Cdc37: a protein kinase chaperone?

The activity of most protein kinases is highly regulated, typically via phosphorylation and/or subunit association. However, the folding of protein kinases into an active state or a form capable of activation is now emerging as another important step through which they can be regulated. The 50-kDa protein Cdc37 and the associated heat-shock protein Hsp90 have been found to bind to, and be required for the activity of, diverse protein kinases, including Cdk4, v-Src, Raf and SEVENLESS. Together, Cdc37 and Hsp90 may act as a general chaperone for protein kinases, in particular those involved in signal-transduction pathways and cell-cycle control.

p21 and retinoblastoma protein control the absence of DNA replication in terminally differentiated muscle cells.

During differentiation, skeletal muscle cells withdraw from the cell cycle and fuse into multinucleated myotubes. Unlike quiescent cells, however, these cells cannot be induced to reenter S phase by means of growth factor stimulation. The studies reported here document that both the retinoblastoma protein (Rb) and the cyclin-dependent kinase (cdk) inhibitor p21 contribute to this unresponsiveness. We show that the inactivation of Rb and p21 through the binding of the adenovirus E1A protein leads to the induction of DNA replication in differentiated muscle cells. Moreover, inactivation of p21 by E1A results in the restoration of cyclin E-cdk2 activity, a kinase made nonfunctional by the binding of p21 and whose protein levels in differentiated muscle cells is relatively low in amount. We also show that restoration of kinase activity leads to the phosphorylation of Rb but that this in itself is not sufficient for allowing differentiated muscle cells to reenter the cell cycle. All the results obtained are consistent with the fact that Rb is functioning downstream of p21 and that the activities of these two proteins may be linked in sustaining the postmitotic state.

Dephosphorylation of Cdk2 Thr160 by the cyclin-dependent kinase-interacting phosphatase KAP in the absence of cyclin.

The activation of cyclin-dependent kinases (CDKs) requires the phosphorylation of a conserved threonine (Thr160 in Cdk2) by CDK-activating kinase (CAK). Human KAP (also called Cdi1), a CDK-associated phosphatase, was shown to dephosphorylate Thr160 in human Cdk2. KAP was unable to dephosphorylate Tyr15 and only dephosphorylated Thr160 in native monomeric Cdk2. The binding of cyclin A to Cdk2 inhibited the dephosphorylation of Thr160 by KAP but did not preclude the binding of KAP to the cyclin A-Cdk2 complex. Moreover, the dephosphorylation of Thr160 by KAP prevented Cdk2 kinase activity upon subsequent association with cyclin A. These results suggest that KAP binds to Cdk2 and dephosphorylates Thr160 when the associated cyclin subunit is degraded or dissociates.

Cell regulation. Innocent bystanders or chosen collaborators?

Cyclins and cyclin-dependent kinases turn out to hve diverse functions, not all directly concerned with the cell cycle; do they provide a link between cell-cycle control and other cellular processes?

MDM2 and MDMX bind and stabilize the p53-related protein p73.

The p53 gene encodes one of the most important tumor suppressors in human cells and undergoes frequent mutational inactivation in cancers. MDM2, a transcriptional target of p53, binds p53 and can both inhibit p53-mediated transcription [1] [2] and target p53 for proteasome-mediated proteolysis [3] [4]. A close relative of p53, p73, has recently been identified [5] [6]. Here, we report that, like p53, p73alpha and the alternative transcription product p73beta also bind MDM2. Interaction between MDM2 and p53 represents a key step in the regulation of p53, as MDM2 promotes the degradation of p53. In striking contrast to p53, the half-life of p73 was found to be increased by binding to MDM2. Like MDM2, the MDM2-related protein MDMX also bound p73 and stabilized the level of p73. Moreover, the growth suppression functions of p73 and the induction of endogenous p21, a major mediator of the p53-dependent growth arrest pathway, were enhanced in the presence of MDM2. These differences between the regulation of p53 and p73 by MDM2/MDMX may highlight a physiological difference in their action.

Regulation of cyclin A-Cdk2 by SCF component Skp1 and F-box protein Skp2.

Cyclin A-Cdk2 complexes bind to Skp1 and Skp2 during S phase, but the function of Skp1 and Skp2 is unclear. Skp1, together with F-box proteins like Skp2, are part of ubiquitin-ligase E3 complexes that target many cell cycle regulators for ubiquitination-mediated proteolysis. In this study, we investigated the potential regulation of cyclin A-Cdk2 activity by Skp1 and Skp2. We found that Skp2 can inhibit the kinase activity of cyclin A-Cdk2 in vitro, both by direct inhibition of cyclin A-Cdk2 and by inhibition of the activation of Cdk2 by cyclin-dependent kinase (CDK)-activating kinase phosphorylation. Only the kinase activity of Cdk2, not of that of Cdc2 or Cdk5, is reduced by Skp2. Skp2 is phosphorylated by cyclin A-Cdk2 on residue Ser76, but nonphosphorylatable mutants of Skp2 can still inhibit the kinase activity of cyclin A-Cdk2 toward histone H1. The F box of Skp2 is required for binding to Skp1, and both the N-terminal and C-terminal regions of Skp2 are involved in binding to cyclin A-Cdk2. Furthermore, Skp2 and the CDK inhibitor p21(Cip1/WAF1) bind to cyclin A-Cdk2 in a mutually exclusive manner. Overexpression of Skp2, but not Skp1, in mammalian cells causes a G1/S cell cycle arrest.

Redistribution of the CDK inhibitor p27 between different cyclin.CDK complexes in the mouse fibroblast cell cycle and in cells arrested with lovastatin or ultraviolet irradiation.

The cyclin-dependent kinase (CDK) inhibitor p27 binds and inhibits the kinase activity of several CDKs. Here we report an analysis of the behavior and partners of p27 in Swiss 3T3 mouse fibroblasts during normal mitotic cell cycle progression, as well as in cells arrested at different stages in the cycle by growth factor deprivation, lovastatin treatment, or ultraviolet (UV) irradiation. We found that the level of p27 is elevated in cells arrested in G0 by growth factor deprivation or contact inhibition. In G0, p27 was predominantly monomeric, although some portion was associated with residual cyclin A.Cdk2. During G1, all of p27 was associated with cyclin D1.Cdk4 and was then redistributed to cyclin A.Cdk2 as cells entered S phase. The loss of the monomeric p27 pool as cyclins accumulate in G1 is consistent with the in vivo and in vitro data showing that p27 binds better to cyclin.CDK complexes than to monomeric CDKs. In growing cells, the majority of p27 was associated with cyclin D1 and the level of p27 was significantly lower than the level of cyclin D1. In cells arrested in G1 with lovastatin, cyclin D1 was degraded and p27 was redistributed to cyclin A.Cdk2. In contrast to p21 (which is a p27-related CDK inhibitor and is induced by UV irradiation), the level of p27 was reduced after UV irradiation, but because cyclin D1 was degraded more rapidly than p27, there was a transient increase in binding of p27 to cyclin A.Cdk2. These data suggest that cyclin D1.Cdk4 acts as a reservoir for p27, and p27 is redistributed from cyclin D1.Cdk4 to cyclin A.Cdk2 complexes during S phase, or when cells are arrested by growth factor deprivation, lovastatin treatment, or UV irradiation. It is likely that a similar principle of redistribution of p27 is used by the cell in other instances of cell cycle arrest.

The role of Cdc2 feedback loop control in the DNA damage checkpoint in mammalian cells.

DNA damage inactivates cyclin-dependent kinases (CDKs) and arrests the cell cycle. Following DNA damage, the G1-S CDKs are inhibited by a mechanism involving p53-dependent induction of p21Cip1/Waf1; but how the Cdc2 is inhibited is less apparent. We found that the signal generated by the DNA damage checkpoint in G2 was dominant over that from the spindle microtubule-assembly checkpoint, because the high Cdc2 activity present in nocodazole or Taxol-arrested cells was reduced by DNA damage. Phosphorylation of the inhibitory residues in Cdc2, Thr14, and Tyr15 coincided with the inactivation of Cdc2 after DNA damage. Interpretation of this result, however, was not straightforward due to the regulation of Thr14/Tyr15 phosphorylation by feedback loops; hence, their phosphorylation can in principle result merely from the inhibition of Cdc2 activity. Consistent with this, Thr14/Tyr15 phosphorylation was induced when Cdc2 kinase activity was inhibited with butyrolactone-I. Given these complications, we undertook a more critical analysis of the mechanisms that regulate Cdc2 after DNA damage. Caffeine reversed the DNA damage-induced inhibition of Cdc2 by causing dephosphorylation of Cdc2, and this dephosphorylation still occurred even when the Cdc2 feedback loops were blocked with butyrolactone-I. These data suggest that the DNA damage checkpoint in part acts through Thr14/Tyr15 phosphorylation by a mechanism independent of Cdc2 activity, and this phosphorylation can be accentuated by the Cdc2 feedback loops involving Thr14/Tyr15 protein kinases and phosphatases. The kinase activity of the Wee1Hu Tyr15 protein kinase was unaltered after DNA damage, but the phosphatase activity of Cdc25C was reduced. Thus, the decrease in Cdc25C activity may in part account for the DNA damage-induced increase in Thr14/Tyr15 phosphorylation.

A possible role of p73 on the modulation of p53 level through MDM2.

MDM2, one of the transcriptional targets of p53, can target p53 for degradation in a negative feedback loop. The p53-related protein p73, however, can bind to MDM2 but is not consequently down-regulated. Here we demonstrate that p73 could transactivate the MDM2 promoter in p53-null cell lines. In p53-null cell lines, the level of MDM2 was increased by p73 due to increases in transcription and protein stability of MDM2. In transient transfection assays, inhibition of the transcriptional activity of p73 required a higher amount of MDM2 than that of p53. This is probably due to the fact that MDM2 can target p53, but not p73, for degradation. We demonstrated further that the level of p53 could be altered by a cooperation between MDM2 and p73, but not by transcriptional inactive mutants of p73. Expression of p73 resulted in a reduction of the ectopically expressed p53 in transient transfections or of the endogenous p53 induced by Adriamycin- or UV-mediated damage. These reductions of p53 were likely to be due to an increase in MDM2-mediated proteolysis. These results suggest the possibility that different levels of p73 in the cell may act as a mechanism to modulate p53 responses after DNA damage and other stresses and that an increase rather than a decrease in p73 may play a role in tumorigenesis.

MDM2 and MDMX inhibit the transcriptional activity of ectopically expressed SMAD proteins.

Transforming growth factor-beta (TGF-beta) inhibits cell proliferation in many cell types, and acquisition of TGF-beta resistance has been linked to tumorigenesis. One class of proteins that plays a key role in the TGF-beta signal transduction pathway is the SMAD protein family. MDM2, a key negative regulator of p53, has recently been shown to suppress TGF-beta-induced growth arrest in a p53-independent manner. Here we show that MDM2 and the structurally related protein MDMX can inhibit the transcriptional activity of ectopically expressed SMAD1, SMAD2, SMAD3, and SMAD4. Immunofluorescence staining indicated that ectopically expressed SMAD4 was present in both the cytoplasm and nucleus, and MDM2 and NIDMX were localized mainly to the nucleus and cytoplasm, respectively. When SMAD4 was coexpressed with either MDM2 or MDMX, nuclear accumulation of SMAD4 was strikingly inhibited. We have no evidence that SMAD4 binds directly to MDM2 or MDMX; hence, the inactivation and nuclear exclusion of SMAD4 by MDM2/MDMX may involve other indirect mechanisms.

Expression of a novel form of p21Cip1/Waf1 in UV-irradiated and transformed cells.

The tumor suppressor p53 and its target the CDK inhibitor p21 (Cip1/Waf1) are key components of the cellular response to DNA damage. Insight into how p21 is regulated in normal cells, and how it may be deregulated in tumor cells is important for the understanding of tumorigenesis. p21 was induced in normal human diploid fibroblasts after UV irradiation-induced DNA damage, but, at a high dose of UV irradiation, a faster mobility form of p21 on SDS-PAGE (designated p21delta) was expressed. Surprisingly, in a variety of growing transformed cell lines, the level of p21 was low but p21delta was prominent. We found that p21delta appeared to be derived through a loss of around 10 amino acids from the C-terminus of p21, which theoretically would remove the PCNA binding domain, a second cyclin binding domain and the nuclear localization signal sequence. Several characteristics distinguish p21 from p21delta. Both the full length p21 and p21delta could be stabilized by a proteasome inhibitor, but only the full length p21 was associated with Cdk2 and PCNA. Consistent with this, gel filtration chromatography revealed that all the full length p21 in the cell was complexed to other proteins, whereas a significant portion of p21delta was in monomeric form. Moreover, p21 was mainly localized to the nucleus, but p21delta was mainly localized to the cytoplasm. We propose that the decrease in p21 and increase in p21delta could contribute to the deregulation of the cell cycle, and could be a mechanism involved in cellular transformation.

Involvement of p21 and p27 in the regulation of CDK activity and cell cycle progression in the regenerating liver.

In tissue culture systems, p21 and p27 inhibit cyclin-dependent kinase (CDK) activity and cell cycle progression in response to numerous stimuli, but little is known about their involvement in cell growth in vivo. We examined the modulation of CDK activity by these proteins after 70% partial hepatectomy (PH), an in vivo model of synchronous hepatocyte cell cycle progression. After PH in BALB/c mice, p21 was induced during the prereplicative (G1) phase and was maximally expressed after peak hepatocyte DNA synthesis. p27 was present in quiescent liver and was minimally induced after PH. p21 and p27 immunoprecipitated with CDK2, CDK4, and cyclin D1 in the regenerating liver. The activity of CDK2-, CDK4- and cyclin D1-associated kinases was upregulated after PH, and maximal activity of these enzyme complexes corresponded to peak DNA synthesis. Immunodepletion experiments suggested that p27 plays a role in downregulating CDK2 activity before and after peak DNA synthesis. Compared to cogenic wild-type mice, p21-/- mice demonstrated evidence of markedly accelerated hepatocyte progression through G1 phase after PH: DNA synthesis, upregulation of cyclin A and PCNA, induction of cyclin D1- and CDK2-associated kinase activity, and appearance of a phosphorylated retinoblastoma protein (Rb) species occurred earlier in the p21-/- mice. These results suggest that p21 and p27 modulate CDK activity in the regenerating liver, and that p21 regulates the rate of progression through G1 phase of the cell cycle in vivo.

G1 versus G2 cell cycle arrest after adriamycin-induced damage in mouse Swiss3T3 cells.

Cell cycle arrest after different types of DNA damage can occur in either G1 phase or G2 phase of the cell cycle, involving the distinct mechanisms of p53/p21(Cip1/Waf1) induction, and phosphorylation of Cdc2, respectively. Treatment of asynchronously growing Swiss3T3 cells with the chemotherapeutic drug adriamycin induced a predominantly G2 cell cycle arrest. Here we investigate why Swiss3T3 cells were arrested in G2 phase and not in G1 phase after adriamycin-induced damage. We show that adriamycin was capable of inducing a G1 cell cycle arrest, both during the G0-G1 transition and during the G1 phase of the normal cell cycle. In G0 cells, adriamycin induced a prolonged cell cycle arrest. However, adriamycin caused only a transient cell cycle delay when added to cells at later time points during G0-G1 transition or at the G1 phase of normal cell cycle. The G1 arrest correlated with the induction of p53 and p21(Cip1/Waf1), and the exit from the arrest correlated with the decline of their expression. In contrast to the G1 arrest, adriamycin-induced G2 arrest was relatively tight and correlated with the Thr-14/Tyr-15 phosphorylation of cyclin B-Cdc2 complexes. The relative stringency of the G1 versus G2 cell cycle arrest may explain the predominance of G2 arrest after adriamycin treatment in mammalian cells.

MDM2 and MDMX can interact differently with ARF and members of the p53 family.

Members of the p53 family of transcription factors have essential roles in tumor suppression and in development. MDM2 is an essential regulator of p53 that can inhibit the transcriptional activity of p53, shuttle p53 out of the nucleus, and target p53 for ubiquitination-mediated degradation. Little is known about the interaction and selectivity of different members of the p53 family (p53, p63, and p73) and the MDM2 family (MDM2 and MDMX). Here we show that the transcriptional activities of p53 and p73, but not that of p63, were inhibited by both MDM2 and MDMX. Consistent with these, we found that MDMX can physically interact with p53 and p73, but not with p63. Moreover, ectopically expressed MDM2 and MDMX could induce alterations in the subcellular localization of p73, but did not affect the subcellular localization of p53 and p63. Finally, we demonstrate that while ARF can interact with MDM2 and inhibit the regulation of p53 by MDM2, no interaction was found between ARF and MDMX. These data reveal that significant differences and selectivity exist between the regulation of different members of the p53 family by MDM2 and MDMX.

Characterization of the cullin and F-box protein partner Skp1.

Skp1 interacts with cullins, F-box containing proteins, and forms a complex with cyclin A-Cdk2 in mammalian cells. Skp1 is also involved in diverse biological processes like degradation of key cell cycle regulators, glucose sensing, and kinetochore function. However, little is known about the structure and exact function of Skp1. Here we characterized the interaction between Skp1 and the F-box protein Skp2. We show that Skp1 can bind to Skp2 in vitro using recombinant proteins, and in vivo using the yeast two-hybrid system. Deletion analysis of Skp1 indicated that most of the Skp1 protein is required for binding to Skp2. In mammalian cell extracts, a large portion of Skp1 appears to associate with proteins other than Skp2. Biochemical analysis indicated that Skp1 is likely to be a flexible, non-spherical protein, and is capable of forming dimers.

On the masking of signals on immunoblots by cellular proteins.

When determining the concentration of a particular protein in a cell extract, or when comparing the amount of a protein in different samples, it is a common practice to use specific antibodies in immunoblotting to compare the samples side by side with known amounts of purified protein. Here we show that with many antibodies, in particular monoclonal antibodies, the sensitivity of detecting the cognate antigen on immunoblots can be significantly reduced when the antigen is in a mixture with other cellular proteins. The signals on the immunoblots are masked by other endogenous proteins in the cell lysate, making the amount of the protein on the immunoblot appear to be less than the actual amount, thus invalidating direct comparison with purified protein.

p53 deficiency enhances mitotic arrest and slippage induced by pharmacological inhibition of Aurora kinases.

A number of small-molecule inhibitors of Aurora kinases have been developed and are undergoing clinical trials for anti-cancer therapies. Different Aurora kinases, however, behave as very different targets: while inhibition of Aurora A (AURKA) induces a delay in mitotic exit, inhibition of Aurora B (AURKB) triggers mitotic slippage. Furthermore, while it is evident that p53 is regulated by Aurora kinase-dependent phosphorylation, how p53 may in turn regulate Aurora kinases remains mysterious. To address these issues, isogenic p53-containing and -negative cells were exposed to classic inhibitors that target both AURKA and AURKB (Alisertib and ZM447439), as well as to new generation of inhibitors that target AURKA (MK-5108), AURKB (Barasertib) individually. The fate of individual cells was then tracked with time-lapse microscopy. Remarkably, loss of p53, either by gene disruption or small interfering RNA-mediated depletion, sensitized cells to inhibition of both AURKA and AURKB, promoting mitotic arrest and slippage respectively. As the p53-dependent post-mitotic checkpoint is also important for preventing genome reduplication after mitotic slippage, these studies indicate that the loss of p53 in cancer cells represents a major opportunity for anti-cancer drugs targeting the Aurora kinases.

Salt-inducible kinase 3 is a novel mitotic regulator and a target for enhancing antimitotic therapeutic-mediated cell death.

Many mitotic kinases are both critical for maintaining genome stability and are important targets for anticancer therapies. We provide evidence that SIK3 (salt-inducible kinase 3), an AMP-activated protein kinase-related kinase, is important for mitosis to occur properly in mammalian cells. Downregulation of SIK3 resulted in an extension of mitosis in both mouse and human cells but did not affect the DNA damage checkpoint. Time-lapse microscopy and other approaches indicated that mitotic exit but not mitotic entry was delayed. Although repression of SIK3 alone simply delayed mitotic exit, it was able to sensitize cells to various antimitotic chemicals. Both mitotic arrest and cell death caused by spindle poisons were enhanced after SIK3 depletion. Likewise, the antimitotic effects due to pharmacological inhibition of mitotic kinases including Aurora A, Aurora B, and polo-like kinase 1 were enhanced in the absence of SIK3. Finally, in addition to promoting the sensitivity of a small-molecule inhibitor of the mitotic kinesin Eg5, SIK3 depletion was able to overcome cells that developed drug resistance. These results establish the importance of SIK3 as a mitotic regulator and underscore the potential of SIK3 as a druggable antimitotic target.

The CDK1 inhibitory kinase MYT1 in DNA damage checkpoint recovery.

Inhibition of cyclin-dependent kinase 1 (CDK1) by phosphorylation is a key regulatory mechanism for both the unperturbed cell cycle and the DNA damage checkpoint. Although both WEE1 and MYT1 can phosphorylate CDK1, little is known about the contribution of MYT1. We found that in contrast to WEE1, MYT1 was not important for the normal cell cycle or checkpoint activation. Time-lapse microscopy indicated that MYT1 did, however, have a rate-determining role during checkpoint recovery. Depletion of MYT1 induced precocious mitotic entry when the checkpoint was abrogated with inhibitors of either CHK1 or WEE1, indicating that MYT1 contributes to checkpoint recovery independently of WEE1. The acceleration of checkpoint recovery in MYT1-depleted cells was due to a lowering of threshold for CDK1 activation. The kinase activity of MYT1 was high during checkpoint activation and reduced during checkpoint recovery. Importantly, although depletion of MYT1 alone did not affect long-term cell growth, it potentiated with DNA damage to inhibit cell growth in clonogenic survival and tumor xenograft models. These results reveal the functions of MYT1 in checkpoint recovery and highlight the potential of MYT1 as a target for anti-cancer therapies.