Harnessing Natural Modularity of Metabolism with Goal Attainment Optimization to Design a Modular Chassis Cell for Production of Diverse Chemicals

Modular
design is key to achieve efficient and robust systems across
engineering disciplines. Modular design potentially offers advantages
to engineer microbial systems for biocatalysis, bioremediation, and
biosensing, overcoming the slow and costly design–build–test–learn
cycles in the conventional cell engineering approach. These systems
consist of a modular (chassis) cell compatible with exchangeable modules
that enable programmed functions such as overproduction of a desirable
chemical. We previously proposed a multiobjective optimization framework
coupled with metabolic flux models to design modular cells and solved
it using multiobjective evolutionary algorithms. However, such approach
might not achieve solution optimality and hence limits design options
for experimental implementation. In this study, we developed the goal
attainment formulation compatible with optimization algorithms that
guarantee solution optimality. We applied goal attainment to design
an Escherichia coli modular cell capable of synthesizing
all molecules in a biochemically diverse library at high yields and
rates with only a few genetic manipulations. To elucidate modular
organization of the designed cells, we developed a flux variance clustering
(FVC) method by identifying reactions with high flux variance and
clustering them to identify metabolic modules. Using FVC, we identified
reaction usage patterns for different modules in the modular cell,
revealing that its broad pathway compatibility is enabled by the natural
modularity and flexible flux capacity of endogenous core metabolism.
Overall, this study not only sheds light on modularity in metabolic
networks from their topology and metabolic functions but also presents
a useful synthetic biology toolbox to design modular cells with broad
applications.