Metabolic Organization in Plants: A Challenge for the Metabolic Engineer

Abstract

Abstract Predictive models of plant metabolism with sufficient power to identify suitable targets for metabolic engineering are desirable, but elusive. The problem is particularly acute in the pathways of primary carbon metabolism, and ultimately it stems from the complexity of the plant metabolic network and the plethora of interacting components that determine the observed fluxes. This complexity is manifested most obviously in the remarkable biosynthetic capacity of plant metabolism, and in the extensive subcellular compartmentation of steps and pathways. However it is argued that while these properties provide a considerable challenge at the level of identifying enzymes and metabolic interconversions ‐ indeed the definition of the plant metabolic network is still incomplete ‐ the real obstacle to predictive modelling lies in identifying the complete set of regulatory mechanisms that influence the function of the network. These mechanisms operate at two levels: one is the molecular crosstalk between effectors and enzymes; and the other is gene expression, where the relationship between fluctuations in expression and network performance is still poorly understood. The tools that are currently available for analysing network structure and performance are described, with particular emphasis on constraints‐based network analysis, metabolic flux analysis, kinetic modelling and metabolic control analysis. Based on a varying mix of theoretical analysis and empirical measurement, all four methods provide insights into the organisation of metabolic networks and the fluxes they support. Specifically they can be used to analyse the robustness of metabolic networks, to generate flux maps that reveal the relationship between genotype and metabolic phenotype, to predict metabolic fluxes in well characterised systems, and to analyse the relationship between substrates, enzymes and fluxes. No single method provides all the information necessary for predictive metabolic engineering, although in principle kinetic modelling should achieve that goal if sufficient information is available to parameterize the models completely. The level of sophistication that is required in predictive models of primary carbon metabolism is illustrated by analysing the conclusions that have emerged from extensive metabolic studies of transgenic plants with reduced levels of Calvin cycle enzymes. These studies highlight the intricate mechanisms that underpin the responsiveness and stability of carbon fixation. It is argued that while the phenotypes of the transgenic plants can be rationalised in terms of a qualitative understanding of the components of the system, it is not yet possible to predict the behaviour of the network quantitatively because of the complexity of the interactions involved.