Transient Hysteresis in CDK4/6 Activity Underlies Passage of the Restriction Point in G1

Abstract

Cells escape the need for mitogens at a restriction point several hours before entering S phase. The restriction point has been proposed to result from CDK4/6 initiating partial Rb phosphorylation to trigger a bistable switch whereby cyclin E-CDK2 and Rb mutually reinforce each other to induce Rb hyperphosphorylation. Here, using single-cell analysis, we unexpectedly found that cyclin E/A-CDK activity can only maintain Rb hyperphosphorylation starting at the onset of S phase and that CDK4/6 activity, but not cyclin E/A-CDK activity, is required to hyperphosphorylate Rb throughout G1 phase. Mitogen removal in G1 results in a gradual loss of CDK4/6 activity with a high likelihood of cells sustaining Rb hyperphosphorylation until S phase, at which point cyclin E/A-CDK activity takes over. Thus, it is short-term memory, or transient hysteresis, in CDK4/6 activity following mitogen removal that sustains Rb hyperphosphorylation, demonstrating a probabilistic rather than an irreversible molecular mechanism underlying the restriction point. In Brief Chung et al. show that the commitment to mitogen-independent cell-cycle progression in G1, termed the restriction point (R), is not executed by an assumed feedback loop from cyclin E-CDK2 to Rb but rather by probabilistic short-term maintenance of CDK4/6 activity.

Stress-mediated exit to quiescence restricted by increasing persistence in CDK4/6 activation

“…We measured the response to a number of different types of stress signals and showed that most trigger first a rapid inactivation of CDK4/6 and that the loss in CDK4/6 activity is then followed by a reduction in CDK2 activity and exit to quiescence. This is also consistent with CDK2 activity and cell-cycle entry remaining sensitive to acute inhibition by the CDK4/6 inhibitor palbociclib until the end of G1 (Chung et al, 2019). We further show that CDK4/6 activity becoming increasingly more resistant to stress towards the end of G1.…”

“…Such a co-requirement for the two kinases has for example been proposed for Rb hyperphosphorylation (Zetterberg et al, 1995). Notably, a recent study by our group showed that CDK4/6 activity can still hyperphosphorylate Rb when all cyclin As and Es genes are missing in MEFs, suggesting that CDK4/6 activity alone can be sufficient for hyperphosphorylating the substrate Rb (Chung et al, 2019). We tested for a potential co-requirement of cyclin E/A-CDK2 activity in regulating the putative CDK4/6 reporter activity by also utilizing a Cre-Lox cell line to conditionally knockout all four cyclin E and A genes ( Figure 1 -supplementary figure 3A).…”

“…We therefore tested how CDK4/6, CDK2, and APC/C Cdh1 activities relate to the phosphorylation of Rb. As a marker of hyperphosphorylated Rb, we measured Rb 316 phosphorylation at Serine 807/811, which shows a bimodal distribution and is correlated with 317 activation of E2F transcription targets (Cappell et al, 2016;Chung et al, 2019;Yang et al, 318 2017). To test for a correlation between CDK4/6 activity and Rb phosphorylation, we used live-319 cell imaging followed by immunostaining and classified cells based on the time since mitosis.…”

“…While NaCl rapidly suppressed CDK4/6 and CDK2 activity at a similar rate, all other stresses and TGF-β1 signaling first inactivated CDK4/6 before CDK2 (Figure 4 -supplementary figure 2C). The typically more rapid stress-mediated inactivation of CDK4/6 is likely functionally relevant for the subsequent CDK2 inactivation as CDK4/6 activity is still needed to sustain the CDK2 activity increase throughout G1 phase ( Figure 1D) until Rb hyperphosphorylation becomes independent of CDK4/6 activity at the onset of S phase (Chung et al, 2019).…”

“…Such trimeric CDK4/6 complexes can be active (Sherr and Roberts, 1999), and tyrosine phosphorylation of p27 can generate active trimeric CDK4/6 complexes (Blain, 2008;Guiley et al, 2019), but studies using double p21/p27 (Cheng et al, 1999) and triple p21/27/p57 (Tateishi et al, 2012) knockout cells came to different conclusions whether binding of CIP/KIP type CDK inhibitors is required for cells to contain active cyclin D-CDK4/6. Addition of the cyclin D-CDK4/6 selective inhibitor palbociclib in late G1 also caused dephosphorylation of hyperphosphorylated Rb in less than 15 minutes (Chung et al, 2019), while an active cyclin D-CDK4 complex with bound tyrosine phosphorylated p27 was unresponsive to palbociclib inhibition (Guiley et al, 2019), raising additional questions how CDK4/6 activity is regulated in cells.…”