In the production of artemisinin precursors, the native Escherichia coli isoprenoid pathway (the deoxyxylulose 5-phosphate (DXP)pathway) was eschewed in favour of a heterologous pathway so as to circumvent the complex regulatory control imposed by the host(Box A2-4). In an alternative method of relieving regulatory control over the large number of DXP pathway components, Wang et al. (2009) diversified and, as a result, optimized the native DXP biosynthetic pathway in E. coli (see the figure, part a). The researchers developed a rapid, automated method for the in vivo directed evolution of pathways, which they termed multiplex automated genome engineering(MAGE). They then applied it to evolve the translational efficiencies of DXP pathway components to achieve maximal lycopene production. Specifically, cells were subjected to cycles of genetic modifications (through oligo-mediated allelic replacement) in an automated fashion to explore sufficient genomic diversity for optimizing biosynthetic pathways at laboratory timescales.
At the other end of the spectrum are strategies that rely on quantitative models and blueprints for the rational design of optimized networks and pathways (part b). Typically, a component of interest (for example, an engineered promoter (P*) or ribosomal binding site (RBS*)sequence) will be built into a simple test network. The network and its input–output data will then be fed into a model, which attempts to determine a parameter set that optimally describes the component’s dynamics within the framework of the model. Finally, the optimized parameter set will be used to forward-engineer new networks and components. For example, stochastic biochemical models have been developed to capture the expression dynamics of synthetically engineered promoters; these models were subsequently used to predict the correct in vivo behaviour of different and more complex gene networks built from the modelled components (Blake et al., 2003; Guido et al., 2006). Similarly, at the level of translation, thermodynamic models that predict the relative translation initiation rates of proteins can be used to rationally forward-engineer synthetic RBS sequences to give desired expression levels (Salis et al., 2009). Such technique sharness modelled genetic parameters (transcriptional or translational) to predict the level of expression of proteins and enzymes in a network. DMAPP, dimethylallyl diphosphate; FPP, farnesyl diphosphate; G3P, glyceraldehyde 3-phosphate; IPP, isopentenyl diphosphate.
From: A2, SYNTHETIC BIOLOGY: APPLICATIONS COME OF AGE
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.