A C-terminal Fragment of Cyclin E, Generated by Caspase-mediated Cleavage, Is Degraded in the Absence of a Recognizable Phosphodegron*

We have previously shown that caspase-mediated cleavage of Cyclin E generates p18-Cyclin E in hematopoietic tumor cells. Its expression can induce apoptosis or sensitize to apoptotic stimuli in many cell types. However, p18-cyclin E has a much shorter half-life than Cyclin E, being more effectively ubiquitinated and degraded by the 26 S proteasome. A two-step process has emerged that regulates accelerated degradation of Cyclin E, with a caspase-mediated cleavage followed by enhanced proteasome-mediated degradation. We show that recognition of p18-Cyclin E by the Skp1-Cul1-Fbw7 (SCF) complex and its interaction with the Fbw7 protein isoforms can take place independently of phosphorylation of p18-Cyclin E at a C-terminal phosphodegron. In addition to the SCFFbw7 pathway, Ku70 binding that facilitates Hdm2 recruitment may also be implicated in p18-Cyclin E ubiquitination. Blocking p18-Cyclin E degradation with proteasome inhibitors increases levels of p18-Cyclin E and enhances its association with Ku70, thus leading to Bax release, its activation, and apoptosis. Moreover, cells expressing p18-Cyclin E are more sensitive to treatment with proteasome inhibitors, such as Bortezomib. By preventing its proteasomal degradation, p18-Cyclin E, but not Cyclin E, may become an effective therapeutic target for Bortezomib and apoptotic effectors in hematopoietic malignancies.

Deregulation of cellular proliferation, together with defects in apoptosis control, are two critical processes responsible for tumor formation (1). Cyclins, in association with their catalytic subunits the cyclin-dependent kinases (Cdks), 2 control cell cycle progression by regulating events that drive the transition between cell cycle phases (2)(3)(4). The Cyclin E-Cdk2 complex plays an essential and rate-limiting role in the transition between the G 1 and S phases of the cell cycle (5,6) and in the initiation of DNA replication (5,7). Regulation of Cyclin E takes place at both transcriptional and post-translational levels. Transcriptionally, Cyclin E is a target of the E2F transcription factor family, whereas post-translationally, Cyclin E is targeted by the ubiquitin-proteasomal pathway to 26S proteasomal degradation (8,9). The required ubiquitination is achieved following recognition by and interaction with the E3 ubiquitin ligase, the F-box tumor suppressor protein, Fbw7 (also known as hCdc4 or Sel10) (10). Fbw7 is a specificity factor for the Skp1-Cul1-F-box protein ubiquitin ligase (SCF) complex that targets a variety of proto-oncogene products for ubiquitin-mediated degradation, including Cyclin E.
The ubiquitin-proteasome pathway is essential for the degradation of the majority of intracellular proteins (11). A number of key regulatory proteins involved in cell proliferation, differentiation, survival, as well as cell death are regulated by proteasome-mediated proteolysis, which results in the activation or inhibition of specific cell signaling pathways (12). Apoptosis is regulated by the ubiquitin/proteasome system at different levels and importantly, inhibition of proteasome function may induce cell death. Recently, inhibition of the proteasomal function with specific inhibitors has evolved as a novel approach for the treatment of various cancers (13,14). Highly specific inhibitors of the 26S proteasome, such as the reversible inhibitor MG132, have been shown to induce apoptosis in many tumor cell lines (14,15). Bortezomib (PS-341; Velcade), a modified dipeptydil boronic acid, is a potent and selective inhibitor of the 26S proteasome, a multisubunit protein complex responsible for the regulation of the turnover of cellular proteins, thus maintaining cellular homeostasis. It controls protein expression and function by degradation of ubiquitinated proteins as well as cleanses the cell of abnormal or misfolded proteins. Bortezomib inhibits the proteasome by binding reversibly to the chymotrypsin-like site in the 20S core of the proteasome. The proteolytic inhibition disrupts homeostasis, leading to cell death (16).
Cyclin E has been shown to be critical for genotoxic stressinduced apoptosis of tumor cells of hematopoietic origin (17). An 18-kDa Cyclin E fragment (p18-Cyclin E) is generated through caspase-mediated proteolytic cleavage of Cyclin E (18). p18-Cyclin E lacks the Cdk2 interaction domain as well as the N-terminal phosphodegron known to be also employed for ubiquitination and degradation of Cyclin E. However, the presence of a single C-terminal phosphodegron is known to be sufficient for Cyclin E degradation. In fact, other Fbw7 substrates (e.g. Myc, Notch) have a single phosphodegron (19). However, the presence of the main C-terminal phosphodegron comprised of the Ser 372 , Thr 380 , and Ser 384 residues would suggest a similar regulatory mechanism for p18-Cyclin E and Cyclin E by the SCF complex.
Here we examine the recognition of p18-Cyclin E by SCF Fbw7 in conjunction with phosphorylation at the Cdc4 (Fbw7) phosphodegron (20) or binding to Cdk2, as well as the contribution of Cul1, and the C terminus of Hsp70-interacting protein (CHIP) to p18-Cyclin E regulation. Our findings indicate a role for Fbw7 that is independent of Cdk2 binding or phosphorylation at Cdc4 phosphodegron. We have recently identified Ku70, known previously for its critical role in nonhomologous end joining DNA repair, as a specific p18-Cyclin E-interacting protein (21). In human umbilical vein endothelial cells, 3 Ku70 was recently found to associate with Hdm2, known previously for its role in regulating the p53 tumor suppressor protein via ubiquitination (22,23). Therefore, we investigated the possible role of Ku70 binding to p18-Cyclin E in its stability, through recruitment of Hdm2. Finally, we tested whether enhancing p18-Cyclin E stability using proteasome inhibitors may potentiate its apoptotic effect, because the half-life of p18-Cyclin E is much shorter than that of Cyclin E. Indeed, blocking its degradation caused a massive elevation of its levels as well as its association with Ku70, leading to Bax activation and enhanced apoptosis. Cells expressing p18-Cyclin E become more sensitive to apoptotic triggers, particularly when its degradation is prevented by proteasome inhibition. These results reveal a two-step process for Cyclin E degradation, with the second step being exquisitely sensitive to proteasome inhibition, possibly identifying it as a novel therapeutic target.

EXPERIMENTAL PROCEDURES
Cells and Treatments-The IM-9 B cell-derived tumor cell line was maintained in RPMI medium with 10% fetal bovine serum (Hyclone, Logan, UT), L-glutamine, and 100 units/ml penicillin-streptomycin. Human embryonic kidney (HEK) epithelial 293T and human non-small cell lung carcinoma NCI-H1299 cells were maintained as above except in Dulbecco's modified Eagle's medium. The cells ϳ75% confluent were irradiated with 5 Gy ( 137 Cs source; fixed dose rate of 2.0 Gy/min) (24). Bortezomib (Velcade) was from Millenium Pharmaceuticals (Cambridge, MA), MG132 and cycloheximide were from Sigma. For the protein half-life study, the cells were treated with 10 M cycloheximide with or without a 2-h preincubation with 100 nM Bortezomib and collected at the indicated times after treatment. HEK293T and H1299 cells stably expressing HA-p18-Cyclin E were generated by infecting cells with a lentivirus carrying HA-p18-Cyclin E as well as enhanced GFP separated by an IRES2 sequence under the control of an EF1␣ promoter.
For determining the ubiquitination status, the cells were treated with 25 M MG132 prior to harvesting and lysed in the same 1% Nonidet P-40 lysis buffer as described above also containing 1ϫ HALT phosphatase inhibitor mixture (Pierce), 5 M MG132, and 5 M N-ethylmaleimide (Sigma). Immunoprecipitation was performed for 3 h at 4°C with HA or FLAG antibody (Sigma) for tagged p18-Cyclin E, with HE-111 (Santa Cruz Biotechnology) for Cyclin E followed by incubation with protein A plus G-agarose beads for 1 h. Immunoblotting was performed with either HA antibody, ubiquitin mouse monoclonal antibody (Chemicon, Temecula, CA), or rabbit polyclonal antibody (Dako, Carpinteria, CA).
Transfections-H1299 cells were transfected with 10 g of p18-Cyclin E-HA and 3 g of FLAG-Fbw7␣, ␤, or ␥. At 24 h post-transfection, the cells were collected and lysed. Immunoprecipitation was performed with anti-HA antibody followed by immunoblotting with anti-FLAG antibody. HEK293T cells were transfected with FLAG-Fbw7␣ and p18-Cyclin E-HA, T380A-p18-Cyclin E-HA, Cyclin E-GFP, or T380A-Cyclin E-GFP. Immunoprecipitation was performed with anti-HA or anti-GFP antibodies followed by immunoblotting with anti-FLAG antibody. For the overexpression of FLAG-tagged Fbw7␣, ␤, ␥, dominant negative Cul1 (25), 3 g of construct and Lipofectamine 2000 (Invitrogen) were used according to the protocol provided by the manufacturer. Knockdown of Cdk2 and Fbw7 was performed using 75-100 nM of siRNA (Smart-Pool; Dharmacon, Lafayette, CO). As a negative control siCONTROL nontargeting siRNA pool #1 (Dharmacon) was used. In both the overexpression and knockdown experiments, cells were collected after 48 h. To effectively knock down Ku70, HEK293T cells were transfected with 100 nM siKu70 or siCON-TROL (Dharmacon). At 24 h post-transfection, the cells were replated followed by a second transfection the following day, with the cells being examined at 72 h after the initial transfection. Overexpression of CHIP (kind gift form Dr. E. Ficker, Metrohealth, Cleveland, OH) was achieved by transfection of an expression construct followed by collection after 24 h. CHIP was knocked down using siRNA (SmartPool, Dharmacon). The transfected cells were collected after 24 and 48 h, respectively.

p18-Cyclin E Is a Short-lived Protein That Is Stabilized by Proteasome Inhibition-Our previous experiments have shown that
Cyclin E is proteolytically cleaved during genotoxic stress, such as that induced by ionizing radiation, generating an 18-kDa C-terminal fragment, p18-Cyclin E (21). To examine the stability of p18-Cyclin E compared with full-length Cyclin E (Cyclin E), the half-life of these two proteins was determined following protein translation inhibition with cycloheximide. The results of these experiments indicate that p18-Cyclin E had a remarkably short half-life of less than 0.5 h. In contrast, the half-life of Cyclin E was longer than 8 h in HEK293T cells stably expressing p18-Cyclin E (Fig. 1A). Furthermore, when protein degradation was inhibited by the proteasome inhibitor Bortezomib, prior to cycloheximide-induced translation inhibition, levels of p18-Cyclin E increased considerably more than those of Cyclin E (Fig. 1B). As expected, following cycloheximide treatment, the half-life life of Cyclin E did not change. In the case of p18-Cyclin E, although there seems to be a substantial decrease in its levels in the first 0.5 h, its degradation was diminished after 2 h, with p18-Cyclin E levels decreasing only ϳ50% by 24 h. In contrast, in the absence of Bortezomib, there was no visible p18-Cyclin E present at 1 h after cycloheximide addition (Fig. 1C). These data indicate that the sustained levels of p18-Cyclin E are a result of its increased levels as well as a decreased degradation rate over time. It is important to note that these cells stably express p18-Cyclin E at low, nontoxic levels. Moreover, Cyclin E cleavage, which would generate additional endogenous p18-Cyclin E, does not take place in these cells.
Strikingly, proteasome inhibition achieved by treatment with either MG132 or the new clinical compound Bortezomib led to a dramatic increase in p18-Cyclin E levels, indicating that this protein is very rapidly turned over through proteasome-mediated degradation. p18-Cyclin E levels increased in both HEK293T and NCI-H1299 cells stably expressing p18-Cyclin E (Fig. 1D). Interestingly, the stabilizing effect of proteasome inhibition on Cyclin E was much less robust than that on p18-Cyclin E.
p18-Cyclin E Is Effectively Ubiquitinated-The above experiments revealed that proteasome inhibition led to a massive increase in the levels of p18-Cyclin E. Polyubiquitination of the target protein is the most common signal for proteasome-mediated degradation. The higher molecular weight, modified forms of p18-Cyclin E were readily visible upon MG132 treatment in the presence of N-ethylmaleimide, an inhibitor of the ubiquitin isopeptidases, indicating that p18-Cyclin E is degraded through the addition of multiple ubiquitin or ubiquitin-like residues that target it for proteasomal degradation ( Fig. 2A). To determine whether p18-Cyclin E is modified by the addition of ubiquitin, FLAG-tagged p18-Cyclin E was cotransfected with an HA-ubiquitin expression construct in HEK293T cells. Immunoprecipitation with anti-FLAG antibody for p18-Cyclin E followed by immunoblotting with anti-HA antibody for ubiquitin shows the characteristic smear (Fig. 2B, top left panel) and the characteristic laddering (Fig. 2B, right panel) of polyubiquitinated p18-Cyclin E. Immunoprecipitation followed by immunoblotting with an anti-HA antibody for the ubiquitinated protein forms revealed that both Cyclin E and p18-Cyclin E were ubiquitinated (Fig. 2B, top left  panel). The addition of polyubiquitin chains was also evident when HA-p18-Cyclin E was immunoprecipitated from HEK293T cells stably expressing it, followed by immunoblotting with two different anti-ubiquitin antibodies (Fig. 2C).
Fbw7 Interacts with p18-Cyclin E Independently of Phosphorylation at the C-terminal Phosphodegron-Cyclin E is known to be degraded through the ubiquitin-proteasomal pathway that uses Fbw7 as a component of the SCF complex (27)(28)(29). There are three splice-derived isoforms of Fbw7, designated as ␣, ␤, and ␥, that localize to different cellular compartments (30).
To determine whether p18-Cyclin E, similar to Cyclin E, is also targeted for proteasomal degradation by the SCF Fbw7 complex, we examined the interaction between p18-Cyclin E and all three Fbw7 isoforms in HEK293T cells transiently transfected with HA-p18-Cyclin E and FLAG-Fbw7␣, ␤, and ␥. Immunoprecipitation with anti-HA to pull down p18-Cyclin E, followed by Western blot performed with an anti-FLAG antibody, indicates that p18-Cyclin E associated with Fbw7␣ and ␥, and to a lesser extent with Fbw7␤ (Fig. 3A). These data indicate that p18-Cyclin E is an ubiquitination target of Fbw7, which regulates its degradation.   C (lanes 2 and 3). NOVEMBER 7, 2008 • VOLUME 283 • NUMBER 45

JOURNAL OF BIOLOGICAL CHEMISTRY 30799
Phosphorylation is known to play an important role in the interaction between the Fbw7 isoforms and Cyclin E (27)(28)(29). To examine the influence of phosphorylation on the expression levels of p18-Cyclin E, the amino acid residues known to be critical for the phosphorylation-dependent turnover of Cyclin E (31) were mutated, alone or in combination. A schematic representation of these residues is shown in Fig. 3B. The following phosphorylation mutants were generated: T380A, S384A, T380A/S384A, S372A/T380A, and S372A/T380A/S384A. Unlike the effect that is expected for these point mutations on Cyclin E (31), transient transfection of these mutant p18-Cyclin E expression constructs has pointed out striking differences. The Thr 380 residue, which has been reported to play a crucial role in the turnover of Cyclin E (8,9), did not affect the levels of p18-Cyclin E. Moreover, the mutation of Ser 384 instead of increasing the levels of p18-Cyclin E had the opposite effect. Only mutation of Ser 372 , in combination with T380A, showed the expected increase in levels of the protein (Fig. 3C), with the triple mutant having an attenuated effect, most likely because of the destabilizing effect of the Ser 384 mutation (Fig. 3C). Strikingly, co-transfection of the HA-tagged triple phosphorylation mutant S372A/T380A/S384A-p18-Cyclin E with the Fbw7␣, ␤, and ␥ isoforms showed their interaction with p18-Cyclin E (Fig.  3D). These data indicate that p18-Cyclin E may interact with Fbw7␣, ␤, and ␥ without phosphorylation at Ser 372 , Thr 380 , or Ser 384 that constitute the C-terminal phosphodegron of Cyclin E. Although not required, phosphorylation at these sites may increase Fbw7 effectiveness. To confirm that p18-Cyclin E interacts with Fbw7 irrespective of its phosphorylation status at the C-terminal phosphodegron, we show that the T380A mutation that prevents phosphorylation does not influence the interaction between p18-Cyclin E and Fbw7␣ (Fig. 3E, bottom  panel). In stark contrast, this mutation dramatically decreases the association between Cyclin E and Fbw7␣ (Fig. 3E, top  panel).
p18-Cyclin E Is Regulated by the SCF Fbw7 Complex Independently of Cdk2-To examine the direct role of Fbw7 in p18-Cyclin E regulation, p18-Cyclin E-expressing HEK293T cells were transfected with the ␣, ␤, and ␥ isoforms of Fbw7. Fbw7 overexpression led to decreased levels of p18-Cyclin E and Cyclin E. Cyclin E is known to be regulated primarily through the Cul1 pathway (27,32,33). Expression of dominant negative Cul1 led to a limited change in p18-Cyclin E levels in 293T-p18-Cyclin E cells and to a larger extent to that of Cyclin E (Fig. 4A). Finally, to clearly establish the role of Fbw7 in the degradation of p18-Cyclin E, RNA interference was used to diminish Fbw7 levels in HEK293T cells stably expressing p18-Cyclin E. Immunoblotting with anti-HA antibody for p18-Cyclin E and anti-Cyclin E (HE-12) antibody for Cyclin E (Fig. 4B) at 48 h posttransfection indicated that their levels increased, establishing a regulatory role for Fbw7.
Cdk2 is known to be responsible for the phosphorylation of Ser 384 on Cyclin E when it is present in complex with Cdk2 (34). To assess whether Cdk2 affects the levels of p18-Cyclin E, Cdk2 was ablated by siRNA-mediated knockdown (Fig. 4B). Despite a robust depletion of Cdk2 levels, there was only a minimal increase in p18-Cyclin E levels. Cyclin E increased to some extent in the absence of Cdk2 but much less compared with when Fbw7 was depleted (Fig. 4B). p18-Cyclin E encompasses the 276 -395 amino acid domain of Cyclin E and thus is lacking the Cdk2-binding domain (35). To confirm that, indeed, Cyclin E binds Cdk2 whereas p18-Cyclin E has lost this ability, HA-p18-Cylin E was immunoprecipitated from HEK293T cells stably expressing it followed by immunoblotting for Cdk2 (Fig.  4C). Conversely, Cdk2 was immunoprecipitated from HEK293T expressing HA-Cyclin E or HA-p18-Cyclin E, respectively, followed by immunoblotting with HA antibody (Fig. 4D). As expected, Cdk2 was found in complex with Cyclin E, whereas p18-Cyclin E did not associate with Cdk2. These experiments, together with our previous report (18) indicate that p18-Cyclin E does not interact with Cdk2, whereas Cyclin E, as expected, does. Taken together, these data suggest that even though p18-Cyclin E does not associate with Cdk2, Fbw7 can recruit it to the SCF complex and promote its ubiquitination.
Ku70 Binding May Regulate p18-Cyclin E Stability through Hdm2 Recruitment-The above studies have shown that p18-Cyclin E does not bind to Cdk2. We have shown that it binds instead to Ku70 (21). To examine the possible role of Ku70 binding in the degradation of p18-Cyclin E, HEK293T cells were transfected with HA-p18-Cyclin E alone or in the presence of FLAG-Ku70. This led to a dramatic reduction in the levels of p18-Cyclin E (Fig. 5A). To compare this change with the effect of Fbw7, HA-p18-Cyclin E was co-transfected with Fbw7␣, ␤, and ␥. Fbw7 overexpression led to a limited decrease in the levels of p18-Cyclin E as compared with that induced by Ku70, indicating that other E3 ligases may be also important (Fig. 5A). To clearly establish the role of endogenous Ku70 in p18-Cyclin E regulation, its levels were substantially depleted by specific siRNA oligonucleotides. Down-regulation of Ku70 by siRNA led to an increase in p18-Cyclin E levels, a change that could not be observed for Cyclin E (Fig. 5B).
Hdm2 is an ubiquitin E3 ligase known to be transcriptionally induced by p53 following DNA damage (22,23). Its role is to polyubiquitinate p53 and thus targets it to proteasomal degradation, therefore limiting its effect. The newly discovered role of Hdm2 in Ku70 regulation 3 raised the question of whether Hdm2 interacts with Ku70 in our experimental system. In IM-9 cells this association was clearly induced following irradiation, corresponding to increased p53 levels, known to lead to transcriptional activation of Hdm2 (Fig. 5C, top panel). Hdm2 was also immunoprecipitated from control or etoposide-treated HEK293T cells, in which the presence of Ku70 was determined by immunoblotting. Ku70 and Hdm2 interacted in both control and etoposide-treated cells (Fig. 5C, bottom panel), because p53 is inactive in these cells, and therefore the levels of Hdm2 are not expected to be induced by DNA-damaging agents.
CHIP is an E3 ligase that partners with chaperones Hsp70, Hsc70, and Hsp90, which target the degradation of misfolded proteins. CHIP has also been shown to participate in the proteasome-mediated degradation of base excision repair proteins, such as XRCC1 and Pol-␤ that have not been recruited to a DNA repair complex (36). We wanted to examine whether CHIP contributes to the degradation of p18-Cyclin E, because, as a truncated protein, it could be misfolded. Moreover, p18-Cyclin E is also associated with a pool of Ku70 that is cytoplasmic and therefore not engaged in the repair of DNA damage and thus could be subjected to the same attrition process as the base excision repair proteins that are in excess, because they are not utilized in DNA repair (36). The levels of p18-Cyclin E determined at 48 h post-transfection of CHIP in HEK293T cells stably expressing HA-p18-Cyclin did not decrease (Fig. 5D). Moreover, knockdown of CHIP by siRNA did not increase the levels of p18-Cyclin E (Fig. 5E). Taken together, these data strongly implicate Ku70 and Hdm2 but not CHIP in a degradation pathway of p18-Cyclin E, the exact role of which and its detailed mechanism will be further explored.
p18-Cyclin E Sensitizes Cells to Bortezomib Treatment through Enhanced Association with Ku70 and Release of Bax from Ku70-Of the commonly used proteasome inhibitors tested, Bortezomib was most effective in stabilizing p18-Cyclin E (data not shown). To examine the dynamics of proteasome inhibition and onset of apoptosis, stably expressing 293T-p18-Cyclin E cells were treated with 100 nM Bortezomib for various times, up to 48 h. Apoptosis, as measured by PARP-1 cleavage, was detectable as early as 8 h, with ϳ10% of PARP-1 being cleaved by 24 h and 50% by 48 h. The levels of p18-Cyclin E were increased in the presence of Bortezomib, because of the inhibition of its proteasomal degradation. p18-Cyclin E levels reached a maximum as early as 2 h, whereas Cyclin E levels did not vary significantly during the same time interval (Fig. 6A).
Because proteasomal degradation is responsible for its very short half-life, enhancing p18-Cyclin E stability may potentiate its apoptotic effect. To determine whether the levels of p18-Cyclin E can determine the response of cells to Bortezomib treatment, HEK293T cells expressing stably either GFP or p18-Cyclin E were examined at 24 h following treatment with increasing concentrations of Bortezomib. Increased PARP-1 cleavage in p18-Cyclin E-expressing cells suggested that p18- A, HEK293T cells were transfected with FLAG-Ku70 and HA-p18-Cyclin E alone or in combination with Fbw7␣, ␤, or ␥ and Ku70 in presence of HA-p18-Cyclin E. The levels of Ku70, Fbw7␣, ␤, and ␥, p18-Cyclin E, and ␤-actin were determined after 24 h post-transfection by Western blotting. B, HEK293T cells stably expressing HA-p18-Cyclin E were transfected with 100 nM siKu70 or siCONTROL. At 48 h post-transfection, a second round of transfection was performed to achieve a satisfactory down-regulation of Ku70. The levels of Ku70, p18-Cyclin E, and ␤-actin used as a loading control were determined by Western blotting at 72 h after the initial transfection. C, IM-9 cells were irradiated with 4 Gy, and Hdm2 was immunoprecipitation after 8 h. The levels of Ku70, Hdm2, and p53 were determined by immunoblotting. Normal mouse IgG was used as a negative control (top panel). HEK293T cells were treated with VP16 for 2 h and Hdm2 was immunoprecipitated (IP) followed by determination by Western blotting of the associated Ku70 levels. D, HEK293T stably expressing HA-p18-Cyclin E cells were transfected with CHIP or empty vector (C) and cells were collected after 48 h. The levels of CHIP, p18-Cyclin E and ␤-actin were determined by Western blotting. E, HEK293T stably expressing HA-p18-Cyclin E were transfected with 100 nM siCHIP or siControl, and the cells were collected after 24 or 48 h. The levels of CHIP, p18-Cyclin E, and ␤-actin were determined by Western blotting.  NOVEMBER 7, 2008 • VOLUME 283 • NUMBER 45

JOURNAL OF BIOLOGICAL CHEMISTRY 30801
Cyclin E may be the mediator of Bortezomib-induced cell death (Fig. 6B). To determine the mechanism responsible for the observed cell sensitivity, we examined the interaction between p18-Cyclin E and Ku70 before and after treatment with 50 nM Bortezomib or irradiation. At 16 h after treatment, HA-p18-Cyclin E was immunoprecipitated with anti-HA antibody followed by immunoblotting for Ku70. Treatment with Bortezomib but not irradiation caused an elevation of the levels of Cyclin E and to an even larger extent those of p18-Cyclin E (Fig.  6C). Increased p18-Cyclin E levels led to an increase in its association with Ku70 (Fig. 6D). To determine whether this association leads to the release of Bax from Ku70, Bax was immunoprecipitated followed by immunoblotting for Ku70. Elevated levels of p18-Cyclin E bound more Ku70, resulting in the release of Bax (Fig. 6E), which correlated with the induction of cell death as indicated by cleavage of PARP-1 (Fig. 6C). Bortezomib treatment also generated endogenous p18-Cyclin E in hematopoietic IM-9 cells at a much higher level compared with treatment with irradiation alone (Fig. 6F). We have shown previously that p18-Cyclin E binds Ku70 leading to release of Bax following DNA damage and its activation (21), an observation that we have now extended to that caused by proteasome inhibition leading to increased levels of p18-Cyclin E. Taken together, these data indicate that by preventing its proteasomal degradation, p18-Cyclin E may become an effective therapeutic target for proteasome inhibitors, such as Bortezomib.

DISCUSSION
The expression of Cyclin E, the master regulatory protein of the cell cycle transition from the G 1 to the S phase, is controlled at the transcriptional as well as post-translational level. Cyclin E is regulated post-translationally by the 26 S proteasome following its ubiquitination by the SCF complex. The Cdk2-bound Cyclin E requires phosphorylation at a conserved phosphodegron to be recognized and efficiently ubiquitinated by the Skp1-Cul1-Fbw7 complex (37). Using knockin mice, it has been recently shown that Cyclin E phosphorylation regulates cell proliferation in hematopoietic and epithelial lineages in vivo (38). Unlike Cyclin E, p18-Cyclin E cannot interact with Cdk2 because it is missing the Cdk2-binding domain.
We have previously shown that Cyclin E is regulated by genotoxic stress caused by ionizing radiation or other chemotherapeutic drugs, which plays a functional role in apoptosis of hematopoietic cells (17). Caspase-3 mediates cleavage of Cyclin E that results in generation of its C-terminal proteolytic fragment, p18-Cyclin E in all hematopoietic tumor cells examined, where most if not all Cyclin E gets converted to its truncated form. Protein synthesis inhibition studies with cycloheximide in HEK293T cells stably expressing p18-Cyclin E have now determined a 16-fold shorter half-life for p18-Cyclin E as compared with Cyclin E. This finding indicates that after being generated, p18-Cyclin E is degraded rapidly through the proteasomal pathway, thus implicating a two-step process in its accelerated turnover. Indeed, the addition of proteasome inhibitors, such as MG132 and Bortezomib, significantly inhibited the degradation of p18-Cyclin E, indicating that p18-Cyclin E is a specific target of the ubiquitin-proteasomal degradation pathway that, on the other hand, has a limited effect on Cyclin E levels. The higher molecular weight ubiquitinated forms of p18-Cyclin E are detectable using proteasome inhibitors, suggesting the susceptibility of these ubiquitinated forms to degradation.
Degradation of p18-Cyclin E is dependent on ubiquitination by the SCF, as indicated by its interaction with the different isoforms of Fbw7, an association directly responsible for the ubiquitin-dependent proteolysis of a number of proto-oncogenes, such as Aurora A, Cyclin E, c-Jun, c-Myc, and Notch (12). Strikingly, we find that the interaction between Fbw7 and p18-Cyclin E takes place even in the absence of its binding to Cdk2 and consequently independently of its phosphorylation at three key residues at the well characterized C-terminal phosphodegron of Cyclin E. These findings may not be unique to Fbw7 and may resemble those reported for another SCF E3 ligase, ␤-TrCP. Although it has been clearly established that ␤-TrCP recognizes a doubly phosphorylated DSG motif, it may also recognize a nonphosphorylated DDG-like motif, such as that found in Cdc25A and Cdc25B (39).
We found that p18-Cyclin E interaction is stronger with the ␣ and ␥ isoforms of Fbw7 and weaker with the ␤ isoform, an observation consistent with an earlier report for Cyclin E (40). It was recently found that the nucleoplasmic Fbw7␣ isoform accounts for almost all Fbw7 activity toward Cyclin E (41). Nevertheless, this result also indicates that when overexpressed, p18-Cyclin E localizes to both the nuclear, nucleolar, and cytosolic compartments. It should be noted that there is a possibility that when p18-Cyclin E is generated in cells following cleavage of Cyclin E during apoptosis, it may be phosphorylated at this C-terminal phosphodegron, whereas these phosphorylation events do not take place in cells stably expressing p18-Cyclin E. The observed difference in degradation efficiency by Cul1 suggests a different regulatory mechanism for p18-Cyclin E than that for full-length Cyclin E, which may involve a parallel pathway independent of Cul1.
We have recently identified a specific p18-Cyclin E-interacting protein, Ku70, known as a critical component of the nonhomologous end joining DNA repair pathway (21). This novel interaction raised the possibility of p18-Cyclin E being targeted to degradation by a pathway involving Ku70. Indeed, overexpression of Ku70 led to a decrease in p18-Cyclin E levels, whereas knockdown of Ku70 by siRNA stabilized the protein. Surprisingly, we have identified Ku70 in a complex with the E3 ubiquitin ligase Hdm2. Taken together, these data suggest a novel degradation mechanism for p18-Cyclin E involving its recruitment, together with Hdm2, to Ku70. But how does Ku70 regulate p18-Cyclin E levels? It is possible that Ku70 acts as a scaffold bringing Hdm2 and p18-Cyclin E together, thus facilitating its ubiquitination and degradation. This may suggest a two-step degradation process of Cyclin E to achieve its accelerated turnover. First, once apoptosis is triggered, Cyclin E is proteolytically cleaved by caspase-3 to avoid unnecessary cell replication and proliferation that would otherwise be sustained by Cyclin E. Then, in a second step of degradation p18-Cyclin becomes more susceptible to proteasome-mediated degradation, a process that is responsible for its accelerated turnover. These observations provide new insights into the regulation of Cyclin E. A similar accelerated proteasome-mediated degrada-tion pathway may regulate turnover of other cellular proteins, such as those created by limited proteolytic cleavage of several hundreds of caspase-3 substrates (42).
We have reported earlier that in hematopoietic cells, following treatment with DNA-damaging agents, such as irradiation and VP16, p18-Cyclin E is generated through proteolytic cleavage of Cyclin E (18,21). By binding to Ku70, p18-Cyclin E causes the dissociation of Bax from Ku70 and its activation, thus leading to apoptosis (21). Therefore, p18-Cyclin E may be an attractive target for cancer therapy. However, the scope of its clinical use is limited because of its labile nature, and therefore enhancing its stability may potentiate its effect. Previously, we have reported that p18-Cyclin E-induced cell death is due to the interaction of Ku70 with p18-Cyclin E, which influences the interaction of Bax with Ku70 and, thereby, activation of Bax. The association of p18-Cyclin E with Ku70 is also enhanced by Bortezomib treatment, most probably because of the increased levels of p18-Cyclin E, as well as counteracting the destabilizing effect of Ku70 on p18-Cyclin E. Importantly, this interaction leads to release of Bax from Ku70, followed by its activation. Cells expressing p18-Cyclin E become more sensitive to cell death stimuli when treated with Bortezomib, as compared with the parental cells, most likely because of the inhibition of p18-Cyclin E proteasomal degradation. Moreover, the combination of Bortezomib with a DNA-damaging agent such as VP16 is much more effective in elevating p18-Cyclin E levels and inducing apoptosis (data not shown). The increased cell death in Bortezomib-treated cells is most likely due to stabilization of p18-Cyclin E leading to its increased levels in hematologic cells, such as IM-9 or in cells where p18-Cyclin E is stably expressed. By preventing its proteasomal degradation, p18-Cyclin E may become an important Bortezomib target.