopsis mesophyll protoplasts transiently expressing the FUS3 gene. By accommodating the current evidence, we propose a hierarchical architecture of the transcriptional network in Arabidopsis seeds in which the oil biosynthetic pathway is integrated through the master transcriptional factor FUS3. In order to identify the targeted genes of those master transcription factors and to resconstruct the transcriptional gene regulatory networks of some fatty acid biosynthetic genes, the experimental data from a mesophyll experiment and the published seed transcriptomic data from AtGenExpress project were analyzed. Transcriptional gene regulatory networks centered at genes of interest were resconstructed through graphic Gaussian models and relevance networks, which were implemented in an R package called GeneNet. From the genes contained in the transcriptional networks centered at five master transcription factor genes and a representative fatty biosynthetic gene, over-represented promoter motifs and gene ontology terms were identified by running a Perl script and a web-based program called EasyGO. In addition, a new transcription factor gene At1g56170 was predicted to be involved in fatty acid biosynthetic process.