Integrative Modeling of Gene Expression and Metabolic Networks of Arabidopsis Embryos for Identification of Seed Oil Causal Genes

Abstract

Fatty acids in crop seeds are a major source for both vegetable oils and industrial applications. Genetic improvement of fatty acid composition and oil content is critical to meet the current and future demands of plant-based renewable seed oils. Addressing this challenge can be approached by network modeling to capture key contributors of seed metabolism and to identify underpinning genetic targets for engineering the traits associated with seed oil composition and content. Here, we present a dynamic model, using an Ordinary Differential Equations model and integrated time-course gene expression data, to describe metabolic networks during Arabidopsis thaliana seed development. Through in silico perturbation of genes, targets were predicted in seed oil traits. Validation and supporting evidence were obtained for several of these predictions using published reports in the scientific literature. Furthermore, we investigated two predicted targets using omics datasets for both gene expression and metabolites from the seed embryo, and demonstrated the applicability of this network-based model. This work highlights that integration of dynamic gene expression atlases generates informative models which can be explored to dissect metabolic pathways and lead to the identification of causal genes associated with seed oil traits.

Keywords: dynamic modeling; fatty acids; gene expression; metabolic networks; plant embryo.

FIGURE 1

Modeling framework and model calibration. (A) The modeling approach to map gene expression dynamics onto the metabolic network. The kinetic equations (left part of A) for modeling the gene expression system and metabolic system were developed based on data for mRNA profiles from microarray and metabolite content from global metabolic analysis for the seven embryo developmental stages. mRNA (red) modulates enzyme synthesis (E1, blue) which in turn catalyzes the production of a product metabolite (M2, green) from a substrate metabolite (M1, black). A sample simulation is shown with a transient mRNA profile (red stars) used as an input (right part of A). The symbol legend for the kinetic equations: k, enzyme reaction constant; k_syn, the rate constant for enzyme synthesis (hr^{− 1}); k_deg, the rate constant for enzyme degradation (hr^{− 1}). μ, the growth rate of the cell (hr^{− 1}). (B) Overview of the metabolic pathways important for Arabidopsis embryo physiological development; a complete set of differential equations and further details for this model are presented in Supplementary Figure 2. Substrates, metabolites and end products are in bold font, cofactors are in gray font, and enzymes and reactions are in white ellipses. Full names of enzymes and metabolites are provided in the Supplementary Tables 3–12. (C) Model calibration using wild-type Arabidopsis physiological data, including seed growth (top row of graphs), carbohydrate metabolism (middle rows of graphs), and FA and storage protein accumulation (bottom rows of graphs). In each row graphs presenting data on carbohydrate metabolism and FA and storage protein accumulation, the first and second columns of graphs depict mRNA data from the seven embryo developmental stages (squares) and simulated enzymes levels (dotted lines), which are used for simulation of content for starch (STA), sucrose (SUC), glucose (GLU), total fatty acids (total FA), storage proteins (SPrt), short-chain fatty acids (SCFA) and triacylglycerides (TAG). The simulation results (solid lines) are shown in the third column. Circles in the simulation graphs in the 3rd column of graphs are experimental data from Baud et al. (2002) to show the simulations are close to experimentally derived results. Key for other abbreviations used in (C): STA syn., genes and enzymes involved in starch synthesis; STA deg., genes and enzymes involved in starch degradation; SUC syn., genes and enzymes involved in sucrose synthesis; GLC trans., genes and enzymes involved in glucose transport; FA syn., genes and enzymes involved in fatty acid synthesis; SPrt genes, genes and enzymes involved in storage protein synthesis; FA elong., genes and enzymes involved in fatty acid elongation; and FA - TAGs, genes and enzymes involved in triacylglyceride formation.

FIGURE 2

Sensitivity analysis of fatty acid to gene modulation. Each reaction in the model was perturbed by changing the gene expression (horizontal axis) and the y-axis shows the % change in fatty acid after 20 days (blue dots). The red lines show the % change in the corresponding metabolic flux between 10 and 20 DAF (i.e., when FA are accumulated).