MEFs harboring null (-) and conditional (floxed, f) alleles of the CSN5/Jab1 gene and immortalized by the null allele of p53 were described previously2122. MEFs, NIH3T3 (Arf-null, p53-wild-type) mouse fibroblasts (provided by Drs C. J. Sherr and M. F. Roussel), and 293T human embryonic kidney (HEK) cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2mM glutamine, 100 units/ml of penicillin, and 100 μg/ml of streptomycin (GIBCO/BRL). Transfection with expression vectors was performed using the calcium phosphate-DNA precipitation method42. In the focus formation assay, transfected cells were maintained in medium containing 5% FBS instead of 10% for approximately 2-3 weeks. In the colony formation assay, transfected cells were selected for resistance to puromycin (5 mg/ml) or G418 (1 mg/ml), depending on the type of vector used for transfection for 2-4 weeks. Both foci and colonies were fixed in 4% paraformaldehyde, and stained in 0.4% crystal violet. In some cases, selected cells were sorted for GFP-positive signals with a FACS Aria flow cytometer (Becton Dickinson).

A cDNA fragment containing the entire coding sequences of mouse CDC6 with or without fusion to the sequence of the estrogen receptor (ER)29 was inserted into the HA-tagged expression vector (a gift from Dr. Junichi Fujisawa).

Cell lysis, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and immunoblotting were performed as described previously4344. A mouse monoclonal antibody to CDC6 (180.2) and rabbit polyclonal antibody to cyclin E (M20) were obtained from Santa Cruz. Mouse monoclonal antibodies to g-tubulin (GTU88) and an HA epitope (12CA5) were purchased from Sigma and Boehringer Mannheim, respectively.

Cells grown on coverslips were fixed in 4% paraformaldehyde, permeabilized in 0.5% Triton X-100, stained with rabbit polyclonal and mouse monoclonal antibodies, and incubated with Alexa Fluor 594-conjugated anti-mouse/rabbit IgG (Invitrogen). Chromosomal DNA was stained by a 2-min incubation in 1 mg/ml of Hoechst 33342. Cell samples were viewed using phase contrast or fluorescence microscopy to enumerate fluorescence-positive cells.

A cDNA fragment containing the entire coding sequence of mouse cyclin E was inserted into the pGEX vector (Pharmacia) in-frame with glutathione S-transferase (GST). The expression and purification of GST-fused proteins and the binding conditions used were as described previously44.

Cells were fixed in 0.25% glutaraldehyde, incubated for 16-24 hours in X-Gal solution containing 0.2% X-Gal, 2 mM MgCl₂, 5 mM K₄Fe(CN)₆, and 5 mM K₃Fe(CN)₆, and viewed with a phase contrast microscope.

Cells (ca 10⁶) were subcutaneously injected into 6-week-old NOD-SCID mice. Mice were sacrificed and the sizes of tumors were measured 2-3 weeks post-injection. All mice were maintained in the Nara Institute of Science and Technology Animal Facility in accordance with Nara Institute of Science and Technology guidelines. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Animal Research Management Committee of Nara Institute of Science and Technology (Permit Number: 1310). All surgery was performed under isoflurane anesthesia, and all efforts were made to minimize suffering.

We previously reported that knockout of the CSN5 gene induced premature senescence in mouse fibroblasts, even in the absence of p5321. We found that cyclin E levels were higher in the cytoplasm of these cells22, suggesting that certain cyclin E-interacting proteins are located in the cytoplasm of CSN5-knockout cells and may trigger the senescence-initiating program. We suspected that CDC6 was one such protein and subsequently investigated this possibility. CDC6 has been shown to play an important role in the regulation of DNA replication by assisting in the assembly of the pre-replication complex (RC) at the G1 phase of the cell cycle24. In order to achieve this, CDC6 exerts its effects in the nucleus with the aim of associating with origin recognition complex (ORC) proteins. After the onset of the S phase, CDC6 is phosphorylated, ubiquitinated, translocated to the cytoplasm, and degraded by the proteasome to avoid re-replication262728. If CDC6 plays a role in senescence and is involved in the cytoplasmic location of cyclin E, it may be located in the cytoplasm of senescent cells.

CDC6 was largely located in the nuclei of proliferating NIH3T3 cells and MEFs harboring p53-/- alleles, which are immortal and do not enter senescence, and cytoplasmic staining was weak (Figure 1a, b). This result was consistent with previous findings in which the CDC6 protein exported from the nucleus in proliferating cells was unstable and rapidly degraded through the ubiquitin-proteasome system262728However, the CDC6-specific fluorescent signal was significantly strong in the cytoplasm of CSN5-null (CSN5-/D) MEFs, which were generated by introducing CRE-ER into CSN5-/floxed p53-/- MEFs followed by a treatment with 4OHT (Figure 1a). Approximately 40% of proliferating MEFs contained nuclear CDC6, whereas less than 5% of senescent MEFs showed nucleus-specific CDC6 staining (Figure 1b).

We utilized replicative-senescent cells to determine whether cytoplasmic CDC6 was restricted to CSN5-knockdown cells. Wild-type MEF cells were cultured in vitro, and the subcellular localization of CDC6 was compared between early (less than 5 passages) and late (more than 10 passages) passages (Figure 1c, d). Proliferating pre-senescent MEF cells (early passage) exhibited similar staining patterns to NIH3T3 cells. However, nuclear staining was not strong in late passage MEFs and the CDC6-specific fluorescent signal was significantly strong in the cytoplasm (Figure 1c). Approximately 50% of proliferating MEFs contained nuclear CDC6, whereas less than 5% of senescent MEFs showed clear nuclear CDC6 staining (Figure 1d). Furthermore, cytoplasmic staining of CDC6 in late passage MEFs was resistant to leptomycin B (LMB), a chemical inhibitor to nuclear export (Figure 1c, d), which indicated that the cytoplasmic location of CDC6 was not due to accelerated nuclear export, but rather to cytoplasmic retention.

In order to explore the possible roles of cytoplasmic CDC6, CDC6 was predominantly expressed in the cytoplasm. Since cytoplasmic CDC6 in senescent cells was refractory to the effects of LMB, we enforced the cytoplasmic retention of CDC6 rather than enhancing its export from the nucleus, and fused CDC6 with an ER tag to achieve this29. ER is a nuclear receptor that is exclusively retained in the cytoplasm in the absence of a ligand. Therefore, ER tagging is a very useful method for controlling the intracellular localization of a fused protein. We constructed an expression vector encoding HA-tagged CDC6 (HA-CDC6) or CDC6 tagged with both HA and ER (HA-CDC6-ER), and transfected 293T cells with these vectors. The cell lysates isolated from the transfectants were analyzed by immunoblotting with an antibody to an HA epitope. Figure 2a shows that HA-CDC6 and HA-CDC6-ER proteins were both detected without significant degradation. We then transfected NIH3T3 cells with a control vector and expression vector encoding HA-CDC6 or HA-CDC6-ER together with a marker plasmid containing green fluorescent protein (GFP) cDNA and a neo-resistant gene. Transiently transfected cells were fixed and stained with an antibody to the HA epitope. Figure 2b shows that HA-CDC6 was predominantly located in the nucleus, whereas HA-CDC6-ER was largely located in the cytoplasm.

Since ER is transferred to the nucleus after binding to its ligand (ex., 4-hydroxytamoxifen (4OHT)), we incubated cells transfected with HA-CDC6-ER in the presence and absence of 4OHT for 2 additional days (Figure 2b). The nuclear signal of CDC6 was stronger in the presence than in the absence of 4OHT; however, a large amount of CDC6 still remained in the cytoplasm, which demonstrated that ligand binding transferred some, but not all ER-fused proteins into the nucleus in this system.

We performed a colony-formation assay in order to examine the effects of cytoplasmic CDC6 on cell proliferation. We transfected NIH3T3 cells with HA-CDC6 and HA-CDC6-ER vectors together with a GFP-neo marker plasmid, and selected cells in the presence of G418. After 2 weeks, cells transfected with the HA-CDC6 vector formed as many colonies as those with an empty vector control, whereas HA-CDD6-ER-transfectants formed none (Figure 3a). The incubation of cells transfected with HA-CDC6-ER and a GFP-neo marker in the presence of 4OHT from the beginning of transfection failed to rescue the formation of colonies. These results indicated that cytoplasmic CDC6 dominated the effects of nuclear CDC6. In order to test this hypothesis, we transfected cells with a constant amount of HA-CDC6-ER together with increasing amounts of HA-CDC6 and allowed them to form colonies. The results shown in Figure 3b demonstrated that nuclear CDC6 did not override the inhibitory effects of cytoplasmic CDC6. Thus, it was not the disappearance of CDC6 from the nucleus, but the cytoplasmic localization of CDC6 that induced the inhibition of proliferation.

Although no colonies were formed, significant number of cells remained attached to the plate transfected with HA-CDC6-ER followed by selection in G418, and we performed an assay to measure senescence-associated-beta-galactosidase (SA-b-gal) activity. Figure 4 shows that transfection with HA-CDC6-ER induced a significant number of SA-b-gal-positive cells, while that with HA-CDC6 or a control vector did not. As expected, incubation in the presence of 4OHT from the beginning of transfection failed to suppress the induction of SA-b-gal activity in cells transfected with HA-CDC6-ER.

ER tagging is a widely used method to control the subcellular distribution of a protein between the nucleus and cytoplasm with few side effects. For example, an ER-tagged CRE recombinase is generally used to induce conditional knockout in mice and cells containing floxed sequences in the loci of interest. We previously used the CRE-ER system to conditionally knockout the CSN5 allele without any undesirable side effects 2230, and successfully applied this system to other proteins22. Therefore, the results obtained with ER-tagged CDC6 were attributed to the cytoplasmic localization of the CDC6 protein, but not to the expression of ER.

Taken together, although cells expressing cytoplasmic CDC6 remained alive (Figure 4), their proliferation was inhibited (Figure 3), and a senescence marker was induced (Figure 4). Furthermore, an increase in the amount of nuclear CDC6 did not overcome the cytoplasmic CDC6-mediated inhibition of proliferation (Figure 3). Therefore, we concluded that the cytoplasmic localization of CDC6 triggered the senescence-initiation program.

Cyclin E was detected in the cytoplasm of CSN5-null cells that underwent senescence, and the enforced expression of cyclin E in the cytoplasm induced senescence22. Since CDC6 interacted with cyclin E in vitro and in vivoFigure 5a, b), we investigated the intracellular localization of cyclin E in cells transfected with HA-CDC6/HA-CDC6-ER by immunofluorescent staining with a specific antibody to cyclin E. Figure 5c shows that cyclin E was exclusively located in the nucleus in some cells that were introduced with GFP and HA-CDC6. In contrast, a significant percentage of HA-CDC6-ER-transfected cells had cytoplasmic cyclin E in addition to nuclear cyclin E (Figure 5c). A quantitative analysis revealed that this difference was significant (Figure 5d), indicating that the cytoplasmic expression of CDC6 converted a certain amount of cyclin E from the nucleus to the cytoplasm, and may have initiated a senescence program together with CDC6.

We next investigated whether cytoplasmic CDC6 prevented transformation by an active oncogene in vitro and in vivo. The introduction of the active form of Ras is known to transform immortalized mouse fibroblasts such as NIH3T3 cells, and makes them lose contact inhibition and anchorage-dependent growth, thereby allowing the formation of foci on a monolayer of cells. We utilized this assay to evaluate the anti-tumorigenic activity of cytoplasmic CDC6. We transfected NIH3T3 cells with an expression vector containing the active form of Ha-Ras together with an HA-CDC6 or HA-CDC6-ER vector, maintained the culture for 2 weeks, and then performed staining with crystal violet (Figure 6a, b). The introduction of Ha-ras into NIH3T3 cells induced the formation of foci as expected, and co-transfection with HA-CDC6 did not significantly affect the number or size of foci induced by Ha-ras. In contrast, the co-introduction with HA-CDC6-ER markedly inhibited the formation of foci.

In order to examine the effects of cytoplasmic CDC6 on tumorigenesis in more detail, we transfected NIH3T3 cells with Ha-ras and HA-CDC6/HA-CDC6-ER together with a GFP vector, collected successful transfectants using the GFP signal with a cell sorter, and subcutaneously injected them into NOD-SCID mice (Figure 6c, d). After 2-3 weeks, control NIH3T3 cells (GFP) failed to proliferate within the bodies of mice, whereas cells transformed by ras formed large tumors in all mice injected. Co-transfection with HA-CDC6 did not significantly affect the efficiency of tumorigenesis, size of tumors, or time kinetics of developing tumors. In contrast, the co-introduction of HA-CDC6-ER markedly impaired the formation of tumors. Although the formation of ras-mediated tumors was often observed 1 week after the injection, no trace of tumor cells was detected 4 weeks after the injection of CDC6-ER-co-transfected cells into mice. Thus, we concluded that the cytoplasmic localization of CDC6 prevented the transformation of mouse fibroblasts by the activated ras oncogene.