In strain design for the purpose of metabolic engineering, enzymes to over or under express can be difficult to ascertain due to the complicated nature of biochemical networks. The development of network-scale kinetic models of these reaction networks is hindered by the overwhelming number of unknown kinetic parameters and functional forms of enzyme kinetics. Previous work based on stoichiometry can only deal with knockouts which alter the network stoichiometry. Since enzyme over or under expression does not alter stoichiometry, such methods do not apply. However, in order to achieve a desired phenotype for increased product formation, adjustment of enzyme expression levels is the key to shift flux toward the desired product.

We present a network-structure driven approach to identify enzymes to target for changes in expression for metabolic engineering strain design. This method allows for the detection of enzymes that require changes in expression levels to lift or impose kinetic constraints to transition between flux distribution states. These enzyme targets can be reached through knowledge of the network connectivity and related flux distributions. We have demonstrated the utility of this approach on the pathway to aromatics biosynthesis, characterizing the necessary enzyme expression level changes that have also been shown experimentally. Also, we show how this approach can be used to construct a kinetic model that can be further perturbed by the user to examine the behavior of the network. With such little input necessary, this is a versatile approach that can be utilized for many different pathways and organisms, given knowledge of the connectivity.