A Dynamical Framework for the All-or-None G1/S Transition

Abstract

The transition from G1 into DNA replication (S phase) is an emergent behavior resulting from dynamic and complex interactions between cyclin-dependent kinases (Cdks), Cdk inhibitors (CKIs), and the anaphase-promoting complex/cyclosome (APC/C). Understanding the cellular decision to commit to S phase requires a quantitative description of these interactions. We apply quantitative imaging of single human cells to track the expression of G1/S regulators and use these data to parametrize a stochastic mathematical model of the G1/S transition. We show that a rapid, proteolytic, double-negative feedback loop between Cdk2:Cyclin and the Cdk inhibitor p27(Kip1) drives a switch-like entry into S phase. Furthermore, our model predicts that increasing Emi1 levels throughout S phase are critical in maintaining irreversibility of the G1/S transition, which we validate using Emi1 knockdown and live imaging of G1/S reporters. This work provides insight into the general design principles of the signaling networks governing the temporally abrupt transitions between cell-cycle phases.

Graphical abstract

Figure 1

Quantifying the Dynamics of G1/S Regulators in Single Cells (A) Stills taken from Movie S1 of HeLa cells stably expressing p27^Kip1-GFP and LSS2-mKate-PCNA. Graph shows quantification of p27^Kip1-GFP from individual cells, aligned to the G1/S transition. Four independent measurements were taken, and 17 cells from one experiment are shown. (B) Stills taken from Movie S2 of HeLa cells stably expressing CyclinE1-GFP and LSS2-mKate-PCNA. Graph shows quantification of CyclinE1-GFP from individual cells, aligned to the G1/S transition. Three independent measurements were taken, and 15 cells from one experiment are shown. (C) Stills taken from Movie S3 of HeLa cells stably expressing CyclinA2-GFP and LSS2-mKate-PCNA. Graph shows quantification of Cyclin A2-GFP from individual cells, aligned to the G1/S transition. Four independent measurements were taken, and 11 cells from one experiment are shown. (D) Stills taken from Movie S4 of HeLa cells stably expressing the CDK2 sensor and LSS2-mKate-PCNA. Graph shows quantification of CDK2 activity from individual cells, aligned to the G1/S transition. Three independent measurements were taken, and nine cells from one experiment are shown. In all cases, each curve is an individual cell. Time on images is shown in minutes and on graphs is shown in hours. Background shading on the graphs shows G1 in red and S phase in green. Scale bars, 10 μm.

Figure 2

Dynamics of the G1/S Transition in HeLa Cells (A) Influence diagram of G1/S regulatory network. Both CyclinE and CyclinA, as well as Emi1 (Hsu et al., 2002), are E2F target genes. The network is characterized by multiple, mutual inhibitory network motifs. (1) APC/C^Cdh1 promotes the degradation of CyclinA, but Cdk2:CyclinA inhibits APC/C activation by Cdh1 (Lukas et al., 1999, Zachariae et al., 1998). (2) p27^Kip1 inhibits both Cdk2 kinase activities, but Cdk2:Cyclin complexes promote p27^Kip1 degradation directly and indirectly, thereby creating a coherent feed-forward loop. The direct effect is through phosphorylation of p27 at T187 that targets p27^Kip1 to the SCF-proteasome system (Montagnoli et al., 1999, Sheaff et al., 1997). The indirect effect is through upregulation of the SCF component, Skp2 by inhibiting its degradation machinery, APC/C^Cdh1. (B) The time courses for each of the cell-cycle reporters and Cdk2 activity was averaged across all cells measured and aligned to S-phase entry (time 0). (C) Numerical simulation of the G1/S control network with the deterministic version of the model (see Supplemental Experimental Procedures for equations and parameter values). (D) The steady-state dependence of p27^Kip1 level on CyclinE and Emi1 (the diagram is computed in the absence of CyclinA since its level is low during the G1/S transition) was calculated as a one-parameter bifurcation diagram with the deterministic version of the model. p27^Kip1 can settle at high and low levels (solid curves) at low and high CyclinE levels, respectively. The two states are separated by two different thresholds whose values are Emi1 dependent. The G1/S transition is characterized by the drop from the high to low level of p27^Kip1.

Figure 3

G1/S Transition in CyclinE1,2-Depleted Cells (A) Stochastic simulations of G1/S transition with 5% residual CyclinE synthesis (with 20-fold increase in mRNA degradation rate). Relative p27^Kip1 levels and free Cdk2:Cyclin complexes are shown. (B) The CyclinA thresholds for p27^Kip1 inactivation and reactivation are plotted as a function of Emi1 levels with the deterministic model. The diagram is divided into three territories: low (green) and high (red) levels of p27^Kip1 and bistability (gray) where both of these states coexist. The trajectory of cell-cycle progression of CyclinE1/2-depleted cells is shown by a blue dashed curve. (C and D) Time courses of p27^Kip1-GFP level (C) and Cdk2 activity (D) in individual CyclinE1/2-depleted (color curves) and control siRNA-treated cells (gray curves). Both time courses are plotted from cell division, and S-phase entry is marked by a diamond on each curve. In (C), three independent experiments were conducted, and 8 (control siRNA) and 11 (CyclinE1/2 siRNA) cells are shown from one experiment. In (D), two independent experiments were conducted, and 10 (control siRNA) and 7 (CyclinE1/2 siRNA) cells are shown from one experiment.

Figure 4

G1/S Transition in CyclinA-Depleted Cells (A) Stochastic simulations of the G1/S transition with 5% residual CyclinA synthesis (with 20-fold increase in mRNA degradation rate). (B and C) Time courses of p27^Kip1-GFP level (B) and Cdk2 activity (C) in individual CyclinA2-depleted (color curves) and control siRNA-treated cells (gray curves). Both time courses are plotted from cell division, and S-phase entry is marked by a diamond on each curve. In (B), three independent experiments were conducted, and ten (control siRNA) and nine (CyclinA2 siRNA) cells are shown from one experiment. In (C), two independent experiments were conducted, and ten (control siRNA) and ten (CyclinA2 siRNA) cells are shown from one experiment.

Figure 5

Emi1 Is Required for an Irreversible G1/S Transition (A) Stochastic simulation of Emi1 depletion. The initial state of the model corresponds to an S-phase cell (CyclinA increasing, CyclinE decreasing, and p27^Kip1 level is low, see Figure 2C). At t = 5h, the level and rate of synthesis of Emi1 were reduced by 90%, causing APC/C^Cdh1 reactivation, loss of CyclinA, and accumulation of p27^Kip1 and CyclinE. Later, the re-accumulation of CyclinE promotes the re-degradation of p27^Kip1. (B) DNA re-replication after Emi1 depletion. FACS plots of control (left) and Emi1-depleted (right) cells at 0, 24, and 48 hr after release from thymidine. Numbers shown on each graph are the percentage of cells with DNA > 4n. (C) Experimental measurement of CyclinA2-GFP in Emi1-depleted cells. CyclinA2-GFP levels continue to increase at the beginning of the experiment, possibly since Emi1 is still functional. Later, CyclinA2-GFP levels decrease and remain low due to reactivation of APC/C^Cdh1 in the absence of Emi1. Two independent measurements were taken, and ten cells from one experiment are shown. (D) Experimental measurement of p27^Kip1-GFP in Emi1-depleted cells. p27^Kip1-GFP is initially low and then shows a transient increase in some cells. Two independent measurements were taken, and eight cells from one experiment are shown. (E) Experimental measurement of CyclinE1-GFP in Emi1-depleted cell. CyclinE1-GFP initially decreases and then continues to accumulate for the duration of the experiment. Two independent measurements were taken, and ten cells from one experiment are shown. (F) Quantification of Cdk2 activity in Emi1-depleted cells. Two cells show a significant dip in Cdk2 activity. Other cells show a more moderate decrease. In all cases, individual curves represent individual cells. Three independent measurements were taken, and 11 cells from one experiment are shown.

Figure 6

G1 Progression in Cells with an Intact Restriction Point (A) G1 control in cells with an intact restriction point including pRb control over E2F activity, as well E2F autoregulation. (B) The steady-state levels of CyclinE (blue) and pRb (red) in the extended model are plotted against each other. Low hypo-phosphorylated pRb corresponds to high E2F activity. These balance curves of CyclinE and hypo-phosphorylated pRb create two qualitatively different steady states (high CyclinE, low pRb labeled “High” and low CyclinE, high pRb labeled “Low”) after cell division. The existence of two different cellular states (fates) is largely dependent on the inverse N-shaped characteristic of the pRb-balance curve (red) caused by the antagonism between Cdk2:CyclinE and CKI. In the absence of CKI, the red curve becomes a hyperbole, and the bistable regime is much reduced. The “dividing the way” behavior is illustrated by two trajectories (dotted arrows). (C and D) Temporal evolution of free Cdk2:CyclinE (C) as well as E2F and hypo-phosphorylated pRb (D) in a cell approaching high and low Cdk2 activity states.