An engineered protein-phosphorylation toggle network with implications for endogenous network discovery

INTRODUCTION

Synthetic biology applies engineering methodologies to biology to build new functions into living cells. Through the assembly and characterization of engineered genetic modules, synthetic biologists have created cells with “prosthetic networks” implementing a variety of behaviors. To date, many of these efforts have focused on regulating gene expression at the transcriptional and translational levels. Because of the time required by cells to transcribe and translate regulatory proteins, synthetic networks based on these types of regulatory modalities operate on time scales of minutes to hours, which is too slow for various real-time sense-and-respond systems.

RATIONALE

By comparison, certain natural cellular processes such as signal transduction and metabolism use reversible protein-protein interaction networks to operate at much faster speeds. Among the various sophisticated behaviors that these networks encode is bistability, that is, exhibiting two distinct stable states and an ability to switch between them. Fast bistable switches are found in a range of important biological processes from cellular cycle progression to innate immune cell activation and memory formation. For protein synthetic biology, engineering bistable switches and, more generally, fast feedback control, represents a foundational tool with great utility. Although synthetic biologists have successfully implemented toggle switch networks using transcriptional and translational regulation, engineering analogous synthetic protein networks capable of switching state in seconds has not yet been achieved.

RESULTS

We built a bistable toggle switch in Saccharomyces cerevisiae with a regulatory network based solely on reversible protein-protein phosphorylation interactions. The network encodes positive feedback regulation through two branches arranged such that they repress each other (mutual cross-repression) and two distinct inputs to switch between the two possible states of the system. The engineered network is built from 11 phospho-in and phospho-out signal transduction elements comprising exogenous chimeric fusion proteins and endogenous proteins from the high-osmolarity mitogen-activated protein kinase (MAPK) pathway.

The resulting toggle network responds to extracellular inputs within seconds and exhibits long-term bistability across cell divisions. The toggle network also demonstrates ultrasensitivity, because cells encoding the toggle network respond to very-low-input signal concentrations that do not elicit observable responses in similar networks lacking feedback regulation. To highlight how protein networks may be advantageous in interfacing with certain natural processes in the cell, we rewired our toggle to control abrogation of cell budding through nuclear localization of a cytoskeletal protein.

Our engineered toggle network’s topology and size are distinct from existing synthetic toggles and well-studied endogenous bistable networks. This motivated us to develop a computational framework for searching endogenous systems for network connectivity similar to our engineered toggle that may also exhibit bistability. A search for toggle networks in endogenous pathways within S. cerevisiae from two to nine nodes in length yielded 109,401 toggle network candidates. Pruning this list to 186 candidates that could be readily experimentally tested, we discovered five previously unreported protein-protein interaction networks that were observed to exhibit bistability.

CONCLUSION

This work demonstrates an engineered fast, bistable toggle network composed solely of reversible protein-phosphorylation interactions and a framework for identifying bistable toggle networks embedded within natural settings. Future synthetic protein-protein networks similar to ours will enable biological engineers to create fast sensing and processing systems for regulating cellular processes that operate in real time. In turn, the design and implementation of these protein networks will enable the discovery of new embedded endogenous networks with prescribed behaviors.