Overview
A comprehensive, multiscale modeling approach is essential for accurately predicting the complex interplay between the metabolic output and the regulatory dynamics of our engineered system. Our device contains two plasmids, an L-cysteine overproduction plasmid and L-cysteine activated kill-switch. These plasmids require different modeling frameworks as one utilizes metabolic engineering and the other encodes a responsive genetic circuit. However, it is important to understand how these two inserts work together, since we ultimately expect to implement them in one plasmid in a dark fermentation bioreactor. Thus, a hybrid multiscale model that incorporates characteristics of both the L-cysteine overproduction and toxin-antitoxin system mechanisms is crucial for understanding the use of our engineered bacterial strain for industrial application and to improve design. The main models were built in Python and the code can be accessed via Github: GitHub Repository
We utilized flux balance analysis (FBA) to understand how the genetic modifications introduced for L-cysteine overproduction affects the overall metabolic network and relevant fluxes under steady state assumptions. Enzyme constraints were introduced to accurately reflect mutant enzymes introduced in the plasmid. This gave us an understanding of how our device changed L-cysteine production and how it could be further optimized by altering media conditions. Click below for more info:
Part of our novel composite part in the toxin-antitoxin system is the binding of L-cysteine to CcdR to then form a transcription factor involved in activating the toxin, CcdB. Given the novelty of the part, several key parameters like binding affinity for oligomerization and ligand binding are not well characterized. Thus, this model gave us the necessary parameters to properly characterize the composite part and to be used in the formulation of the mathematical model in the following section. Click below for more info:
To maximize of biohydrogen production, the device depends on the activation of the kill-switch to prevent toxic over-supplementation of L-cysteine in feedstocks. This process is highly time dependent and random, requiring ordinary differential equations (ODEs) and stochasticity to accurately model the relevant molecular interactions. This model informed the expected time for the activation of the kill switch. Click below for more info:
The engineered system is inherently time dependent, but the current models treat each plasmid as an isolated framework. Dynamic flux balance analysis (dFBA) addresses these issues by formulating a final model encompassing all parts of the design interweaving the FBA and ODEs to inform each other and increase accuracy of the model. The model tracks changes in extracellular metabolites, growth, and secretion of L-cysteine over time and refines the timing of the activation of the kill switch to understand how the device will behave when the plasmids are co-transformed and in a bioreactor. Click below for more info: