ABSTRACT
DDB1, a subunit of the damaged-DNA binding protein DDB, has been shown to function also as an adaptor for Cul4A, a member of the cullin family of E3 ubiquitin ligase. The Cul4A-DDB1 complex remains associated with the COP9 signalosome, and that interaction is conserved from fission yeast to human. Studies with fission yeast suggested a role of the Pcu4-Ddb1-signalosome complex in the proteolysis of the replication inhibitor Spd1. Here we provide evidence that the function of replication inhibitor proteolysis is conserved in the mammalian DDB1-Cul4A-signalosome complex. We show that small interfering RNA-mediated knockdown of DDB1, CSN1 (a subunit of the signalosome), and Cul4A in mammalian cells causes an accumulation of p27Kip1. Moreover, expression of DDB1 reduces the level of p27Kip1 by increasing its decay rate. The DDB1-induced proteolysis of p27Kip1 requires signalosome and Cul4A, because DDB1 failed to increase the decay rate of p27Kip1 in cells deficient in CSN1 or Cul4A. Surprisingly, the DDB1-induced proteolysis of p27Kip1 also involves Skp2, an F-box protein that allows targeting of p27Kip1 for ubiquitination by the Skp1-Cul1-F-box complex. Moreover, we provide evidence for a physical association between Cul4A, DDB1, and Skp2. We speculate that the F-box protein Skp2, in addition to utilizing Cul1-Skp1, utilizes Cul4A-DDB1 to induce proteolysis of p27Kip1.
The Cul4A gene is amplified and overexpressed in breast and hepatocellular carcinomas (6, 42). Also, Cul4A is essential for mammalian development (18). It encodes a protein of the cullin family. The cullins are central components of several E3 ubiquitin ligases (11). Cul4A associates with the damaged-DNA binding protein DDB (22, 32). DDB consists of two subunits: DDB1 and DDB2. The DDB2 subunit is mutated in xeroderma pigmentosum (complementation group E) (reviewed in reference 35). Cul4A participates in the ubiquitination of the DDB2 subunit of DDB and induces its proteolysis through the ubiquitin-proteasome pathway (22). Recent studies indicated that the DDB1 subunit of DDB functions as an adaptor for substrate binding by Cul4A in a manner similar to how Skp1 functions in the Skp1-cullin1-F-box (SCF) complex (15). However, unlike the case for Skp1, there are instances where DDB1 directly targets a substrate without additional adaptor proteins. For example, Cul4A has been implicated in the proteolysis of the replication licensing protein Cdt1 following DNA damage (14, 44). It was shown that the interaction between Cul4A and Cdt1 is mediated by DDB1 (15). In other examples, Cul4A-DDB1 interacts with additional adaptors to target a specific protein. The DDB1-Cul4A complex associates with hDET1, an ortholog of Arabidopsis De-etiolated-1, and hCOP1, an ortholog of Arabidopsis constitutively photomorphogenic-1 (COP1), to induce proteolysis of the c-Jun protein through the ubiquitin-proteasome pathway (40). In that study, the authors proposed that the hDET1-hCOP1 functioned as the heteromeric substrate adaptor and, in keeping with the SCF E3 ligase, proposed the name DCXhDET1-COP1 as the ligase for c-Jun (40). Similarly, it was shown that the paramyxovirus V protein associated with DDB1 (37). Moreover, the V protein formed a complex with DDB1-Cul4A to induce ubiquitination and proteolysis of the STAT proteins (37). In that study, the authors proposed a role of the viral V protein in linking the STAT proteins to the DDB1-cullin 4A ligase complex and, based on analogy with the SCF complex, termed the V-DDB1-Cul4A complex the VDC complex (37). The interaction of DDB1 with multiple secondary adaptor proteins is not surprising, because DDB1 possesses 17 WD40-like motifs that are involved in protein-protein interaction. Cul4A has been shown to participate in the MDM2-dependent proteolysis of p53 (23). Moreover, Cul4A is involved in the proteolysis of HOXA9 (43). However, the role of DDB1 in the proteolysis of p53 and HOXA9 is yet to be established.
The functions of Cul4A-DDB1 are linked to the COP9 signalosome (CSN) (13). CSN, an eight-subunit protein complex, was first characterized from Arabidopsis thaliana as a regulator for light-dependent development (reviewed in references 30 and 31). More recently, CSNs from a variety of species, ranging from yeasts to humans, has been characterized. CSN possesses significant structural homology with the 19S lid complex of the 26S proteasome and, to a lesser extent, with the eukaryotic translation initiation factor 3 (31). The structural homology with the19S lid complex is interesting because CSN has been shown to participate in proteolysis involving the ubiquitin-proteasome pathway (see reference 29 and references therein). CSN associates with several proteins involved in the ubiquitination pathway, including deubiquitinating enzymes and E3 ubiquitin ligases (45). The plant E3 ligase COP1 associates with CSN (31). The cullin family of E3 ligases found in yeasts to humans associates with CSN (11). It was shown that CSN could regulate the functions of the cullins by removing the NEDD8 modification (see reference 8 and references therein). The CSN subunit CSN5 possesses a metalloprotease activity that appears to be involved in deneddylating the cullins. In addition, fission yeast CSN was shown to suppress the activities of cullins (Pcu1p and Pcu3p) through recruitment of the deubiquitylating enzyme Ubp12p (45). Despite the observations on the negative regulation of the cullins by CSN in vitro, mounting evidence suggests a role of CSN also in positively cooperating with the functions of cullins in mediating ubiquitination and proteolysis in vivo. It was proposed that CSN served as an assembly and maintenance platform for cullin ubiquitin ligases, and the CSN-mediated deubiquitylation and deneddylation were prerequisite for the activation of cullins (45). Moreover, recent protein cross-linking studies suggested the possibility that CSN plays a role in linking the E3 ligases to the proteasome (27).
Cul4A-DDB1 remains bound to CSN (13). Moreover, it was shown that CSN is required for the Cul4A-mediated proteolysis of Cdt1 (14). Studies with fission yeast provided clear genetic evidence for a functional link between CSN and Cul4A-DDB1. For example, it has been shown that the Cul4A, DDB1, and CSN homologs in fission yeast are involved in the DNA damage checkpoint-dependent and -independent proteolysis of the replication inhibitor Spd1 (19). Spd1 is an inhibitor of the ribonucleotide reductase, which is degraded in S phase and in response to DNA damage in fission yeast (2, 19). It was shown that mutations in the CSN subunits CSN1 and CSN2 resulted in strains that were unable to degrade Spd1 (19). It was suggested that the fission yeast CSN cooperated with Pcu4, the Schizosaccharomyces pombe homolog of Cul4A, to induce proteolysis of Spd1 (19). We observed that a fission yeast strain lacking Ddb1, the S. pombe homolog of DDB1, exhibited slow replication and defective progression through S phase. Also, the strain was highly sensitive to UV irradiation in S phase (3). Since Ddb1 associates with CSN, we investigated whether the S phase defects in our strain were a consequence of an inability to degrade Spd1 in S phase. Indeed, the fission yeast strain lacking Ddb1 expression failed to degrade Spd1 in S phase and following UV irradiation, suggesting a role of the Ddb1-Pcu4/CSN interaction in the regulated proteolysis of Spd1 (2).
Since the fission yeast Pcu4 and Ddb1 are involved in the proteolysis of a replication inhibitor, we investigated whether the function is conserved in mammalian cells. In mammalian cells, the replication inhibitor p27Kip1 is degraded by the ubiquitin-proteasome pathway, involving different E3 ligases. For example, in G1 phase of the cell cycle, the proteolysis of p27Kip1 involves KPC1/KPC2 (17). On the other hand, the proteolysis of p27Kip1 in S phase involves the F-box-containing protein Skp2-linked ligase. It was shown that Skp2 links p27Kip1 to the E3 ligase Cul1 through an interaction with Skp1 (reviewed in references 4 and 41). Consistent with that, the Skp2−/− mice accumulate p27Kip1 (24). However, the Cul1−/− embryos do not exhibit accumulation of p27Kip1 (10). These observations suggest the possibility that Skp2 can utilize another E3 ligase(s) to target p27Kip1. Here, we provide evidence that Cul4A and DDB1 associate with Skp2 to target p27Kip1 for proteolysis.
MATERIALS AND METHODS
Cell culture.HeLa, T98G, and Caski cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. The Skp2-deficient Caski cell line was generated by stably transfecting a plasmid that constitutively expresses small hairpin RNA (shRNA) against Skp2. We used the pSUPERretroneo+gfp vector from OligoEngine for expression of Skp2 shRNA. The target sequence for Skp2 was 5′-GAGGAGCCCGACAGTGAGA-3′. The pSUPER-RNAi-SKP2 plasmid was transfected in Caski cells by using Lipofectamine, and the clones were selected with G418. Pools of G418-resistant clones were isolated and were analyzed by Western blot assay to monitor downregulation of Skp2. Independent clones expressing different levels of Skp2 were isolated. The construction of the Flag-Cul4A-expressing 293 cells has been described (23).
Expression vectors and virus.A plasmid expressing T7 epitope-tagged DDB1 has been described before (22). The V5 epitope-tagged Skp1 and Skp2 plasmids were generated by subcloning the respective cDNAs into PCDNA3.1/V5-His vector in frame with the V5 epitope-coding sequence. Recombinant adenovirus expressing the T7-tagged DDB1 was generated following the procedure used for construction of the Adp48 virus (9).
Antibodies.The V5 monoclonal antibody was from Invitrogen. The Skp1 antibody, Skp2 antibody, p27 polyclonal antibody, CSN1 antibody, and α-tubulin antibody were from Santa Cruz Biotechnology. The anti-p27 monoclonal antibody used for immunostaining was from BD Transduction Laboratory. The Alexa Fluor 594-conjugated goat anti-mouse antibody was from Molecular Probes. The T7-Tag antibody and T7-Tag antibody agarose beads were from Novagen.
siRNA transfection.Small interfering RNAs (siRNAs) were used at a final concentration of 100 nM. Logarithmically growing HeLa or T98G cells were seeded at a density of 5 × 105cells/10-cm dish. Cells were transfected using Oligofectamine from Invitrogen. After 8 to 10 h of transfection, the medium was replaced with fresh medium containing 10% fetal bovine serum, and 24 h after transfection, the cells were divided. Cells were harvested between 48 and 60 h after transfection. The siRNAs were synthesized by Dharmacon with the following target sequences: Cul4A, 5′-GAAGAUUAACACGUGCUGGdTdT-3′; Cul4B, 5′-AAGCCUAAAUUACCAGAAAUU-3′; Cul1, 5′ UAGACAUUGGGUUCGCCGUdTdT; CSN1, 5′-AAGUACGCCUCAUGUCUCAAG-3′; DDB1, 5′-AACGGCUGCGUGACCGGACAC-3′; and control scramble siRNA sequence, 5′-AACAGUCGCGUUUGCGACUGG-3′.
Immunoprecipitation and Western blot assay.Western blot analyses were performed following a previously described procedure (23). The procedure for immunoprecipitation of Flag-tagged Cul4A and associated protein and elution with Flag-peptide has been described previously (23). For the coimmunoprecipitation of DDB1 and Skp2, C33A cells were grown in Dulbecco's modified Eagle's medium (GIBCO-BRL) supplemented with 10% fetal bovine serum and were transfected at 60% confluence with T7 epitope-tagged DDB1 (10 μg), Skp2 (3 μg), and V5-tagged Skp1 (5 μg) expression plasmids by using the calcium phosphate coprecipitation method as described previously (22). DNA precipitates were removed 18 h after transfection, and the cells were harvested 30 h later. The cells were treated with MG132 (5 μM) for 6 hours prior to harvesting. Cell lysates were prepared by using extraction buffer containing 50 mM Tris (pH 7.4), 0.1% Triton X-100, 0.25 M NaCl, 5 mM EDTA, 50 mM NaF, 0.1 mM Na-orthovanadate, and aprotinin (4 μg/ml), leupeptin (10 μg/ml), pepstatin (4 μg/ml), and 2 mM phenylmethylsulfonyl fluoride. Cell lysates (1.0 mg) were precleaned with 1 μg of mouse immunoglobulin G (IgG) and protein G-Sepharose and were immunoprecipitated with T7-Tag antibody-agarose beads (20 μl). The immunoprecipitates were washed three times with 0.5% NP-40-containing buffer and resolved by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (10%), followed by immunoblotting with anti-V5 antibody, anti-Skp2 antibody, and anti-T7 antibodies.
p27Kip1 decay rate analysis.HeLa cells were infected with recombinant adenovirus expressing DDB1 (Ad-DDB1) or control virus (Ad-TA). The infected cells were treated with 50 μg/ml of cycloheximide for different time periods of between 45 min and 8 h as described in the figure legends. The cells were lysed in radioimmunoprecipitation assay buffer containing protease inhibitors and phosphatases inhibitors; 50 to 200 μg of cell extract was separated by 10% SDS-polyacrylamide gel electrophoresis and blotted to a nitrocellulose membrane, which was probed with p27 antibody.
Immunostaining for p27Kip1.The cells were grown on glass coverslips and transfected with the control siRNA, DDB1 siRNA, and Cul4A siRNA as described above. Cells were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 20 min at room temperature, washed once with 0.1 M glycine in PBS, and permeabilized for 5 min with 0.1% Triton X in PBS. After fixation, the cells were washed four times with PBS (5 min each) and blocked with 5% goat serum for 1 h at room temperature. Cells were incubated with the monoclonal p27 antibody (1:100) for 4 h at room temperature. The cells were then washed five times with PBS (5 min each), incubated with a 1:500 dilution of Alexa Fluor 594-conjugated goat anti-mouse antibody for 40 min at room temperature, and then washed 10 times with PBS (2 min each). The cell nuclei were labeled with bisbenzamide (2 μg/ml) in PBS for 3 min at room temperature. After a final wash in PBS, the cells were mounted onto slides with Vectashield (Vector) mounting medium and viewed with a Nikon microscope.
RESULTS
siRNA-mediated knockdown of DDB1, CSN1, or cullin 4A causes an accumulation of p27Kip1.The fission yeast Ddb1 and Pcu4 participate in the same pathway with CSN to induce proteolysis of the replication inhibitor Spd1 (2, 19). Spd1 inhibits replication by inhibiting the activity of ribonucleotide reductase (19). In mammalian cells, ribonucleotide reductase is regulated by different mechanisms, and no inhibitor has been identified (see reference 5 and references therein). Therefore, to investigate whether the function of DDB1, Cul4A, and CSN is conserved in mammalian cells, we analyzed the inhibitors of the S phase in mammalian cells. We employed siRNA to knock down the levels of DDB1, Cul4A, or CSN1 to investigate their role in regulating the levels of replication inhibitors in mammalian cells. We decided to target the CSN1 subunit of CSN by siRNA because the yeast mutant lacking CSN1 exhibited a phenotype similar to that observed in a strain lacking expression of Ddb1 (3, 19). HeLa cells were transfected with a control siRNA or with an individual siRNA against DDB1, CSN1, or Cul4A. Forty-eight hours after transfection, cells were harvested and cell extracts were analyzed by Western blot assays. The siRNA-transfected cells exhibited a significant reduction in the levels of the target proteins (Fig. 1). The extracts were also analyzed for the levels of two replication inhibitors, p21Cip1 and p27Kip1. We detected an increase in the level of p21Cip1, but we think that was an indirect effect of p53 stabilization because we did not see any increase in p21Cip1 in p53-negative cells (data not shown). However, the extracts depleted of DDB1, CSN1, and Cul4A consistently exhibited a robust increase in the steady-state levels of p27Kip1 (Fig. 1). Analysis of the DDB1-depleted HeLa cells by flow cytometry showed only about a 10% increase in the G1 population (see Fig. S1 in the supplemental material). The accumulation of p27Kip1 was also investigated by immunostaining experiments. HeLa or T98G (glioblastoma) cells, grown on coverslips, were transfected with control siRNA or siRNA specific for Cul4A or DDB1. Forty-eight hours after transfection, the cells on the coverslips were fixed and subjected to immunostaining using a monoclonal antibody against p27Kip1. As shown in Fig. 2, siRNA-mediated knockdown of Cul4A or DDB1 increased the nuclear abundance of p27Kip1. The T98G glioblastoma cell line does not express functional p53; therefore, the accumulation of p27Kip1 is independent of p53.
siRNA-mediated knockdown of DDB1, Cul4A, or CSN1 increases accumulation of p27Kip1. HeLa cells were transfected with scrambled siRNA (control siRNA) or with siRNA specific for DDB1, Cul4A, or CSN1. Forty-eight hours after transfection, cells were harvested. The cell lysates (200 μg) were subjected to Western blot analyses. The blots were probed with antibodies for Cul4A, DDB1, and CSN1 to assess the extent of knockdown. The blots were also probed with a polyclonal antibody against p27Kip1.
Nuclear accumulation of p27Kip1 in cells transfected with siRNA for Cul4A or DDB1. HeLa cells or T98G glioblastoma cells were grown on coverslips and transfected with control siRNA or siRNA for Cul4A or DDB1. The transfected cells were fixed and subjected to immunostaining. Cells were first incubated with a monoclonal p27 antibody and then incubated with Alexa Fluor 594-conjugated goat anti-mouse antibody, as described in Materials and Methods. The cell nuclei were labeled with bisbenzamide and then viewed with a Nikon fluorescence microscope.
DDB1 expression decreases the steady-state level of p27Kip1 through increased proteolysis.Since the knockdown of the DDB1-Cul4A-CSN complex increased the level of p27Kip1, we predicted that an increase in DDB1 would decrease the level of p27Kip1. To test that prediction, we employed a recombinant adenovirus (DDB1 virus) to express an epitope (T7)-tagged DDB1. At different times after infection, cells were harvested and the extracts were analyzed for the levels of p27Kip1. Expression of DDB1 became detectable 8 to 10 hours following infection, and the level reached its peak by 16 hours (Fig. 3A). Coinciding with the increase in the level of DDB1, there was a decrease in the level of p27Kip1. To investigate whether the decrease in the level of p27Kip1 was a result of increased proteolysis, we compared the decay rates of p27Kip1 in control virus- and DDB1 virus-infected cells. HeLa cells were infected with the TA- or the DDB1-expressing virus. Sixteen hours following infection, the cells were treated with cycloheximide to inhibit protein synthesis. At different times following cycloheximide treatment, cells were harvested and the extracts were analyzed for the level of p27Kip1 by Western blot assays. The p27Kip1 protein exhibited a relatively long half-life in the control virus-infected cells (Fig. 3B). However, in the DDB1-expressing cells p27Kip1 exhibited a much higher decay rate. The increased decay is consistent with an enhanced proteolysis of p27Kip1 in DDB1-expressing cells.
Expression of DDB1 increases the rate of decay of p27Kip1. (A) HeLa cells were infected with a recombinant adenovirus expressing T7 epitope-tagged DDB1. At the indicated time points following infection, cells were harvested and the lysates (200 μg) were subjected to Western blot analyses. (B) HeLa cells were infected with a control adenovirus (TA virus) or DDB1 virus at equal multiplicities of infection (100 PFU/cell). Fourteen hours after infection the cells were treated with cycloheximide, and the cells were harvested at the indicated time points. Cell extracts (200 μg) were subjected to Western blot assays using p27Kip1 antibody or tubulin antibody.
To investigate whether the enhanced proteolysis of p27Kip1 resulted from an effect on the cell cycle progression, we carried out experiments with synchronized cells. We used HaCat cells because these cells can be easily synchronized to G0/G1 phases by serum starvation. The serum-starved HaCat cells synchronously enter S phase at about 18 h following serum supplementation (data not shown). We synchronized the cells by maintaining the cells in medium containing 0.1% fetal bovine serum for 60 h. Following that, the cells were infected with control adenovirus or adenovirus expressing DDB1. Immediately after infection, the cells were stimulated by adding fetal bovine serum (10%) in the culture medium. Sixteen hours after infection and serum stimulation, aliquots of cells infected with the control and the DDB1 expression viruses were subjected to flow cytometric analysis. As shown in Fig. 4, expression of DDB1 had no significant effect on the serum-stimulated progression to S phase. At 16 hours, the cells were also treated with cycloheximide, and at different time points cells were harvested. Extracts of the cycloheximide-treated cells were analyzed for p27Kip1 and p130. Expression of DDB1 reduced the half-life of p27Kip1 in the synchronized cells (Fig. 4). These results are inconsistent with the notion that DDB1-induced proteolysis of p27Kip1 is an indirect effect of cell cycle progression. Interestingly, the half-life of p130, which is degraded by the SCF-Skp2 pathway, was not affected by DDB1 expression, suggesting that DDB1 does not enhance the function of the SCF-Skp2 complex.
DDB1 induces proteolysis of p27Kip1 without stimulating progression of the cell cycle. HaCat keratinocytes were synchronized to G0/G1 phases by maintaining the cells in medium containing 0.1% fetal bovine serum for 60 h. The synchronized cells were then infected with control virus or DDB1 expression virus at 100 PFU/cell. The infected cells were maintained in medium containing 10% fetal bovine serum for 16 h. (A) Cells were treated with cycloheximide for the indicated time periods. Extracts (0.2 mg) of the harvested cells were subjected to Western blot analyses using antibodies against p27Kip1, p130, and tubulin. (B) Aliquots of the serum-stimulated, infected cells at the 16 hours were subjected to propidium iodide staining and flow cytometry. The distribution of cells in different phases of the cell cycle is shown.
DDB1-induced proteolysis of p27Kip1 requires Cul4A and signalosome, but not Cul1.DDB1 exist in a complex with Cul4A and CSN (13). Moreover, we observed that siRNA-mediated knockdown of any of the three components resulted in an accumulation of p27Kip1. Therefore, we predicted that the DDB1-induced proteolysis of p27Kip1 would depend upon availability of Cul4A and CSN. To investigate that possibility, we employed the siRNAs to knock down the level of Cul4A (Fig. 5A) or CSN1 (Fig. 5B) in HeLa cells. Forty-eight hours after transfection of the specific siRNA, the cells from each set of transfections were pooled and replated equally into four plates. Following replating, the cells were infected with a control virus or adenovirus expressing DDB1. Sixteen hours after infection, cycloheximide was added to the culture medium to inhibit protein synthesis. At different times after cycloheximide addition, cells were harvested and the extracts were analyzed for the levels of p27Kip1, Cul4A, and CSN1. Transfection of the specific siRNA caused a significant depletion of the levels of Cul4A (Fig. 5A) and CSN1 (Fig. 5B). Moreover, the depletion of Cul4A or CSN1 blocked DDB1-induced proteolysis of p27Kip1. DDB1 expression increased the decay rate of p27Kip1 in cells transfected with control siRNA, but it failed to increase the decay rate of p27Kip1 in cells transfected with Cul4A siRNA or CSN1 siRNA. These observations are consistent with the notion that Cul4A and CSN are essential in the DDB1-mediated proteolysis of p27Kip1.
DDB1-induced proteolysis of p27Kip1 requires Cul4A and CSN1. HeLa cells were transfected with control siRNA or Cul4A siRNA (A) or CSN1 siRNA (B). Twenty-four hours after siRNA transfection, the transfected cells from each group were pooled and split equally into five plates. Twelve hours after replating, the cells were infected with control virus (TA virus) or DDB1 virus at 100 PFU/cell. Fourteen hours after infection, the cells were treated with cycloheximide, and at the indicated time points cells were harvested and the extracts (200 μg) were analyzed by Western blot assays.
To further investigate whether the DDB1-induced proteolysis of p27Kip1 is independent of the SCF pathway, we investigated p27Kip1 half-life in Cul1 knockdown cells. Cul1 is the central component of the SCF complex. HeLa cells were transfected with control siRNA, Cul1 siRNA, or Cul4A siRNA. The transfected cells were infected with adenovirus expressing DDB1. Sixteen hours after infection, the cells were treated with cycloheximide, and at different time points, the cells were harvested and the extracts were analyzed for p27Kip1 by Western blotting. As shown in Fig. 6, DDB1 virus infection failed to enhance the decay of p27Kip1 in Cul4A siRNA-transfected cells. In contrast, Cul1-depleted cells, like the control siRNA-transfected cells, failed to block DDB1-induced proteolysis of p27Kip1. The observation is consistent with the notion that the Cul4A-DDB1 pathway of p27Kip1 proteolysis is distinct and that it does not involve Cul1.
DDB1 induces proteolysis of p27Kip1 in Cul1-depleted cells. HeLa cells were transfected with control siRNA, Cul1 siRNA, or Cul4A siRNA. Thirty-six hours after transfection, the cells were infected with the DDB1 expression virus at 100 PFU/cell. At the indicated time points, cells were harvested and the extracts were analyzed by Western blots for the levels of p27Kip1.
DDB1 specifically increases neddylation of Cul4A.All cullins are modified by neddylation, in which a small ubiquitin-like molecule, NEDD8, is conjugated to a lysine residue in the cullins (8). It is believed that neddylation is important for the activity of the cullins. The neddylated forms of the cullins exhibit slower mobility during SDS gel electrophoresis, and as a result, the cullins migrate as a doublet in which the slower-migrating form represents the NEDD8-modified cullin. We observed that DDB1 could modulate the neddylation status of Cul4A. For example, infection of cells with adenovirus expressing DDB1 caused an increase in the abundance of the slower-migrating species of Cul4A (Fig. 7A). The effect was specific for Cul4A, because we did not detect any significant change in the neddylation status of Cul1 in the same extracts. To investigate a potential role of DDB1 in altering the neddylation of the other cullins, we transfected cells with plasmids that expressed the V5 epitope-tagged cullins. The cells were then infected with a control virus or with virus expressing DDB1. Extracts of the infected cells were analyzed for the two forms of the cullins by use of a Western blot that was probed with V5 antibody. Clearly, expression of DDB1 caused a relative increase of the slower-migrating form of Cul4A and had no effect on the two forms of the other cullins (Fig. 7B), indicating that DDB1 specifically increases the abundance of the neddylated form of Cul4A. The results are consistent with those of a previous study (15), in which the authors showed that DDB1 binding prevented association of Cul4A with CAND1, an inhibitor of cullin neddylation.
Expression of DDB1 specifically increases the neddylation of Cul4A. (A) HeLa cells were infected with increasing levels of DDB1 virus (0, 50, and 100 PFU/cell) or DDB2 virus (100 PFU/cell) for 16 hours. Extracts (200 μg) of the infected cells were analyzed for Cul4A by Western blot assay. (B) HeLa cells were transfected with a plasmid expressing one of the V5 epitope-tagged cullins. Twelve hours after transfection, the cells were split equally into two plates and infected with the control virus (TA virus, 100 PFU/cell) or a virus expressing T7-tagged DDB1 (100 PFU/cell). Sixteen hours after infection, cells were harvested and the extracts (200 μg) were analyzed for the cullins by using a V5 antibody (Ab) and for the expression of DDB1 by using T7 antibody.
DDB1-induced proteolysis of p27Kip1 is blocked by the depletion of Skp2.The SCF complex involving Skp2 is involved in the proteolysis of p27Kip1 (reviewed in references 4 and 21). The proteolysis of p27Kip1 by the SCF pathway of ubiquitination has been studied extensively. That pathway does not involve Cul4A and DDB1. Also, the experiments in Fig. 6 suggested that Cul1, the central component of the SCF complex, is not involved in the DDB1-induced proteolysis of p27Kip1. To investigate whether or not the DDB1-induced proteolysis depends upon Skp2, we sought to analyze the proteolysis of p27Kip1 in cells depleted of Skp2. We generated a cell line, derived from the human cervical carcinoma Caski cells, which stably expressed shRNA against Skp2 (see Materials and Methods). This is the only cell line in which we could stably express Skp2 shRNA (data not shown). Expression of the shRNA reduced the level of Skp2 in that cell line (Fig. 8) and raised the level of p27Kip1. To determine the effect of Skp2 depletion on the DDB1-induced proteolysis of p27Kip1, we infected the Skp2-depleted cells with adenovirus expressing DDB1. For a comparison, the parental Caski cells (not depleted by shRNA expression) were infected by the DDB1 expression virus. Following 16 hours of infection, the infected cells were treated with cycloheximide. Cells were harvested at different times after cycloheximide addition. Extracts of the infected cells were analyzed for the levels of p27Kip1. As seen in Fig. 8, expression of DDB1 caused an increased decay in the control Caski cells. However, DDB1-induced decay of p27Kip1 was blocked in cells depleted of Skp2. These results suggest that the DDB1-mediated proteolysis of p27Kip1 depends upon Skp2. Similar results were obtained in experiments in which we compared DDB1-induced proteolysis of p27Kip1 in wild-type and Skp2−/− mouse embryo fibroblasts (see Fig. S2 in the supplemental material).
The accelerated proteolysis of p27Kip1 by DDB1 requires Skp2. We generated a Caski cervical cancer cell line that stably expresses the Skp2 siRNA. Skp2-depleted cells or parental cells were infected with a control virus (TA) or virus expressing DDB1. Sixteen hours following infection, the cells were treated with cycloheximide. At the indicated time points the cells were harvested, and the extracts were analyzed for p27Kip1 by Western blot assays. The level of Skp2 was assessed using Skp2 antibody.
Cul4A and DDB1 associate with Skp2.Our analysis indicated that the DDB1-induced proteolysis required Cul4A and Skp2. To investigate whether these proteins could physically associate, we generated a cell line (derived from 293) that expressed a Flag epitope-tagged Cul4A. This cell line expressed Cul4A at about the same level as the parental 293 cells (Fig. 9B). The epitope-tagged Cul4A could be easily immunoprecipitated by a monoclonal antibody against the Flag peptide. We immunoprecipitated the Flag-Cul4A-expressing cells and the parental cells with Flag antibody. The immunoprecipitates from the Flag-Cul4A-expressing cells contained Flag-Cul4A, DDB1, and the CSN subunit CSN1 (Fig. 9A). However, we failed to detect any significant level of Skp2 or p27Kip1 (not shown). We considered the possibility that the association with Skp2 and p27Kip1 might not be stable in the presence of the active proteasome. The CSN complex is very similar to the 19S lid complex of the 26S proteasome. It was suggested that the CSN complex could bring a substrate directly to the proteasome for degradation (29). Therefore, we carried out the immunoprecipitation experiment with cells that were treated with MG132, an inhibitor of the 26S proteasome, for 5 h before harvesting the cells. Interestingly, the Flag-antibody immunoprecipitates from the Flag-Cul4A-expressing cells contained significant levels of the Skp2 and p27Kip1 proteins (Fig. 9). Also, we could easily detect an interaction between DDB1 and Skp2 by coimmunoprecipitation assays with extracts from cells transfected with a plasmid expressing T7 epitope-tagged DDB1 and Skp2 (Fig. 10). To investigate whether the interaction between DDB1 and Skp2 involved Skp1, we looked at binding of DDB1 to Skp1, an essential component of the SCF complex. HeLa cells were transfected with plasmids expressing T7-tagged DDB1, V5-tagged Skp2, and V5-tagged Skp1. The extracts of the transfected cells were immunoprecipitated with T7 antibody or with control IgG. The control IgG immunoprecipitates did not contain any Skp2 (Fig. 10) or Skp1. The T7 immunoprecipitate contained Skp2, but no Skp1 was detected (Fig. 10). Therefore, DDB1 interacts with Skp2 in the absence of Skp1, suggesting that the interaction does not involve the SCF complex. Together the results suggest that Cul4A and DDB1 associate with Skp2 to target p27Kip1 for proteolysis.
Cul4A associates with Skp2 and p27Kip1. A Flag-Cul4A-expressing 293a cell line (23) and the parental 293a cell line were used in this experiment. The cells were incubated with MG132 for 5 h before harvesting. Extracts (2 mg) were subjected to immunoprecipitation (IP) using Flag antibody (Ab)-containing beads (M2 beads). After washing of the beads, the bound proteins were eluted with Flag peptide (23). The immunoprecipitates were analyzed for Flag-Cul4A, CSN,1, and DDB1 (A) or for Skp2 and p27Kip1 (B) by Western blot assays. The extracts (0.1 mg) of the parental 293a and the Flag-Cul4A-expressing cell lines were also analyzed with Cul4A antibody to compare the total levels of Cul4A expressed by those cells.
DDB1 binds to Skp2 but not Skp1. HeLa cells were transfected with the indicated combinations of plasmids expressing T7-tagged DDB1, untagged Skp2, and V5-tagged Skp1. Extracts (2 mg) of the transfected cells were subjected to immunoprecipitation (IP) using T7 antibody (Ab) or Flag antibody. The immunoprecipitates were analyzed with V5 antibody to detect the presence of Skp1 and with Skp2 antibody to detect Skp2. Extracts (100 μg) were also analyzed to detect the level of expression of the transgenes.
DISCUSSION
The work presented here is significant in several ways. First, we show that Cul4A and DDB1 are involved in the proteolysis of the cell cycle inhibitor p27Kip1. The DDB1-induced proteolysis of p27Kip1 is dependent upon the COP9 signalosome. We provide evidence that DDB1 specifically increase the abundance of the active form of Cul4A. We show that the DDB1-induced proteolysis of p27Kip1 requires Skp2. Moreover, Cul4A and DDB1 associate with Skp2 but not Skp1. Our results provide evidence for a potential new pathway of Skp2 in which Skp2 associates with Cul4A and DDB1, possibly in a manner similar to how it associates with Cul1 and Skp1, to target p27Kip1 for proteolysis.
The inhibitor p27Kip1 is regulated mainly at the level of posttranslational modification and degradation by the ubiquitin-proteasome pathway (26). The protein level of p27Kip1 is high in growth-arrested cells. Growth factor- or mitogen-stimulated entry into the cell cycle is associated with a decrease in the protein level of p27Kip1 (7). The decrease in the level of p27Kip1 is a result of increased proteolysis (20). The proteolysis of p27Kip1 in early G1 phase involves CRM1-dependent export of the protein to the cytoplasm (16). It was shown that phosphorylation of the S10 residue of p27Kip1 is critical for the growth factor- and mitogen-stimulated export of p27Kip1 (16). Recent studies identified a cytoplasmic E3 ubiquitin ligase complex of two proteins, KPC1 and KPC2, that is responsible for the proteolysis of p27Kip1 in G1 phase (17). KPC1/KPC2 failed to cause proteolysis of a mutant p27Kip1 that lacks a nuclear localization signal, suggesting that a modification in the nucleus is necessary for the KPC1/KPC2-mediated proteolysis of p27Kip1 in the cytoplasm. However, a clear link between S10 phosphorylation and KPC1/KPC2-mediated proteolysis of p27Kip1 is yet to be established. In addition to CRM1-mediated nuclear export of p27Kip1, a subcomplex of the COP9 signalosome also has been implicated in the nuclear export of p27Kip1. It was shown that the CSN5/JAB1 subunit of signalosome associates with p27Kip1 (36), and a subcomplex of signalosome containing subunits CSN5 to CSN8 is involved in the nuclear export of p27Kip1. The nuclear export function of CSN5 was also linked to the cytoplasmic proteolysis of p27Kip1 (36). In this study, we show that the CSN1 subunit of the signalosome, which is not a component of the export subcomplex of CSN, is involved in the proteolysis of p27Kip1.
In late G1 and in S/G2 phases, p27Kip1 is believed to be degraded in the nucleus through ubiquitination by the Skp2-containing SCF complex (4, 26, 28). The proteolysis by the SCF complex involves phosphorylation of p27Kip1 and T187 residue by the cyclin-Cdk kinases. The F-box-containing protein Skp2 binds p27Kip1 only when it is phosphorylated at T187 residue (33). Skp2 is more abundant in the late G1 and S/G2 phases of the cell cycle when cyclin-Cdks actively phosphorylate p27Kip1. Ubiquitination studies in vitro confirmed the notion that cyclin E-Cdk2- or cyclin A-Cdk2-mediated phosphorylation at T187 is critical for ubiquitination of p27Kip1 (21, 46). Moreover, it was shown that p27Kip1 bound to cyclin A/E-Cdk2 is a better substrate for ubiquitination in vitro. In addition to cyclin-Cdk2, efficient targeting of p27Kip1 by Skp2 also involves a cofactor called Cks1 (12, 34). The Skp2-mediated proteolysis is predominant in late G1 and in S phase, because in early G1 phase both Skp2 and Cks1 are maintained at low levels through proteolysis by the APC/Cdh1 complex (1, 39). It is noteworthy that the nuclear accumulation of DDB1 is cell cycle regulated (23a). We showed that in regenerating mouse liver, in which the hepatocytes synchronously enter the cell cycle, the nuclear abundance of DDB1 becomes detectable at the G1/S boundary, increases in early S phase, and decreases near the end of S phase. Moreover, it was shown that the Cul4A level also increases near the G1/S boundary of the cell cycle (6a). Therefore, we speculate that Cul4A-DDB1 participates in the Skp2-dependent proteolysis of p27Kip1 during the early S phase.
Studies with mice provided genetic evidence for a role of Skp2 in targeting p27Kip1 for proteolysis. Cells from Skp2−/− mice contain enlarged nuclei with polyploidy and multiple centrosomes, and they exhibit a reduced growth rate (24). In addition, the cells exhibit accumulation of p27Kip1. More interestingly, many of the defects in the Skp2−/− mice are reversed by the elimination of p27Kip1 in a Skp2−/− p27−/− double-knockout strain (25). Those results suggested that p27Kip1 might be the critical target of Skp2. If Skp2 targets p27Kip1 only through the SCF complex, then the Cul1−/− cells are expected to exhibit accumulation of p27Kip1. Surprisingly, however, the embryos of the Cul1−/− genotype did not exhibit accumulation of p27Kip1 (10). Therefore, it is likely that Skp2 can participate in another pathway(s) to target p27Kip1. Our observation that Skp2 associates with Cul4A and DDB1 is significant in that regard. The interaction with Skp2 also explains how Cul4A and DDB1 participate in the proteolysis of p27Kip1. However, further in vitro experiments will be important to demonstrate that the Cul4A-DDB1-Skp2 complex functions as an E3 ubiquitin ligase for p27Kip1.
We observed that siRNA-mediated knockdown of the signalosome subunit CSN1 resulted in the accumulation of p27Kip1. Moreover, DDB1-induced decay of p27Kip1 involved signalosome, because the decay of p27Kip1 was inhibited by the knockdown of CSN1. We do not think that the lack of p27Kip1 proteolysis by DDB1 depends upon signalosome-mediated export of p27Kip1 out of the nucleus, because DDB1-induced proteolysis of p27Kip1 was not inhibited by leptomycin B (data not shown). One possibility is that CSN1 and signalosome participate in the proteolysis of p27Kip1 by linking the Cul4A-DDB1-Skp2-p27Kip1 complex to the proteasome. Signalosome is structurally similar to the 19S regulatory subunit of the 26S proteasome. Based on the structural similarities, it was proposed that signalosome could function as a chaperone that brings ubiquitinated proteins to the proteasome (29). Support for such a model was obtained by protein cross-linking studies, in which it was shown that a variety of E3 ligases could be found cross-linked to signalosome and proteasome (27). Studies with lower organisms and plants provided evidence for distinct functions of the signalosome subunits (19, 38). For example, the phenotypes of fission yeast mutants lacking csn1 or csn2 are distinctly different from those of mutants lacking csn5 (19), suggesting that the subunits of signalosome have functions that are independent of the signalosome complex. Interestingly, the phenotypes of the fission yeast csn1 and csn2 mutants are very similar to those of the Ddb1 (homolog of DDB1) mutant, indicating that they function in the same pathway. This observation also suggests that in fission yeast Ddb1, Pcu4, csn1, and csn2 function in a distinct pathway that may or may not involve the other subunits of signalosome to degrade the replication inhibitor Spd1. Our observation that DDB1 and CSN1 are involved in the proteolysis of p27Kip1 confirms that the pathway in yeast is conserved in mammals. It is possible that CSN1 plays a role in the assembly of the Cul4A-DDB1-Skp2-p27Kip1 complex. Clearly, further work will be needed to establish the mechanism by which Cul4A and DDB1 collaborate with CSN1 and Skp2 to induce proteolysis of p27Kip1.
ACKNOWLEDGMENTS
We thank Peter Jackson, Stanford University, for the Skp1 and Skp2 plasmids. We also thank Sudhakar Baluchami and Tanya Stoyanova for providing reagents and technical help.
This work was supported by grants from NCI (CA 088863) and NIA (AG 024138) to P.R. and S.B.
FOOTNOTES
- Received 6 July 2005.
- Returned for modification 18 October 2005.
- Accepted 6 January 2006.
- Copyright © 2006 American Society for Microbiology
