Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Mar 26.
Published in final edited form as: Mol Cell. 2010 Mar 26;37(6):753–767. doi: 10.1016/j.molcel.2010.02.023

Quantitative Reconstitution of Mitotic CDK1 Activation in Somatic Cell Extracts

Richard W Deibler 1, Marc W Kirschner 1,*
PMCID: PMC2882237  NIHMSID: NIHMS188932  PMID: 20347419

Summary

The regulation of mitotic entry in somatic cells differs from embryonic cells, yet it is only for embryonic cells that we have a quantitative understanding of this process. To gain a similar insight into somatic cells, we developed a human cell extract system that recapitulates CDK1 activation and nuclear envelope breakdown in response to mitotic cyclins. As cyclin B concentrations increase, CDK1 activates in a three-stage nonlinear response, creating an ordering of substrate phosphorylations. This response is established by dual regulatory feedback loops involving WEE1/MYT1, which impose a cyclin B threshold, and CDC25, which allows CDK1 to escape the WEE1/MYT1 inhibition. This system also exhibits a complex response to cyclin A. Cyclin A promotes WEE1 phosphorylation to weaken the negative loop and primes mitotic entry through cyclin B. This observation explains the requirement of both cyclins A and B to initiate mitosis in somatic cells.

Introduction

During mitosis there is little in the cell that is unaltered: the nucleus disintegrates, the Golgi vesiculate, the chromosomes condense into non-functioning masses, and the cytoskeleton rearranges for a new purpose. As these changes are incompatible with the functioning of the interphase cell, the transition into mitosis should be sharp, complete, and as brief as possible.

The in vitro frog egg system has illuminated a core molecular circuitry to explain how the mitotic transition takes place (Kim and Ferrell, 2007; Pomerening et al., 2005; Pomerening et al., 2003; Solomon et al., 1990). Cyclin dependent kinase 1 (CDK1) is the master regulator of mitosis, but it exhibits no kinase activity on its own (Desai et al., 1992). It becomes activatable once bound to a cyclin, predominantly cyclin B in mitosis. CDK1 bound to cyclin B is phosphorylated on residue T161 by CDK activating kinase (CAK) to stabilize the cyclin B-CDK1 interaction and to induce the conformational rearrangements needed for kinase activity (Larochelle et al., 2007; Russo et al., 1996). However, the WEE1 and MYT1 kinases (WEE1/MYT1) rapidly inactivate the activatable CDK1 by phosphorylating residues T14 and Y15, thereby blocking ATP binding and hydrolysis. Consequently, at low levels of cyclin B, CDK1 is inactive (Solomon et al., 1990). Once cyclin B concentrations exceed a threshold, CDK1 activates after a 10–20 min lag (Solomon et al., 1990). This activation is abrupt and occurs through positive and double negative feedback loops. Cyclin B-CDK1 phosphorylates and activates the CDC25 phosphatase allowing CDC25 to remove the inhibitory T14 and Y15 phosphorylations on CDK1. Cyclin B-CDK1 is also a negative regulator of both MYT1 and WEE1 as these two kinases are inactivated upon cyclin B-CDK1 phosphorylation (Okamoto and Sagata, 2007; Watanabe et al., 1995). However, one weakness of this description is that it is unclear how the activity of inactivated CDK1 increases in the lag phase to initiate the positive and negative feedback loops, i.e. what the exact trigger is for mitotic entry.

Despite the conservation of the mitotic circuitry, it is clear that cell division in the frog egg differs significantly from that in proliferating somatic cells. In the egg, which lacks a recognizable G2 phase, the cytoplasmic state is the sole determinant of the cell cycle stage. Without any nuclear control, the frog egg lacks many critically important mitotic features seen in somatic cells, such as sensitivity to checkpoint-inducing stresses. In addition, other cyclins (usually cyclin A but also cyclin E in certain cells, such as mouse fibroblasts (Kalaszczynska et al., 2009)) are essential in somatic cell division but dispensable in the early frog embryo. Further clouding the issue is that many of the core components, such as WEE1 and CDC25, have diverged in their regulatory sequences between frog embryos and mammalian somatic cells (Kim and Ferrell, 2007; Kim et al., 2005; Okamoto et al., 2002; Okamoto and Sagata, 2007).

These findings argue that the mechanism that drives the G2/M transition in somatic cells is related to, yet distinct from, that used in frog eggs. While we have detailed and quantitative information of mitotic entry in eggs, this same basic information is lacking for somatic cells. To address this deficit, we have reconstituted a biochemically tractable cell-free system from human somatic cells at the G2 stage of the cell cycle that recapitulates mitotic entry and preserves the network of interactions that leads to CDK1 activation in response to physiological levels of cyclins A and B. We first explore the features of the response to cyclin B to understand how the cell buffers itself from entering mitosis as cyclin B levels slowly rise in S and G2 phase. Next, we examine the role of the dual mitotic feedback loops in this process, and ask whether the system has only two stable states (interphase and mitosis) or multiple stable states. Finally, we describe the critical role of cyclin A in mitosis and show how it is able to feed into the cyclin B circuitry. Fom these considerations we can draw a more realistic and quantitative picture of mitotic entry in somatic cells in vivo and understand why the two cyclins are localized into different cellular compartments during interphase..

Results

Establishing a Somatic Cell Extract System for Mitotic Entry

To create a frame of reference for reconstituting an in vitro mitotic entry system (Figure 1A) using cell extracts and purified cyclins (Figure 1B) we examined several in vivo parameters as cells progressed through the cell cycle. HeLa S3 cells were synchronized at the G1/S transition with a double thymidine block. These cells, which synthesize less mitotic cyclins during this arrest than other cell types (Sherwood et al., 1994), entered mitosis nine to ten hours after G1/S release as determined by two independent markers, the phosphorylation of CDC27 and the phosphorylation of histone H3 (Figure 1C). During this time cyclin A levels rose three-fold and cyclin B increased 10–20 fold (Figure 1C). We used quantitative immunoblotting with recombinant cyclins A and B as standards to determine their concentration in mitotic cells. In mitosis cyclin B was estimated to be at 975 ± 230 nM (at 2.5 pL cell volume by Coulter counter) and cyclin A to be 277 ± 49 nM. This high cyclin B measurement is in accord with other estimates of mitotic cyclin B levels (Xu and Chang, 2007) and is likely to be a lower limit to the amount of cyclin B owing to the slight asynchrony with which cells pass through mitosis.

Figure 1.

Figure 1

Open in a new tab

Overview of cyclin induced mitotic entry

(A) Schematic depicting experimental approach. See text for details.

(B) Coomassie blue-stained polyacrylamide gels containing purified recombinant His6-cyclin B and His6-cyclin A.

(C) HeLa S3 lysates (20 µg) prepared at the indicated times following G1/S synchronization (Time = 0) were separated by SDS-PAGE and were immunoblotted with the shown antibodies (top). Recombinant cyclin A and B were added to quantify their concentration in the lysate. Shown is the quantification of three independent experiments with the error bars signifying ± one standard deviation (bottom).

(D) Graphical representation of the DNA content of asynchronous cells and cells 5.5 hours after release from a double thymidine block as determined by FACS analysis.

(E) Checkpoint status in somatic cell extracts. Cells were treated with 25 μM etoposide for 4 hours, lysed, and compared to the extracts used in the described in vitro system made with (NH4)2SO4 by immunoblotting.

These observations suggested the optimal time to make a cell extract would be approximately six hours after the start of S phase when most cells have terminated DNA replication but before cyclin B levels rapidly rise. Flow cytometry confirmed that ~70% of cells at this time possess a 4n or late S phase DNA content (Figure 1D). The lack of mitotic markers indicated that the 4n cells at this time were in G2 rather than M phase. It was possible that these cells could have active DNA replication or damage checkpoints. However, when we assessed the activation/phosphorylation of checkpoint kinases ATM, CHK1, and CHK2, we observed little signal for any of these proteins compared to when the DNA had been damaged by the addition of the DNA damaging agent etoposide (Figure 1E). Thus, this period of the cell cycle was suitable for in vitro analysis of the effect of cyclin B accumulation on the G2/M transition.

A consideration when making cell extracts is that cell cycle components often preferentially localize to certain compartments of the cell. However, these localizations are also very dynamic with component exchange from one compartment to another. For example cyclin B is predominantly cytosolic but can enter and leave the nucleus. CDK1 is activated in the cytosol but is phosphorylated by the predominantly nuclear CAK and WEE1 kinases. For this reason, and for experimental simplicity, we developed total cell extract conditions that mimic the mitotic state in which the nuclear membrane is dissolved and nuclear and cytoplasmic contents are thoroughly mixed. The HeLa nucleus contributes ~15% of the cell volume and would not be expected to significantly dilute the cytosolic contents. Indeed, the levels of the core cell cycle components in the extract were in close agreement with their relative abundances in cells, indicating that this complex lysate reflects the in vivo composition (Figure S1).

A Three-Stage Activation Profile of CDK1 by Cyclin B

To mimic the slow rise in cellular cyclin B levels thought to drive mitotic entry we added increasing concentrations of purified recombinant human cyclin B1 tagged with an N-terminal His6 epitope. The level of endogenous cyclin B in the extract was very low (40 ± 11 nM) compared to the amount added (see below). As this system is not supplemented with cyclin A (and the diluted endogenous cyclin A produces little relative CDK activity) it approximates cytosolic CDK1 activation in the cell. CDK1 activity was assayed by measuring its ability to phosphorylate histone H1 20 min after addition of cyclin B.

Cyclin B, alone, produced high levels (12-fold over background) of CDK1 activity (Figure 2A), suggesting that no other proteins need to be synthesized during interphase for cyclin B to activate CDK1. This observation is in accord with in vivo experiments with intact cells that showed the activation of CDK1 through the overexpression of only cyclin B1 and CDC25C is sufficient to promote premature mitosis in G2 cells (Karlsson et al., 1999). We refer to this H1 kinase activity as CDK1 activity as 28 μM of the CDK1 specific inhibitor RO-3306 reduced it by ~85% (Figure S2A). Interestingly, the dependence of CDK1 activation on the concentration of cyclin B was neither linear nor simply switchlike (ultrasensitive), like the classic picture observed in frog egg extracts. Instead, CDK1 activation occurred, in a highly reproducible manner, in three distinct stages. At low cyclin B concentrations, up to ~100 nM, the CDK1 activity of the extract increased proportional to the amount of recombinant cyclin B added (hereafter referred to as stage 1). This activation (see also Figure S2B) was statistically significant by Welch’s t-test (p<0.05). Next, from 125 to 400 nM cyclin B there was a plateau of CDK1 activity (stage 2): despite a 3.2-fold increase in cyclin B levels, CDK1 activity remained ~3.5-fold over the background activity of the extract. At >400 nM cyclin B the CDK1 activity of the extract again increased until saturation (stage 3). This change in CDK1 activity during stage 3 was abrupt, increasing 3.3-fold as the cyclin B concentration increased from 400 to 600 nM, and was statistically significant (p=0.00005). When fit as a Hill plot the stage 2/3 transition had a Hill coefficient greater than 7; using the least-squares (χ2) test we found that the nonlinear fit was three orders of magnitude better than the linear fit. The sharp activation of CDK1, in which the increase in CDK1 activity greatly exceeded the increase in cyclin B, suggested an ultrasensitive rather than a linear activation. However, the stage 2 plateau of partially active CDK1 (three-fold over background) is quite different from the simple switchlike activation of CDK1 in the Xenopus system. Although all extract preparations showed this behavior, there was some variation (± 2-fold) at the concentration at which the transitions occurred (Figure S2C). For the purpose of comparison, all data shown is from extracts with a stage 1 to 2 transition at ~125 nM cyclin B and stage 2 to 3 transition at 400–500 nM cyclin B. That ≥600 nM cyclin B was needed for the transition to fully active CDK1 is consistent with our observation that mitotic cells are capable of normally producing ~1 μM cyclin B. Here and in all following experiments we could use full-length cyclin B (rather than non-degradable mutants), because the anaphase-promoting complex (APC) was not activated in these extracts, despite high CDK1 activity. (Figure S2D).

Figure 2.

Figure 2

Open in a new tab

Biochemical characterization of the response of somatic cell extracts to His6-cyclin B

(A) Autoradiogram of a dried gel containing radioactively phosphorylated histone H1 (top). His6-cyclin B was added to somatic cell extracts at the indicated concentrations and incubated for 20 min at room temperature. Kinase assays were performed to determine the amount of P32 incorporated into the CDK1 substrate histone H1. Quantification (n=2–9) is shown (bottom) with error bars representing ± one standard deviation. The coefficient of variation was constant at 30%.

(B) The effect of cyclin B on cellular proteins was analyzed by immunoblotting (phospho-CDK1, CDC25 and actin as a loading control) or by adding radiolabeled in vitro translated proteins to the extract (WEE1 and securin).

(C) Gel filtration of extracts following His6-cyclin B addition. Following a 20 min incubation at room temperature with the indicated cyclin B concentrations, extracts were loaded onto a Superdex 200 column, and fractions were analyzed by immunoblotting for CDK1, cyclin B, and WEE1. Shown are apparent molecular weight positions.

(D) Cyclin B-CDK1 binding to mitotic regulators. His6-cyclin B bound to Ni-NTA agarose was added to the extract at a concentration below (100 nM) or above (2000 nM) the level for robust CDK1 activity. Following a 30 min incubation, the extract was centrifuged and the Ni-NTA beads were collected. The beads were washed three times with extract dialysis buffer and the bound proteins were analyzed by SDS-PAGE and immunoblotting.

CDK1 is Regulated by Y15 Phosphorylation and Exhibits an Ordering of Substrate Phosphorylations

To understand the mechanism underlying these intricate CDK1 dynamics we focused on the phosphorylation of the Y15 residue of CDK1, a critical regulatory element of the known dual mitotic feedback loops. As in the frog egg extract, addition of cyclin B increased the phosphorylation of Y15 (Figure 2B), confirming the presence of an operative WEE1/MYT1 negative feedback loop in the somatic cell extracts. The levels of CDK1 Y15 phosphorylation increased until ~400 nM cyclin B (Figure 2B and see Figure 3), the point at which there is a switch from stage 2 to stage 3 of CDK1 activation. At higher cyclin B concentrations the amount of Y15 phosphorylation progressively decreased until it reached background levels (quantification of a representative experiment is shown in Figure S3A). This result suggests that the transition from stage 2 to stage 3 corresponds to the activation of the CDC25 feedback loop and the inactivation of the WEE1/MYT1 feedback loop (see below). Unlike the frog system that shows a sharp loss of signal, as the cyclin B concentration was increased beyond 400 nM there was a proportional linear decrease in CDK1 Y15 phosphorylation. Additionally in contrast to the frog egg (Devault et al., 1992; Solomon et al., 1990) there was substantial CDK1 activity concurrent with high levels of CDK1 Y15 phosphorylation (stage 2) in the somatic system. Though these results differ from the frog, they are consistent with somatic cell observations that show that the dephosphorylation of Y15 is a gradual process with active and inactive cyclin B-CDK1 complexes coexisting in the same cell (Lindqvist et al., 2007). These findings further indicate a highly nonlinear relationship between the amount of CDK1 Y15 dephosphorylation and CDK1 activation.

Figure 3.

Figure 3

Open in a new tab

WEE1 and CDC25 determine the nature of the response of the extract to cyclin B

(A) Effect of WEE1 and CDC25 inhibition on CDK1 activity. Shown is an autoradiogram of dried gel containing radioactively phosphorylated histone H1 (top) and quantification (bottom). Activities corresponding to stages 2 and 3 are also shown.

(B) Effect of WEE1 and CDC25 inhibition on CDK1 Y15 phosphorylation.

(C) Effect of WEE1 and CDC25 inhibition on CDK1 substrate phosphorylation. Shown is an immunoblot CDC25C in the presence and absence of W2.

(D) Effect of CDK1 inhibition on WEE1 phosphorylation (top) and CDK1 Y15 phosphorylation (bottom).

To examine CDK1 activation further, we monitored phosphorylation of CDK1 substrates in the extract by electrophoretic mobility shifts (Figure 2B). In the Xenopus system there is a cyclin B threshold: at cyclin B concentrations below a certain level no phosphorylation occurs, whereas above the threshold all relevant CDK1 substrates are phosphorylated eventually (Georgi et al., 2002). The somatic cell extract system, however, exhibited very different and graded properties. With only 27 nM added cyclin B, a detectable portion of WEE1 protein was shifted by gel electrophoresis; >50% of WEE1 shifted at 55 nM cyclin B. No other CDK1 substrates examined shifted at these low stage 1 cyclin B concentrations. Securin was phosphorylated next with >50% of the protein showing an electrophoretic shift by the addition of 161 nM cyclin B near the plateau of CDK1 activity in stage 2. Thus, both WEE1 and securin showed extensive modification well below the sharp upswing of histone H1 kinase activity occurring at cyclin B concentrations >400 nM. It was not until the sharp transition at 400 nM that the CDK1 Y15 phosphatase, CDC25C, manifested its mobility shift. Importantly, the concentration at which CDC25C shifted also corresponds to the concentration of cyclin B at which the CDK1 Y15 phosphorylation signal began to decrease. CDC27, a component of the APC, required even higher concentrations of cyclin B (600 nM) before its electrophoretic mobility was reduced.

The in vitro phosphorylation results corresponded well to those observed in intact cells (see Figure 1C). Our in vivo measurements of synchronized cells show that WEE1 is partially phosphorylated six hours after the start of S phase with the bulk phosphorylated by eight hours. CDC25C showed little phosphorylation at seven hours, but reached its highest levels of phosphorylation by eight hours. Between seven and eight hours the in vivo concentration of cyclin B increased from 200 to 700 nM, which brackets the amount of cyclin B needed for a similar phosphorylation in vitro. In addition, as observed in vitro, this cyclin B concentration drives dephosphorylation of CDK1 in vivo. Phosphorylation of the CDC27 occurred in vivo at a time when the cyclin B level had reached 1 μM at nine hours, which is less than two-fold higher than the cyclin B concentration needed for in vitro phosphorylation.

Collectively, these results indicate that there is an ordering of specific phosphorylation sites in CDK1 substrates with certain sites being stably phosphorylated at lower cyclin B levels (and, hence, lower CDK1 activity) than others. A consequence of this ordering could be that a CDK1 regulator, such as CDC25, cannot become phosphorylated until CDK1 has achieved a sufficient activity and has already phosphorylated more favorable sites on other substrates. The idea that competition among CDK1 phosphorylation sites influences the ultrasensitive nature of CDK1 activation has been suggested for the Xenopus system (Kim and Ferrell, 2007).

Cyclin B Forms a Complex with WEE1 that is Disrupted at High CDK1 Activity

To explore the mechanism by which cyclin B activates CDK1 we examined the binding of cyclin B to CDK1 using size exclusion chromatography and binding assays. By gel filtration, in the absence of exogenously added cyclin B, the bulk of CDK1 was monomeric, eluting at 35 kD, except for a small portion co-eluting with and bound to cyclins (Figure 2C). Increasing cyclin B reduced the level of free CDK1: when 1250 nM cyclin B was added to fully activate CDK1, all CDK1 in the extract was in a complex. At lower concentrations (125, 250, 375 nM) of cyclin B, not all CDK1 was bound. Interestingly, the size of the cyclin B-CDK1 complex, 176 kD, (calculated from a diffusion coefficient of 4.1 × 10−7 cm2/s by gel filtration and a sedimentation coefficient of 7.7S by sucrose gradient sedimentation) was larger than would be expected based on the predicted sizes of the two proteins, suggesting that other proteins might be complexed with CDK1-cyclin B. This size is nearly the same (191 kD) as would be expected for a complex comprising cyclin B (49 kD), CDK1 (34 kD) and CKS1 (9 kD), which is normally bound to CDK1, and WEE1 (98 kD). Accordingly, we found that WEE1 co-eluted with cyclin B-CDK1 (Figure 2C). To determine whether WEE1 was actually in a complex with cyclin B-CDK1 we performed co-precipitation experiments. Cyclin B bound to Ni-NTA agarose was added to the extract at two different concentrations (100 nM and 2000 nM) and re-isolated (Figure 2D). As suggested by the gel filtration data at low levels (100 nM) of cyclin B, WEE1 was bound to cyclin B. However, when the concentration of cyclin B was increased 20-fold to completely activate CDK1, less WEE1 was pulled down by the cyclin B, suggesting the interaction with WEE1 was weakened. To test whether this dissociation was promoted by CDK1 activity, we added 28 μM RO-3306 (shown above to inhibit CDK1) and found that blocking CDK1 activity increased the efficiency of cyclin B in pulling down WEE1 (Figure S3B). These results suggest that during stages 1 and 2 of the activation, cyclin B shows saturation binding to CDK1. The bound CDK1 undergoes inhibitory phosphorylations, but by gel shift assays, only part of the population is phosphorylated. Therefore, the stage 2 plateau in activity is likely to be the result of a mixture of inhibited and uninhibited cyclin B-CDK1 kinases. During this stage when cyclin B-CDK1 is partially inhibited, WEE1 is bound to the complex. At higher cyclin B levels, WEE1 dissociates, which could result in a reduced efficiency in inhibiting CDK1 and increased CDK1 activation.

The WEE1-cyclin B-CDK1 interaction has not been previously reported and has not been detected in vivo. To establish binding in vivo, we treated cells released from a double thymidine block with the CDK1 inhibitor RO-3306 to allow cyclin B accumulation without CDK1 activation. We then assayed for complexes by two methods: immunoprecipitation of complexes with an antibody to cyclin B and gel filtration of the cell lysate. In RO-3306-treated cells cyclin B antibodies co-immunoprecipitated both CDK1 and WEE1 (Figure S3C). Furthermore, we also found that WEE1 migrated as a 191 kD complex in these cells (Figure S3D). Thus the complexes we observed in vitro are also present in cells and are likely to be important for understanding the kinetics of mitotic entry. Although they generally possess different localizations in vivo, there must be some interaction in order for WEE1 to inhibit CDK1 activation. When CDK1 is inactive WEE1 binds tighter to cyclin B-CDK1, which could facilitate inhibition, and the weakening of this interaction upon CDK1 activation could help reduce inhibition by WEE1.

WEE1 is Responsible for the Nonlinear Plateau Response to Cyclin B, whereas CDC25 Mediates the Stage 2 to Stage 3 Transition

To explore the relationship of CDKY15 dephosphorylation to CDK1 activity we perturbed the WEE1 and CDC25 feedback loops. Even in the well-studied frog system there is not a full understanding of how the WEE1/MYT1 and CDC25 feedback loops set the amplitude of CDK1 activity and the threshold of the response to cyclin B. To address these issues we used small molecule inhibitors of WEE1 (WEE1 Inhibitor 2; W2)(Palmer et al., 2006) and CDC25 (BN82002)(Brezak et al., 2004) to perturb these feedback loops.

We found that regulation of CDK1 Y15 phosphorylation and dephosphorylation determines the characteristic features of the extract response to cyclin B. Exposure to 24 μM of the WEE1 inhibitor, W2, converted the three-stage H1 kinase activation curve to a simple linear one with saturation (Figure 3A) and abolished Y15 phosphorylation (Figure 3B). The carrier DMSO had no effect on CDK1 activation (Figure 3A). Absent a stage 2 plateau, only 44 nM cyclin B was needed for detectable phosphorylation of CDC25C (as compared to 422 nM in the absence of W2), >80% of CDC25C showed an electrophoretic shift in response to 200 nM cyclin B (as compared to 1890 nM in the absence of W2) (Figure 3C). The conversion of CDK1 activation to a simple saturation curve casts doubt on the role of other proteins, such as CDK inhibitors, in the cyclin B dose response curve, as had been previously suggested (Amador et al., 2007; Lee and Kirschner, 1996; Marlovits et al., 1998). If other factors were responsible for the nonlinear response, these should have still exerted their control in the absence of Y15 phosphorylation, which was not seen.

The feedback inhibition of CDK1 through Y15 phosphorylation is lifted by CDK1 and its inhibition of WEE1, constituting one of the well studied double feedback loops in CDK1 activation. However in yeast it has been argued that there is an additional loop by which an initial phosphorylation of WEE by CDK1 activates WEE1 followed by the well known inactivating phosphorylation of WEE1 (Harvey et al. 2005). If such a feedback existed in mammalian cells it could help explain why CDK1 activation shows a plateau stage rather than linear activation, since the activity of WEE1 would increase as CDK1 is activated, compensating for the increased activity of CDK1 in response to increasing levels of cyclin B. One clue to the existence of such an activation process for WEE1 in somatic cells is our discovery of WEE1 phosphorylation at low levels of cyclin B. To demonstrate a WEE1 activation loop, we incubated the cell extracts with RO-3306 to inhibit CDK1 (or DMSO, as a control) and added low concentrations of cyclin B. Active CDK1 in the presence of cyclin B promoted Y15 phosphorylation (Figure 3D). By contrast, when RO-3306 blocked the phosphorylation of WEE1 by CDK1, it also suppressed the phosphorylation of CDK1 by WEE1 on Y15, arguing for activation of WEE1 mediated by CDK1. At 300 nM cyclin B, when CDK1 phosphorylation should be near maximum, the level of phosphorylation on CDK1 after RO-3306 treatment was approximately half (43 ± 18%, n=6) of what is was in the absence of CDK1 inhibition. It seemed unlikely that RO-3306 is inhibiting WEE1 as is quite specific for CDK1 over similar serine/threonine kinases and WEE1 is an unrelated tyrosine kinase. However, we tested this possibility with an order of addition experiment in which RO-3306 was either before or after WEE1 phosphorylation. When RO-3306 was added before WEE1 phosphporylation, Y15 phosphorylation was reduced, yet there was no effect on Y15 phopshorylation if WEE1 was phosphorylated by CDK1 before RO-3306 was added (Figure S3E). This result argues against RO-3306 inhibiting WEE1 directly.

These observations suggest that Y15 phosphorylation is fully responsible for the plateau in CDK1 activation (phase 2). We presumed that CDC25 must play a role in the escape from the plateau (phase 3. In the presence of 1.2 mM BN82002 to block CDC25 and Y15 dephosphorylation (Figure 3B), cyclin B was able to activate CDK1, but only to the low level as achieved at the plateau stage (phase 2) (Figure 3A). A switch to full CDK1 activation was not observed. Specifically, 4000 nM cyclin B in BN82002-treated extracts produced less CDK1 activity than 614 nM cyclin B in control extracts. Hence the activation of CDC25 and removal of the Y15 phosphorylation is required for the final step of CDK1 activation, the phase 2-to-phase 3 transition.

Kinetics of CDK1 Activation

In Xenopus embryos positive feedback by CDC25 and negative feedback by WEE1 prevent low levels of cyclin B from activating CDK1. There is also a temporal delay in the process so that even when cyclin B concentrations exceed a threshold, there is a pause of 15 min before CDK1 activation. In the abbreviated embryonic cell cycles this delay makes a major contribution to mitotic timing; a similar delay should be much less significant in the somatic cell cycle. When either a low concentration of cyclin B (100 nM), promoting partial activation, or a high concentration of cyclin B (1000 nM), promoting full activation, was added there was a delay of 8 to 16 min before steady-state activity appeared (Figure 4A). The explanation for this delay in the embryonic system is that it corresponds to the of time for CDK1 to be phosphorylated on Y15 and removed by CDC25. This cannot be the explanation in somatic extracts, since there is a lag upon addition of 100 nM cyclin B, an amount insufficient to activate CDC25.

Figure 4.

Figure 4

Open in a new tab

The kinetics of the extract response

(A) Autoradiogram of a dried polyacrylamide gel showing the time course of histone H1 kinase activation. Cell extract was mixed with 1000 nM cyclin B on ice and then incubated at room temperature. Aliquots were removed at the indicated time points, frozen in liquid nitrogen, and kinase activity was analyzed as in Figure 1.

(B) Immunoblot showing time course of CDK1 Y15 phosphorylation at the indicated concentrations of cyclin B.

(C) Rate of binding of cyclin B to CDK1. Cyclin B bound to Ni-NTA was added to the extract at a concentration of 1000 nM and isolated at the indicated time points. Bounds proteins were analyzed by immunoblotting for cyclin B and CDK1.

To probe further the mechanism of this delay in somatic cell extracts, we measured the rate of Y15 phosphorylation in the absence of CDC25 activation by using a stage 2 concentration of cyclin B (250 nM) (Figure 4B, bottom). Y15 phosphorylation was first detectable over background levels at 2 min and maximal levels were reached by 8 min, a rate similar to what had been observed in the frog system in the absence of active CDC25 (Solomon et al., 1990). We also attempted to measure the rate of Y15 phosphorylation in the presence of 950 nM cyclin B (stage 3) where CDC25 is activated (Figure 4A, top), but we saw no increase in Y15 phosphorylation at any time between cyclin B addition and CDK1 activation (0–8 min). Rather, we found a gradual loss of the Y15 signal over time (Figure 4B, top). Thus, the time delay for CDK1 activation is observed even when not enough cyclin B is added to activate CDC25 and occurs without increased Y15 phosphorylation during the lag. These results strongly imply that the dephosphorylation of Y15 cannot mediate the delay in somatic cells.

An alternative explanation for the delay in CDK1 activation is slow binding of cyclin B to CDK1. To test this idea, 1000 nM His6-cyclin B bound to Ni-NTA agarose was added to the extract, and at various times the beads were re-isolated and immunoblotted for cyclin B and CDK1. CDK1 was found associated with cyclin within one minute of incubation (Figure 4C). The amount of bound CDK1 increased over time reaching a maximum at 20 min (Figure 4C). Complete binding of cyclin B and CDK1 was much slower than would be achieved by diffusion, which should be complete within seconds, and suggests additional inter- or intramolecular events preceding their stable interaction. The slow rate of stable association would explain the 16 min required to achieve maximum activity. Presumably, CDK1 T161 phosphorylation by CAK or conformational rearrangements upon cyclin B binding, possibly mediated by chaperones could be responsible for the lag before full CDK1 activation. Thus, the frog and somatic cell extracts show a similar time delay between cyclin B addition and CDK1 activation, but the underlying mechanisms appear to be different.

The Response to Cyclin A in Somatic Cell Extracts

As discussed below, the role of cyclin B independent of cyclin A in mitosis is relevant to early cytosolic events that occur in intact cells, as cyclin A is initially sequestered in the nucleus. At some point, however, mitotic entry involves both cyclin A and cyclin B acting together. The role of cyclin A in mitosis is enigmatic: it is not required for cell division in the frog embryo until after gastrulation and, similarly, it is not required for the in vitro cell cycles of egg extracts. By contrast, blocking cyclin A function in mammalian cells causes an arrest at the G2/M transition (Fung et al., 2007; Furuno et al., 1999). We examined the function of cyclin A in the initiation of mitosis in cell extracts. As with cyclin B, extracts were prepared before cyclin A levels reached their mitotic maximum (Figure 1C). Previous experiments had shown that when cyclin A is added to an interphase frog egg extract, histone H1 kinase activity becomes linear with respect to cyclin A concentration until saturation (Clarke et al., 1992; Minshull et al., 1990). By contrast we find that the response to cyclin A in somatic extracts is nonlinear (Figure 5A). At low concentrations the H1 kinase activity is weak: as cyclin A concentration was increased ~4.5-fold, from 22 to 100 nM, CDK1 activity increased by only 34%. There is a steep increase in H1 kinase activity beyond this ~100 nM threshold, after which CDK1 activity was nearly proportional to the amount of cyclin A added until it saturated at ≥200 nM.

Figure 5.

Figure 5

Open in a new tab

Cyclin A activation of CDKs

(A) Activation of CDKs by His6-cyclin A. Recombinant His6-cyclin A was added at the indicated concentrations to the somatic cell extract and incubated for 20 min at room temperature. Histone H1 kinase activity was measured. Shown is an autoradiogram of a dried gel containing radiolabeled histone H1 (top) and the resulting (n=3) quantification (bottom). Error bars represent ± one standard deviation.

(B) Binding of CDK1 and CDK2 to cyclin A. BSA or 100 nM His6-cyclin A was added to the extracts. Ni-NTA agarose beads were added to both samples. The beads were washed three times in buffer. Shown are immunoblots for His6 (cyclin A), CDK1, and CDK2.

(C) Immunodepletion of CDKs. Extracts were immunodepleted for CDK2 (bottom), CDK1 (middle), or mock (top), and cyclin A was added. H1 kinase activity was measured, and an autoradiogram is shown.

(D) CDK1 sets sensitivity to cyclin A. Buffer (top) or CDK1 (bottom) was added to extract along with cyclin A and H1 kinase activity was measured. Shown is an autoradiogram.

(E) Phosphorylation of CDK substrates by cyclin A-CDK complexes. WEE1 (top) or securin (bottom) were synthesized and labeled with S35 using in vitro translation. His6-cyclin A was added to the extracts at the indicated concentrations. After 20 min, the extracts were subject to SDS-PAGE and autoradiography.

Cyclin A can bind and activate both CDK2 and CDK1. To determine whether this dual specificity was recapitulated in the extract we added 100 nM recombinant His6-tagged cyclin A (or, as a control, BSA) to the extract, incubated the reaction for 20 min, and re-isolated the recombinant cyclin A using Ni-NTA agarose. Cyclin A binds to both CDK2 and CDK1 (Figure 5B). To measure the contribution of CDK2 or CDK1 to cyclin A-stimulated activity, we assayed the activation profiles after immunodepletion of either kinase (Figure 5C). The overall response to cyclin A was similar when CDK2 was immunodepleted, indicating that CDK1 is clearly the major determinant of the nonlinear cyclin A activity (Figure 5C). A minor role for CDK2 is consistent with CDK1 being a much more potent kinase towards histone H1. The immunodepletion of CDK1 confirmed the dominant role of this kinase, as higher levels of cyclin A were needed to produce activity with CDK2 as the lone cyclin A binding partner (Figure 5C). Although cyclin A was able to bind to both CDK1 and CDK2, the total CDK activity was less than half of that produced by cyclin B (compare Figure 5A to Figure 1C). Thus, the nature of the bound cyclin influences the magnitude of the kinase activity, using histone H1 as a substrate. The level of CDK sets the response to cyclin A. When we supplemented the endogenous CDK in the extract with recombinant 1 μM CDK1 the increased level sensitized the response so that less than half the amount of cyclin A (90 nM compared to 227 nM), produced full CDK activity (Figure 5D).

The cyclin subunit of the cyclin-CDK complex is thought to affect substrate affinity and specificity (Minshull et al., 1990). The somatic cell extract system allowed us to test specificity differences between cyclin A and B kinases under in vivo-like conditions with competing substrates and phosphatases. Cyclin A produces sufficient kinase activity potentially to phosphorylate securin or WEE1. Indeed, cyclin A promoted substantial WEE1 phosphorylation but did not promote securin phosphorylation (Figure 5E). Thus, cyclin A may not merely generate additional kinase activity but may provide a different spectrum of substrate specificities..

Cyclin A Primes both the Enzymatic activity of Cyclin B – CDK1 and Nuclear Membrane Breakdown

The cyclin A requirement for mitotic entry can be shown not to be absolute. Recently it had been found that mouse embryonic fibroblasts can enter mitosis with a small G2 delay without cyclin A: it appears that these cells compensate for the loss of cyclin A by upregulating cyclin E and maintaining through G2, well after it is normally degraded (Kalaszczynska et al., 2009). Also, in cultured cells, the depletion of cyclin A can be overcome through either the overexpression of cyclin B or through the concomitant knockdown of WEE1 (Fung et al., 2007). These results suggest that physiological levels of cyclin B alone are insufficient for mitotic entry: cells must be sensitized to these cyclin B levels by another cyclin, usually A, or by increased cyclin B expression. Attempts cyclin A or cyclin B dependent kinase activity have produced conflicting data in the Xenopus egg system (Devault et al., 1992; Walker and Maller, 1991). Because cyclin A-dependent activity peaks prior to cyclin B-dependent activity in vivo (Goldstone et al., 2001), we asked whether a high level of cyclin A affects the response of the extract to cyclin B. We measured H1 kinase activity as a function of increasing cyclin B levels in extracts containing a concentration of cyclin A (200 nM) sufficient to generate high cyclin A-CDK activity. As shown in Figure 6A, at this concentration, cyclin A converted the complex three phase cyclin B response to a simple hyperbolic saturation curve, displaced to low concentrations of cyclin B.

Figure 6.

Figure 6

Open in a new tab

Cyclin A enhances cyclin B-CDK1 activity

(A) Representative autoradiogram showing cyclin B-dependent H1 kinase activity in the presence or absence of added 200 nM cyclin A (top). Graphical representation of CDK activity for a typical extract containing cyclin A and cyclin B (bottom) with error bars representing ± one standard deviation (n=2).

(B) Shown is an immunoblot for CDK1 Y15 phosphorylation levels for extract containing buffer (top) or cyclin A (bottom) as a function of increasing cyclin B concentration.

(C) The envelopes of nuclei was stained with rhodamine-conjugated DiI (pink) and the DNA with DAPI (blue). The nuclei were added to extract containing the indicated treatments at 30° C and imaged with a 60X objective 60 min later (top). The quantification of nuclear disintegration from three independent experiments (n=50 per trial) is shown ± one standard deviation (below).

Cyclin A drives WEE1 phosphorylation (Figure 5E), suggesting this phosphorylation may affect the cyclin B-CDK1 dose response this phosphorylation. Cyclin A linearized the cyclin B response, exactly the same effect produced by WEE1 inactivation (see Figure 3A). To confirm this effect is through inhibition of WEE1, we tested the efficiency of cyclin B at driving CDK1 Y15 phosphorylation both in the presence and absence of cyclin A. As the concentration of cyclin B was increased, higher levels of CDK1 Y15 phosphorylation were generated (Figure 6B, top). In contrast, when cyclin A was present (see Figure S4 for combined cyclin WEE1 phosphorylation), increasing levels of cyclin B, not only failed to increase CDK1 Y15 phosphorylation, but actually decreased the level (Figure 6B, bottom). Thus, cyclin A facilitates the activation of cyclin B-CDK1, by acting to inhibit WEE1 activity, with no indication that it acts to activate WEE1 in the first place.

Cyclin A may also mediate cyclin B activity by stimulating nuclear envelope breakdown to allow cellular mixing of cyclin A and cyclin B (Gong et al. 2007). To observe this effect nuclei were isolated from HeLa S3 cells; their membranes and DNA were first stained with lipophilic DiI-rhodamine and DAPI, respectively; and the stained nuclei were incubated in extracts to which were added buffer or mitotic cyclins. For the two hour course of the experiment, virtually all of the nuclei remained intact when the extract was supplemented with buffer (PBS) or cyclin A (300 nM) as evidenced by the continuous ring of membrane signal and its enclosed DNA (Figure 6C). By contrast, stage 3 cyclin B levels (2000 nM) caused extensive disintegration of the nucleus as seen by holes in and vesiculation of the envelope and extrusion of decondensed DNA into the extract. Some form of envelope breakdown was seen in 37% of the nuclei within 60 min, compared to 3% with PBS or 6% with cyclin A (Figure 6C). Nuclear breakdown was significantly augmented through the addition of both cyclin A and cyclin B to the extract, which nearly doubled the total nuclear envelope breakdown within 60 min; ~90% of the nuclei were broken down by 120 min. At these concentrations of cyclins A and B together produced the same level of CDK activity as cyclin B alone (compare Figure 2A to Figure 6A), revealing the specific effect of cyclin A. Finally these experiments show that the extract system is capable of recapitulating one of the major morphological events of mitosis with appropriate regulation by mitotic cyclins.

Discussion

The lack of a biochemical system for understanding CDK1 regulation in somatic cells has been a barrier to distinguishing which features of this highly integrative and robust regulatory network may general and which may be the specific specializations related to the unique lifestyle of eggs. To overcome this limitation we have established a complementary in vitro system for mammalian somatic cells. Here we consider the biochemical architecture of the somatic cell mitotic regulatory network as it exists in homogeneous extracts and speculate on how these basic findings can be applied to a compartmentalized and heterogeneous cell.

The Biochemical Network of Mitotic Regulation

Frog blastomeres are well provisioned and have an autonomous cell cycle that does not respond to signals produced by other cells or even to drastic environmental signals that cause DNA or spindle damage. Moreover, they divide very rapidly with very little variability so that cell division remains in tight balance with neighboring cells, despite the lack of communication. By contrast, slowly dividing somatic cells communicate with neighboring cells, respond to distant cellular signals, and arrest in the presence of environmental stressors. These fundamental differences between early embryos and somatic cells suggest that common regulatory genes could be deployed differently.

We found that high levels of cyclin B in somatic cell extracts are sufficient to promote mitotic entry autonomously: it induces nuclear disintegration as well as the standard biochemical markers of mitotic entry including the activation of CDK1, inhibition of WEE1, activation of the CDC25 feedback loops and phosphorylation of mitotic substrates. We interpret our data in somatic cells as sequential steady-state events driven by the slow cyclin B accumulation, in contrast to the rapid kinetic events of cyclin B-CDK1 complex formation and the Y15 phosphorylation/dephosphorylation characteristic of eggs. As cyclin B accumulates in somatic cells it forms a complex with CDK1 and WEE1. This leads to an observed suppression of CDK1 activation mediated by the WEE1/MYT1 negative feedback loop and phosphorylation of CDK1 on Y15. The tight WEE1 binding to CDK1 could promote its inhibition of CDK1. Because of these inhibitory events, cyclin B levels can increase over a large range (100–400 nM), with little increase in CDK1 activity. WEE1/MYT1 suppression is incomplete, however, and is stimulated by low levels of CDK1 activity, as has been suggested in budding yeast (Harvey et al., 2005). The measurable low level of CDK1 at these early stages avoids a longstanding paradox from the frog extract experiments: how can Y15-phosphorylated CDK1 initiate the double feedback loops if it is completely inactive? In the somatic cell system it is partially active and in a dynamic balance. At a defined threshold (~400–500 nM), a small increase in cyclin B levels causes a dramatic increase in CDK1 activity, mediated by CDC25. Our results suggest cells can remain in G2 indefinitely in the presence of low levels of CDK1, as long as CDC25 is inhibited. This role for CDC25 underscores its importance in the DNA damage control and in the response to incomplete DNA replication as CDC25 and is degraded or sequestered from CDK1 during cell cycle checkpoints.

Mitosis requires that events occur in a particular order; the complex pattern of CDK1 activation found in the in vitro system provides a means to accomplish this (Figure 7A). We observed that a low concentration of cyclin B leads to WEE1 phosphorylation; such partial WEE1 phosphorylation has been observed in synchronized cell populations prior to mitotic entry (Figure 1C)(Watanabe et al., 1995). In our concentrated (in vivo-like) extracts this ordering may be a result of the combined effects of competing substrates and endogenous phosphatases. Phosphorylation of the APC does not occur until late in mitosis, consistent with our finding that CDC27 phosphorylation requires high CDK1 activity (Lindqvist et al., 2007). We easily recorded CDK1 activation at cyclin B concentrations that did not completely eliminate CDK1 Y15 phosphorylation. Although this is not seen in frog eggs, early events such as centrosome separation can take place when most of CDK1 is Y15 phosphorylated (Lindqvist et al., 2007). Also, in somatic cells in vivo it has been observed that dephosphorylation of CDK1 Y15 is gradual and starts before nuclear envelope breakdown and continues through prometaphase, which encompasses a series of cellular changes that require active CDK1 (Lindqvist et al., 2007). Thus, a combination of quantitative measurements in both in vitro and in vivo situations provide key clues to the regulation of mitotic entry and suggest that the dynamic changes are very different from the embryonic systems.

Figure 7.

Figure 7

Open in a new tab

Model of mitotic entry.

(A) Schematic showing the relationship between CDK1 activity and substrate phosphorylation. See text for details.

(B) In G2 cyclin B accumulates throughout the cytoplasm and forms inactive complexes with CDK1. The concentration of cyclin B locally increases through localization at the centrosomes near the nucleus. This increased concentration activates cyclin B-CDK1 complexes. These complexes then can enter the nucleus. Cyclin A in the nucleus prevents WEE1 from inactivating these complexes.

Cyclin B is the central regulator of mitosis, but understanding the contribution of cyclin A to this process is critical (Gong et al., 2007; Pagano et al., 1992). Cyclin A is present at the G1/S transition but continues to accumulate through the S and G2 phases (Figure 1C)(Goldstone et al., 2001). Cyclin A and cyclin B are each inadequate inducers of mitosis. Even high levels of cyclin A produces less CDK activity than cyclin B will achieve. Additionally, cyclin A-CDK1 cannot generate the complete set of mitotic phosphoproteins. Cyclin B, without cyclin A, can induce mitotic entry, but only at high levels in vitro or artificially high levels generated by overexpression in vivo. These observations indicate an essential coupling between cyclin A and cyclin B in somatic cells. We suggest that active cyclin A-CDK weakens the WEE1/MYT1 negative feedback to permit the activation of cyclin B at lower concentrations. The use of two cyclins and connected regulatory networks requires that multiple criteria are met prior to mitotic entry to prevent a premature cell cycle transition.

Implications for Compartmentalized Cells

Our findings can be extended to from this homogenous system to generate a more mechanistic understanding of mitosis in compartmentalized, intact cells (Figure 7B). We suggest differential localization perturbs the basic biochemical features observed in the cell-free system, imparting new features, without changing the molecular relationships. In late S and G2 phases, cyclin B accumulates in the cytoplasm segregated from the cyclin A in the nucleus. The combination of the low cytoplasmic concentrations of cyclin B with the fact that cyclin A at this point is largely nuclear keeps the cell out of mitosis (Pines and Hunter, 1991). As cyclin B accumulates in the cytosol, the cell may acquire a high local concentration in some subcompartment that could itself trigger a localized stage 2 to stage 3 transition. There is evidence that cyclin B and CDC25 can accumulate at the G2/M transition at the centrosome (Jackman et al., 2003; Lindqvist et al., 2005; Lindqvist et al., 2007). Thus, even at global concentrations of cyclin B too low to propel the cell into mitosis, the high local concentration could activate CDK1 through a cyclin A-independent mechanism by raising the cyclin B concentration above the phase 2-to-phase 3 transition. In the absence of cyclin A, cyclin B-CDK1 seems to be activated on the centrosomes as judged by CDK1 Y15 and MPM2 (a marker of mitotic phosphorylations) staining but not in the nucleus (De Boer et al., 2008). Once locally activated, CDK1 could generate even more active CDK1 through the established positive feedback loops to drive the rest of the cytoplasm into a mitotic state. As the cytoplasm becomes generally more mitotic, cyclin B-CDK1 will be imported into the nucleus (Heald et al., 1993; Lindqvist et al., 2007). The concentration of WEE1 is higher in the nucleus than in the cytoplasm and would initially suppress the imported cyclin B-CDK1 (Heald et al., 1993). However, because cyclin A is present and active in the nucleus, it could drive WEE1 inactivation through phosphorylation, which would then allow active cyclin B to accumulate without immediate inhibition by WEE1. Indeed, we observed that cyclin A promotes the disintegration of nuclei by cyclin B, an observation supported by in vivo studies (Gong et al. 2007). With sufficient activation, cyclin B mixes with cyclin A in the nucleus, the nuclear envelope would begin to disintegrate, and the entire cell will be quickly driven into mitosis. Thus compartmentation, the rising concentrations of cyclin A and cyclin B at mitosis, and the quantitative features of the various mitotic circuits described here may provide the essential features to understand the logic and robustness of this process. These biochemical reconstitution and perturbation studies should ultimately lead to quantitative dynamic models, as has occurred in frog egg extracts (Pomerening et al. 2005), and thus we should be able to achieve a more predictive description of how concerted cell cycle transitions are achieved in somatic cells and how they are affected by environmental conditions.

Experimental Procedures

Purification of Recombinant Proteins and Size Exclusion Chromatography

Human His6-Cyclin A2 (from M. Rape) and human His6-tagged-Cdk1 (from M. Springer) clones were individually transformed into BL21 E. coli (Stratagene). IPTG was added to induce protein expression for 3 hr. Human His6-cyclin B1 baculovirus (from R. King), amplified, and used to infect 1.5 L of Sf9 cells grown in suspension (1:25 dilution) for 48 hr. Recombinant proteins from both sources were purified by Ni-NTA immobilized metal affinity chromatography following instructions manufacturer (Qiagen) instructions. Protein concentration was determined by performing SDS-PAGE on a series of sample dilutions, staining the gel with Gelcode blue (Pierce), and comparing the signal obtained with that of a BSA standard using a LI-COR Odyssey. For gel filtration, recombinant protein or extract was applied to a Superdex 200 column equilibrated in running buffer (20 mM Tris, pH 7.7, 150 mM NaCl, 1 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT).

Tissue Culture, Flow Cytomtery, and Somatic Cell Extract Preparation

HeLa S3 cells growing in DMEM (Gibco) supplemented with 10% fetal bovine serum (Cell Gro), 100 I.U./ml penicillin (Cell Gro), and 100 μg/ml streptomycin (Cell Gro) were arrested in S phase by the addition of 2 mM thymidine (Sigma) 24 hr. Cells were washed 2 times in warmed DMEM, and released into fresh medium. After 8–10 hr of growth in thymidine free DMEM, 2 mM was again added. Cells then were washed and released in fresh medium for 4–6 hr. For flow cytometry the cells were washed in PBS, fixed in cold a 50% EtOH/PBS solution, and stored at −20° C. The fixed cells were then permeablilized, treated with RNase A, and stained the propidium iodide. DNA content was analyzed on a FACSCalibur. To prepare extracts, HeLa S3 cells were harvested and washed 3 times with ice cold PBS. The cell pellet was resuspended in twice the pellet volume with hypotonic lysis buffer (20 mM Hepes, pH 7.7, 5 mM KCl, 1.5 mM MgCl2, 1 mM DTT, 1 EDTA free complete protease inhibitor cocktail tablet). After swelling for 30 min on ice, the cells then were disrupted by seven strokes of a loose-fitting Dounce homogenizer on ice. A two-pellet volume of Buffer B (50 mM Hepes, pH 7.7, 10 mM MgCl2, 2 mM DTT, 25% sucrose, 50% glycerol) was added to the lysed cells. The lysate was stirred at 4° C, and (NH4)2SO4 was added to a final concentration (w/v) of 10%. After 30 min the extract was cleared by centrifugation at 5,000 G. The supernatant was removed and dialyzed against extract dialysis buffer (25 mM Hepes, pH 7.7, 100 mM K-glutamate, 5 mM MgCl2, 0.5 mM EDTA, 2 mM DTT, 250 mM sucrose, and 10% glycerol) at 4° C. The extract was concentrated until reaching the volume of the initial cell pellet (typically 10–20 mg/ml), frozen in liquid N2, and stored at −80° C until use.

Nuclear Isolation and Envelope Breakdown Assay

HeLa S3 cells were washed twice in cold PBS. Buffer (10 mM Tris, pH 7.5; 2 mM MgCl2; 250 mM sucrose; 0.2% NP-40; 1 mM PMSF) was added to the cells, which were then incubated on ice for 30 min. Following centrifugation at 200 G for 5 min, the nuclear pellet was washed in the same buffer and stored at −80° C. For the nuclear envelope breakdown assays the nuclei were thawed and incubated with 2.5 μM DiI conjugated to rhodamine (Invitrogen) first at 37° C for two minutes and then at 4 °C for 15 min to label the nuclear membrane. The nuclei were washed with PBS to remove unbound DiI and added to the somatic cell extract at 100–1000 nuclei per μl. The extracts were incubated at 30° C with samples taken at various time points and stained with DAPI (1 μg/ml). Images of nuclei were captured with a Hamamatsu CCD camera and analyzed with MetaMorph software.

Immunodepletions and Immunoblotting

Immunodepletions were performed as in Ho et al. (2004) using 7 μg of antibody or control normal rabbit IgG and 50 μl of protein G beads. For immunoblotting, 2–5 μl of extract was subjected to SDS-PAGE. The protein was transferred to Protran nitrocellulose membrane (Scleicher and Schuell) using semi-dry transfer according to the manufacturer instructions (BioRad). The membrane was incubated in blocking buffer (0.1% casein, 0.2X PBS or 5% milk, 1X TBST) for 1hr at room temperature. The primary antibodies were diluted in blocking buffer and incubated 1hr at room temperature. The follow dilutions were used: 1:750 αWEE1 (Santa Cruz); 1:500 αCDC27 (Santa Cruz); 1:1000 anti-CYCLIN A (Santa Cruz), 1:1000 anti-CYCLIN B (Santa Cruz), anti-CHK1 (Santa Cruz), anti-CHK1-p-S345 (Cell Signaling), anti-CHK2 (Upstate), anti-CHK2-p-T68 (Santa Cruz), anti-ATM-p-S1981 (Cell Signaling), anti-H3-p-S10 (Santa Cruz), anti-ACTIN (Sigma), anti-alpha TUBULIN (Sigma), 1:10,000 αCDK1-p-Y15 (Cell Signaling); 1:2000 αCDK1 (Santa Cruz); 1:2000 αCDK2 (Cell Signaling); 1:500 αCDC25C (Santa Cruz) 1:2000 αHis6 (Santa Cruz). The blots were washed 3–4 times with PBS and incubated for 45–60 min with the conjugated secondary antibodies (Molecular Probes) at a dilution of 1:2000–1:10,000 at room temperature. The blots again were washed 4 times and were scanned into a LI-COR Odyssey for analysis or visualized by ECL.

H1 Kinase Assays

H1 kinase activity was determined following the protocol of Stukenberg et al. (2001). A 0.5 μl aliquot was taken from the extract, diluted into 10 μl EB (15 mM MgCl2, 20 mM EGTA, 80 mM K-β-glycerophosphate, 10 mM DTT, and 10 μg/ml each of leupeptin, pepstatin, chymostatin, pH =7.3), snap frozen in liquid N2, and stored at −80° C until use. The diluted aliquots were thawed on ice, and 6 μl of H1 assay mix (267 μg/ml histone H1, 0.5 mM ATP, and 0.125 mu;Ci/μl ATPγP32 diluted in EB) was added. The mix was incubated at room temperature for 15 min, and the reaction was stopped with Laemmli buffer. Following SDS-PAGE gels then were fixed (30 min in 20% ethanol, 6% acetic acid), dried, and exposed to a PhosphorImager screen for 1–24 hr. The exposed screen was scanned into a BioRad Molecular Imager FX, and the BioRad Quantity One software was used to perform the quantification. Kinase activity was presented by dividing values by the background activity at 0 nM added cyclin.

Supplementary Material

01

Acknowledgments

We thank Michael Springer for his insights in the early part of this investigation and Allon Klein for help with the statistical analyses. We are grateful for Randy King, Michael Rape, and Sashank Reddy for kindly providing reagents. This research was supported by a Jane Coffin Childs Memorial Fellowship (to RWD) and grants from the National Institutes of Health (to MWK).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Amador V, Ge S, Santamaria PG, Guardavaccaro D, Pagano M. APC/C(Cdc20) controls the ubiquitin-mediated degradation of p21 in prometaphase. Mol Cell. 2007;27:462–473. doi: 10.1016/j.molcel.2007.06.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Brezak MC, Quaranta M, Mondesert O, Galcera MO, Lavergne O, Alby F, Cazales M, Baldin V, Thurieau C, Harnett J, et al. A novel synthetic inhibitor of CDC25 phosphatases: BN82002. Cancer Res. 2004;64:3320–3325. doi: 10.1158/0008-5472.can-03-3984. [DOI] [PubMed] [Google Scholar]
  3. Clarke PR, Leiss D, Pagano M, Karsenti E. Cyclin A- and cyclin B-dependent protein kinases are regulated by different mechanisms in Xenopus egg extracts. EMBO J. 1992;11:1751–1761. doi: 10.1002/j.1460-2075.1992.tb05227.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. De Boer L, Oakes V, Beamish H, Giles N, Stevens F, Somodevilla-Torres M, Desouza C, Gabrielli B. Cyclin A/cdk2 coordinates centrosomal and nuclear mitotic events. Oncogene. 2008;27:4261–4268. doi: 10.1038/onc.2008.74. [DOI] [PubMed] [Google Scholar]
  5. Desai D, Gu Y, Morgan DO. Activation of human cyclin-dependent kinases in vitro. Mol Biol Cell. 1992;3:571–582. doi: 10.1091/mbc.3.5.571. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Devault A, Fesquet D, Cavadore JC, Garrigues AM, Labbe JC, Lorca T, Picard A, Philippe M, Doree M. Cyclin A potentiates maturation-promoting factor activation in the early Xenopus embryo via inhibition of the tyrosine kinase that phosphorylates cdc2. J Cell Biol. 1992;118:1109–1120. doi: 10.1083/jcb.118.5.1109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Fung TK, Ma HT, Poon RY. Specialized roles of the two mitotic cyclins in somatic cells: cyclin A as an activator of M phase-promoting factor. Mol Biol Cell. 2007;18:1861–1873. doi: 10.1091/mbc.E06-12-1092. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Furuno N, den Elzen N, Pines J. Human cyclin A is required for mitosis until mid prophase. J Cell Biol. 1999;147:295–306. doi: 10.1083/jcb.147.2.295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Georgi AB, Stukenberg PT, Kirschner MW. Timing of events in mitosis. Curr Biol. 2002;12:105–114. doi: 10.1016/s0960-9822(01)00662-5. [DOI] [PubMed] [Google Scholar]
  10. Goldstone S, Pavey S, Forrest A, Sinnamon J, Gabrielli B. Cdc25-dependent activation of cyclin A/cdk2 is blocked in G2 phase arrested cells independently of ATM/ATR. Oncogene. 2001;20:921–932. doi: 10.1038/sj.onc.1204177. [DOI] [PubMed] [Google Scholar]
  11. Gong D, Pomerening JR, Myers JW, Gustavsson C, Jones JT, Hahn AT, Meyer T, Ferrell JE., Jr Cyclin A2 regulates nuclear-envelope breakdown and the nuclear accumulation of cyclin B1. Curr Biol. 2007;17:85–91. doi: 10.1016/j.cub.2006.11.066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Heald R, McLoughlin M, McKeon F. Human wee1 maintains mitotic timing by protecting the nucleus from cytoplasmically activated Cdc2 kinase. Cell. 1993;74:463–474. doi: 10.1016/0092-8674(93)80048-j. [DOI] [PubMed] [Google Scholar]
  13. Jackman M, Lindon C, Nigg EA, Pines J. Active cyclin B1-Cdk1 first appears on centrosomes in prophase. Nat Cell Biol. 2003;5:143–148. doi: 10.1038/ncb918. [DOI] [PubMed] [Google Scholar]
  14. Kalaszczynska I, Geng Y, Iino T, Mizuno S, Choi Y, Kondratiuk I, Silver DP, Wolgemuth DJ, Akashi K, Sicinski P. Cyclin A is redundant in fibroblasts but essential in hematopoietic and embryonic stem cells. Cell. 2009;138:352–365. doi: 10.1016/j.cell.2009.04.062. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Karlsson C, Katich S, Hagting A, Hoffmann I, Pines J. Cdc25B and Cdc25C differ markedly in their properties as initiators of mitosis. J Cell Biol. 1999;146:573–584. doi: 10.1083/jcb.146.3.573. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kim SY, Ferrell JE., Jr Substrate competition as a source of ultrasensitivity in the inactivation of Wee1. Cell. 2007;128:1133–1145. doi: 10.1016/j.cell.2007.01.039. [DOI] [PubMed] [Google Scholar]
  17. Kim SY, Song EJ, Lee KJ, Ferrell JE., Jr Multisite M-phase phosphorylation of Xenopus Wee1A. Mol Cell Biol. 2005;25:10580–10590. doi: 10.1128/MCB.25.23.10580-10590.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Larochelle S, Merrick KA, Terret ME, Wohlbold L, Barboza NM, Zhang C, Shokat KM, Jallepalli PV, Fisher RP. Requirements for Cdk7 in the assembly of Cdk1/cyclin B and activation of Cdk2 revealed by chemical genetics in human cells. Mol Cell. 2007;25:839–850. doi: 10.1016/j.molcel.2007.02.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lee TH, Kirschner MW. An inhibitor of p34cdc2/cyclin B that regulates the G2/M transition in Xenopus extracts. Proc Natl Acad Sci U S A. 1996;93:352–356. doi: 10.1073/pnas.93.1.352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lindqvist A, Kallstrom H, Lundgren A, Barsoum E, Rosenthal CK. Cdc25B cooperates with Cdc25A to induce mitosis but has a unique role in activating cyclin B1-Cdk1 at the centrosome. J Cell Biol. 2005;171:35–45. doi: 10.1083/jcb.200503066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Lindqvist A, van Zon W, Karlsson Rosenthal C, Wolthuis RM. Cyclin B1-Cdk1 activation continues after centrosome separation to control mitotic progression. PLoS Biol. 2007;5:e123. doi: 10.1371/journal.pbio.0050123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Marlovits G, Tyson CJ, Novak B, Tyson JJ. Modeling M-phase control in Xenopus oocyte extracts: the surveillance mechanism for unreplicated DNA. Biophys Chem. 1998;72:169–184. doi: 10.1016/s0301-4622(98)00132-x. [DOI] [PubMed] [Google Scholar]
  23. Minshull J, Golsteyn R, Hill CS, Hunt T. The A- and B-type cyclin associated cdc2 kinases in Xenopus turn on and off at different times in the cell cycle. EMBO J. 1990;9:2865–2875. doi: 10.1002/j.1460-2075.1990.tb07476.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Okamoto K, Nakajo N, Sagata N. The existence of two distinct Wee1 isoforms in Xenopus: implications for the developmental regulation of the cell cycle. EMBO J. 2002;21:2472–2484. doi: 10.1093/emboj/21.10.2472. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Okamoto K, Sagata N. Mechanism for inactivation of the mitotic inhibitory kinase Wee1 at M phase. Proc Natl Acad Sci U S A. 2007;104:3753–3758. doi: 10.1073/pnas.0607357104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Pagano M, Pepperkok R, Verde F, Ansorge W, Draetta G. Cyclin A is required at two points in the human cell cycle. EMBO J. 1992;11:961–971. doi: 10.1002/j.1460-2075.1992.tb05135.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Palmer BD, Thompson AM, Booth RJ, Dobrusin EM, Kraker AJ, Lee HH, Lunney EA, Mitchell LH, Ortwine DF, Smaill JB, et al. 4-Phenylpyrrolo[3,4-c]carbazole-1,3(2H,6H)-dione inhibitors of the checkpoint kinase Wee1. Structure-activity relationships for chromophore modification and phenyl ring substitution. J Med Chem. 2006;49:4896–4911. doi: 10.1021/jm0512591. [DOI] [PubMed] [Google Scholar]
  28. Pines J, Hunter T. Human cyclins A and B1 are differentially located in the cell and undergo cell cycle-dependent nuclear transport. J Cell Biol. 1991;115:1–17. doi: 10.1083/jcb.115.1.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Pomerening JR, Kim SY, Ferrell JE., Jr Systems-level dissection of the cell-cycle oscillator: bypassing positive feedback produces damped oscillations. Cell. 2005;122:565–578. doi: 10.1016/j.cell.2005.06.016. [DOI] [PubMed] [Google Scholar]
  30. Pomerening JR, Sontag ED, Ferrell JE., Jr Building a cell cycle oscillator: hysteresis and bistability in the activation of Cdc2. Nat Cell Biol. 2003;5:346–351. doi: 10.1038/ncb954. [DOI] [PubMed] [Google Scholar]
  31. Russo AA, Jeffrey PD, Pavletich NP. Structural basis of cyclin-dependent kinase activation by phosphorylation. Nat Struct Biol. 1996;3:696–700. doi: 10.1038/nsb0896-696. [DOI] [PubMed] [Google Scholar]
  32. Sherwood SW, Rush DF, Kung AL, Schimke RT. Cyclin B1 expression in HeLa S3 cells studied by flow cytometry. Exp Cell Res. 1994;211:275–281. doi: 10.1006/excr.1994.1087. [DOI] [PubMed] [Google Scholar]
  33. Solomon MJ, Glotzer M, Lee TH, Philippe M, Kirschner MW. Cyclin activation of p34cdc2. Cell. 1990;63:1013–1024. doi: 10.1016/0092-8674(90)90504-8. [DOI] [PubMed] [Google Scholar]
  34. Walker DH, Maller JL. Role for cyclin A in the dependence of mitosis on completion of DNA replication. Nature. 1991;354:314–317. doi: 10.1038/354314a0. [DOI] [PubMed] [Google Scholar]
  35. Watanabe N, Broome M, Hunter T. Regulation of the human WEE1Hu CDK tyrosine 15-kinase during the cell cycle. EMBO J. 1995;14:1878–1891. doi: 10.1002/j.1460-2075.1995.tb07180.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Xu N, Chang DC. Different thresholds of MPF inactivation are responsible for controlling different mitotic events in mammalian cell division. Cell Cycle. 2007;6:16. doi: 10.4161/cc.6.13.4385. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS188932-supplement-01.pdf (268.9KB, pdf)

RESOURCES