Experimental evidnece for hysteresis in the cell cycles of Xenopus Laevis egg extracts

In 1993, Novak and Tyson published a comprehensive mathematical model of the regulation of M-phase promoting factor (MPF) activity in Xenopus laevis eggs and egg extracts. Although this model was in agreement with existing and subsequent experimental data, fundamental predictions that the cell cycle is driven by a hysteresis loop have never been validated experimentally. The model's predictions of bifurcations that create and destroy MPF activity, indicative of hysteresis, were tested in this study.

Prediction 1: The threshold concentration of cyclin B required to activate MPF is measurably higher than the threshold concentration required to inactivate MPF. The difference in thresholds implies that the MPF control system is hysteretic and bistable. To measure these thresholds, extracts in interphase or M-phase were supplemented with varying concentrations of non-degradable human cyclin B1 protein. MPF activity was determined by the morphology of sperm nuclei and by assays of histone H1 kinase activity. Consistent with the model, the activation threshold was determined to be 40 nM, which is two-fold higher than the inactivation threshold, 20 nM.

Prediction 2: For cyclin levels marginally above the activation threshold concentration of cyclin B, there is a dramatic "slowing-down" in the rate of MPF activation. Supra-threshold concentrations of nondegradable cyclin B1 were added to cycloheximide-treated CSF-released extracts, and samples taken at various time-points were analyzed for MPF activity. At 40 nM cyclin B1, just above the activation threshold, the lag time for MPF activation was 45 - 60 minutes; at 50 nM cyclin B1, the lag time was between 30 - 45 minutes; and at 60 nM or higher concentrations of cyclin B1, the lag time was 20 - 30 minutes, thus confirming the prediction of the Novak-Tyson model.

Prediction 3: DNA replication checkpoint increases the activation threshold concentration of cyclin B by increasing the hysteresis loop. Cycloheximide-treated, CSF-released extracts containing 1200 sperm nuclei/μl were treated with aphidicolin, then supplemented with varying concentrations of nondegradable cyclin B1. The activation threshold was 100 nM, 2.5 fold higher than in extracts lacking aphidicolin.

Conclusions: These studies confirm three predictions of the Novak-Tyson model and indicate that hysteresis underlies cell cycle control in Xenopus egg extracts. These experiments validate use of mathematical models to study complex biological control systems such as the eukayotic cell cycle.