(333c) Loads and Modularity in Synthetic and Natural Switches: A Computational Study

We are on the cusp of significant breakthroughs in the use of synthetic gene circuits for a variety of applications. However the mathematical design of simple synthetic circuits is often made under assumptions of modularity, in particular the idea that these modules will behave identically in isolation and in connection. Similarly analysis of signal transduction or gene transcription networks is also carried out under the assumption of modularity, allowing analysis of motifs in isolation from the rest of the network. A limitation of this modularity assumption is the observation that downstream protein circuits affect the properties of an upstream circuit, without explicit feedback, because of binding between the interconnected parts of the two modules (Sontag et. al. Mol. Sys. Biol., 4:161). Thus understanding the effects of adding additional elements to the output of these technologically important network modules is required for a thorough understanding of the challenges of scaling up synthetic networks to higher levels of complexity. In this study we examine the qualitative behavior of the synthetic genetic toggle switch, as well as a toggle switch motif that we identify in the mammalian apoptosis network and other natural systems, and finally positive feedback switch motifs that have been shown to play a significant role in T cell activation. We show that the addition of a simple binding partner, i.e. a load, to the output of these circuits has the potential to change the qualitative behavior of the circuits. For the simple genetic toggle switch we find that loads affect the dynamics of switching between the two states. We construct the quasi-potential landscape of the toggle switch and show that loads skew the potential landscape dramatically. We show that some naturally occurring toggle switches have an additional motif that in principle allows the system to tune the quasi-potential landscape to deal with differing loads. We also show that for some naturally occurring signal transduction switches based on positive feedback, loads can abrogate switch-like behavior completely. Our analysis underscores the importance of analyzing the effect of connecting elements when studying network modules.