Ensemble Modeling Enables Quantitative Exploration of Bacterial Nitric Oxide Stress Networks

Bacterial nitric oxide (NO•) response networks are extensive and highly complex, owing to the broad reactivity of NO• and its reaction products. Deeper quantitative understanding of these NO• defense networks would greatly aid the development of antivirulence therapies aimed at sensitizing bacteria toward immune-generated NO•. In recent work, we have demonstrated that kinetic models can be used to quantitatively interpret and predict the dynamics of the NO• response network in Escherichia coli. Here, we postulate that such models can be used in an ensemble approach to identify the underlying mechanisms of how novel genetic mutations or chemical inhibitors perturb NO• defenses. Specifically, biochemical measurements from genetically or chemically perturbed bacteria exposed to NO• are obtained, and an ensemble of models is generated, where each individual model can quantitatively capture the system dynamics. The ensemble represents a collection of potential mechanisms, with each model differing in parameter values and/or network structure. To discriminate between these potential mechanisms, in silico simulations are performed in order to guide experiments to those that can discriminate between ensemble members. Experiments are then performed, and the results used to eliminate incompatible mechanisms. In this manner, an iterative computational and experimental methodology can be employed that ultimately converges on a single model that describes the mechanism underlying the effect of the genetic or chemical perturbation to the NO• response network. We propose that this integrated and iterative computational and experimental technique will be a powerful approach to quantitatively delineate the mechanisms underlying new leads for antivirulence therapeutics that target pathogen NO• response networks.