Two Bistable Switches Govern M Phase Entry

Abstract

The abrupt and irreversible transition from interphase to M phase is essential to separate DNA replication from chromosome segregation. This transition requires the switch-like phosphorylation of hundreds of proteins by the cyclin-dependent kinase 1 (Cdk1):cyclin B (CycB) complex. Previous studies have ascribed these switch-like phosphorylations to the auto-activation of Cdk1:CycB through the removal of inhibitory phosphorylations on Cdk1-Tyr15 [1, 2]. The positive feedback in Cdk1 activation creates a bistable switch that makes mitotic commitment irreversible [2-4]. Here, we surprisingly find that Cdk1 auto-activation is dispensable for irreversible, switch-like mitotic entry due to a second mechanism, whereby Cdk1:CycB inhibits its counteracting phosphatase (PP2A:B55). We show that the PP2A:B55-inhibiting Greatwall (Gwl)-endosulfine (ENSA) pathway is both necessary and sufficient for switch-like phosphorylations of mitotic substrates. Using purified components of the Gwl-ENSA pathway in a reconstituted system, we found a sharp Cdk1 threshold for phosphorylation of a luminescent mitotic substrate. The Cdk1 threshold to induce mitotic phosphorylation is distinctly higher than the Cdk1 threshold required to maintain these phosphorylations-evidence for bistability. A combination of mathematical modeling and biochemical reconstitution show that the bistable behavior of the Gwl-ENSA pathway emerges from its mutual antagonism with PP2A:B55. Our results demonstrate that two interlinked bistable mechanisms provide a robust solution for irreversible and switch-like mitotic entry.

Keywords: Greatwall kinase; PP2A; cyclin-dependent kinase; hysteresis; mitosis; robustness; threshold.

Graphical abstract

Figure 1

CycB Threshold of Mitotic Phosphorylation in Xenopus Egg Extracts Persists in the Presence of an Inhibitor of Wee1 (A) Interphase egg extract was supplemented with luminescent probe, CycB-ΔN (144 nM) and cycloheximide (CHX) at 0 min. p27^Kip1 (450 nM) was added at 65 min for mitotic exit. Aliquots were analyzed by immunoblotting for phosphorylation (Pho’n) of CDK substrates (subst.), endogenous Fizzy-Ser50, and the probe. The probe was only weakly phosphorylated when CycB-ΔN was not added (Figure S1D). (B) Probe luminescence of (A) was plotted. (C) Interphase egg extract was supplemented with increasing amount of CycB-ΔN in the absence (left) or presence (right) of PD166285 (1 μM). After 80 min incubation, aliquots were analyzed by immunoblotting. (D) Probe luminescence of the samples shown in (C) was plotted. (E) Interphase egg extract immunodepleted of ENSA was analyzed as in (D). The inset is a magnification of the low x-axis range. Experiments were repeated three to five times, and similar results were obtained. The S50-1G4 probe was used in this figure. See also Figure S1.

Figure 2

The Gwl-ENSA Pathway Creates a Cdk Activity Threshold of Mitotic Phosphorylation in a Reconstituted System (A) Simple Cdk:Cyc/PP2A:B55 antagonism. Red and blue represent mitotic and anti-mitotic effects, respectively. (B) Time-course analysis of probe phosphorylation (Pho’n) shown as in (A), with increasing Cdk1:CycB levels ([PP2A:B55] = 50 nM). A sample without PP2A:B55 was used as 100% control (ctr.; open black squares). (C) Cdk:Cyc/PP2A:B55 antagonism with the Gwl-ENSA pathway. (D) Time-course analysis of probe phosphorylation shown as in (C) ([PP2A:B55] = 50 nM, [Gwl] = 20 nM, and [ENSA] = 300 nM). (E) Endpoint analyses of (B) (black) and (D) (red). (F) Aliquots of (B) and (D) were analyzed for Gwl (top), ENSA-S67 phosphorylation (middle; upper arrowhead indicates S67-phosphorylated form of ENSA), and ENSA-T28 phosphorylation (bottom). Slower and faster migrating forms of Gwl are indicated by square brackets. Samples free from PP2A:B55 are shown in lanes 9 and 18 as fully phosphorylated controls. Representative result of five experiments is shown here. (G and H) Simulation analyses of probe phosphorylation of (B) and (D). (I) The steady states of phosphorylated probe in (G) and (H), which can be directly compared with the endpoint analysis panel in (E). The T50-NCP probe was used in this figure. See also Figures S2 and S3 and Tables S1 and S2.

Figure 3

The Gwl-ENSA Pathway Defines a Bistable Switch with Two Distinct Thresholds (A) Reconstituted system was supplemented either with p27^Kip1 (450 nM) or staurosporine (Sts) (10 μM) 5 min before or 25 min after adding a supra-threshold level of Cdk activity (20 nM of Cdk2:CycA). Pho’n, phosphorylation. (B) Endpoint analysis of the samples shown in (A) was done for Gwl and ENSA-S67 phosphorylation by immunoblotting. (C) Simulation of Gwl/PP2A:B55 antagonism with staurosporine titration suggests bistability of the reconstitution system. (D) Experimental confirmation of bistability suggested in (C). Different concentrations of staurosporine were added to the reconstituted system before (black) or after (red) the addition of a supra-threshold level of Cdk activity (20 nM of Cdk2:CycA). Data after 60 min incubation were plotted. Representative data of two experiments are shown. The S50-1G12 probe was used in this figure. See also Figures S4A–S4F and Tables S1 and S2.

Figure 4

The Effect of Gwl Concentration on Cdk Thresholds and the Length of Time Delay before PP2A:B55 Reactivation (A) Steady states of the pS67 form of ENSA predicted from the model for different Gwl concentrations. cond., condition. (B) Model in (A) was experimentally tested in the reconstituted system. Samples of lower Gwl (0–20 nM) and higher Cdk (100 nM of Cdk2:CycA) concentrations were incubated for 16 min. Aliquots of samples were taken before (0 min, left) and 95 min after (right) the addition of p27^Kip1. Numbers in the ENSA blot (phos-tag, second from top) indicate concentrations of S67-phosphorylated ENSA in nanomolar (nM). PP2A:B55 is 50 nM. The effect of ENSA T28 phosphorylation is described in Figure S4 and the Supplemental Information. (C and D) Simulation (C) and experimental (D) analyses of the time course of probe phosphorylation (Pho’n) shown in (B) were done. Representative data of four experiments are shown. (E) Schematic diagram of two interlinked bistable switches. The S50-1G12 probe was used in this figure, including the model. See also Figures S4G and S4H and Tables S1 and S2.