A Population Model Reveals Surprising Role of Stochastic Cell Division in Epigenetic Memory Systems

Epigenetic memory systems can store transient environmental signals in bacteria in form of DNA methylation patterns. A synthetic zinc finger protein (ZnF4) binds to the DNA in a methylation-dependent manner and represses the expression of the DNA methyltransferase CcrM. The ON-state of these systems is characterized by high CcrM expression, high methylation levels, and low ZnF4 binding, but the mechanisms ensuring long-term ON-state stability remain unclear. Measurements showed a gradual shift of cell populations from ON to OFF starting after about four days of cultivation. We use a hybrid modeling approach integrating flow-cytometry data and bulk methylation measurements to test the hypothesis that stochastic cell division is a key factor in this transition. Interestingly, model parameters cluster into two groups with opposite effects of cell division rates on ON-state stability. Experiments under varying growth conditions show that faster cell division increases memory stability -- an initially unexpected result. Model simulations provide a potential explanation for this observation and deepen our understanding about the mechanisms and timing of the ON/OFF switch in individual cells.