A synthetic gene–metabolic oscillator

Abstract
Autonomous oscillations found in gene expression and metabolic, cardiac and neuronal systems1,2,3,4 have attracted significant attention both because of their obvious biological roles and their intriguing dynamics. In addition, de novo designed5,6,7,8,9,10,11,12 oscillators13,14 have been demonstrated, using components that are not part of the natural oscillators. Such oscillators are useful in testing the design principles and in exploring potential applications not limited by natural cellular behaviour15. To achieve transcriptional and metabolic integration characteristic of natural oscillators, here we designed and constructed a synthetic circuit in Escherichia coli K12, using glycolytic flux to generate oscillation through the signalling metabolite acetyl phosphate. If two metabolite pools are interconverted by two enzymes that are placed under the transcriptional control of acetyl phosphate, the system oscillates when the glycolytic rate exceeds a critical value. We used bifurcation analysis to identify the boundaries of oscillation, and verified these experimentally. This work demonstrates the possibility of using metabolic flux as a control factor in system-wide oscillation, as well as the predictability of a de novo gene–metabolic circuit designed using nonlinear dynamic analysis.